AI Absorption Ledger / ACN

Accenture

ACN · Q2 2026 · reported 2026-06-18 · revenue $18.72bn

Assessment

Accenture's second covered quarter (Q3 FY2026, ended 2026-05-31) again carries no dollar figure for AI. The advanced AI bookings and revenue figures, withdrawn on the last call before coverage, stay absent. The count of clients starting advanced AI projects is repeated at , the data attach rate at no less than , and the partner bookings measure unchanged, now with the partners named. The CEO adds that AI projects are still small with average size rising, and the framing moved from AI as a tailwind today to AI as a tailwind as it scales. Revenues were , up in local currency, with consulting up .

Every size remains the ledger's inference. Revenue: advanced AI services at ; data projects that follow are measured only by an attach rate on a count of projects and are left unsized, as are modernization attributed to AI readiness and the AI-enabler businesses, which credit AI beside other causes (broader transformations; geopolitical risk). The one new number is not an AI figure: cybersecurity services of in fiscal 2025, given to frame the purchase of a group of operational technology security companies, which lifts the year's acquisition plan to about . Those companies are not AI companies and nothing of their cost is credited to AI; Faculty, the one acquisition that is, closed in the quarter and is not mentioned in any of its sources. The ledger puts its income-statement cost at .

The internal story thinned. The statements of a quarter earlier, that AI improves delivery efficiency and that internal use shows in operating income, are not repeated, and no training hours or headcount of AI professionals are given. The filing's shape changed with them: payroll costs fell as a share of revenues, against the prior-year rate, but non-payroll costs including subcontractors were above theirs, gross margin slipped to from and the workforce grew by in the quarter. The company's own token use is described for the first time, as something it meters and optimizes with an internal platform, without a cost.

Steps from the prior quarter: delivery savings move from directional to described and are unsized, the quarter's one passage being a client example whose savings go to the client; corporate-function savings from described to not-mentioned, unsized, since a silent quarter carries no new size; training from quantified to withdrawn, the hours given on the last call before coverage and the prior call no longer given; and a new toll opens, client budget diverted to AI infrastructure and tokens, which the CEO bounds as not material while saying budgets are not increasing. Pricing is reported as relatively stable where it had improved in some areas, the 10-Q no longer calls managed services demand growing, and consulting revenue lagged consulting bookings, all explained by management without reference to AI. The ledger's toll ranges are , and at the point, each starting from nothing.

Sized channels against the income statement, Q2 2026

9 of 14 channels sized

Each blue mark is one channel's dollars for the quarter; a bar is the range of an estimate. Grey marks are the company's reported lines. The distance between them is the point: how large the AI channel is next to the line it sits in.

New money and old money, Q2 2026

3 new8 expanded3 relabelled

Each channel is tagged once for whether its money existed before language models, from the company's annual report and call at the start of the period. A bar splits one flow's sized dollars by that tag. The incremental total is the part that would not be there without the models: a new channel counts in full, an expanded one only for what AI added, a relabelled one at zero.

Paid for AI$74mn to $1.01bn sized

new $2.4mn to $168mnexpanded $71mn to $843mn

Incremental total $2.9mn to $1.01bnpoint $25mn$262mn in 2 channels has no traced baseline
Revenue arriving through AI$704mn to $1.58bn sized

new $704mn to $1.58bnexpanded not sizedrelabelled not sized

Incremental total $704mn to $1.58bnpoint $1.28bn
Cost imposed, or revenue lost, by others’ AI$0 to $561mn sized

new $0 to $94mnexpanded $0 to $280mnrelabelled $0 to $187mn

Incremental total $0 to $373mnpoint $65mn

The sized total counts every channel the company credits to AI, including relabelled money that existed before language models and the ledger's own estimates for it. The incremental total counts relabelled channels at zero. A flow is split by layer where its dollars sit at more than one: end use, compute sold to builders, and hardware. The same dollar can be a buyer's spend, a cloud's revenue and a chipmaker's revenue, so the layers are never added together.

Paid for AI4 channels · $74mn to $1.01bn sized · $2.9mn to $1.01bn incremental

other

Training and reskilling the workforce for AI

0.22% to 3.1% of the quarter’s revenue

Incremental total: counts at zero at the point and in full at the high end, with no traced baseline.

our inferencedisclosure: withdrawn· motive: narrative-defensive· before LLMs: expanded

Training hours were given on the last call before coverage, in the quarter ended 2025-11-30 (anchor claim acn-anchor-c37), and on the prior call, in Q2 FY2026; this quarter's call, release and 10-Q do not give them, nor the course completions or the count of AI and data professionals given a quarter earlier (, ). A metric given in consecutive quarters up to and including Q2 FY2026, counting the reference quarter, and absent now meets the definition of withdrawn, applied by the same rule as advanced AI revenue; no stop was announced, and the rule does not ask for one. The state steps from quantified to withdrawn. The CFO speaks of continuing to invest to strengthen relevance in the age of AI (claim c32). The size is the ledger's own, , carrying the hours stated on the prior call with a wider range.

Evidence: 2 quotes, 4 figures, 2 confounds, 4 from before coverage

What the company spends to train its people for AI work: training hours, an agentic AI curriculum built with a university, and the build-out of its AI and data workforce. The cost sits in Cost of services as learning and professional development and as the paid time of the people trained; the filing does not separate it.

Why this motive

Carried. The performance-evaluation requirement of Q2 FY2026 is neither repeated nor withdrawn; this quarter's only words are the CFO's, investing to strengthen relevance in the age of AI for long-term leadership (claim c32), which read alone would be exploratory. A change of motive is not taken from an omission.

Before LLMs: expanded

Learning and professional development was in fiscal 2024 for training hours, about an hour, and the annual report attributed that year's rise in hours predominantly to generative AI training; the company had said it was investing in AI over three years. The budget existed before LLMs; its size in a quarter before them is not in the anchor, so no baseline figure is registered. The size is the whole of an activity that existed before.

“We are progressing towards our goal of doubling our deeply skilled data and AI practitioners from 40,000 to 80,000, with an additional 5,000 practitioners as of Q1.”
CEO, prepared remarks, earnings call, 2023-12-19
“With our digital learning platform, we delivered approximately 44 million training hours, an increase of 10% compared with fiscal 2023, predominantly due to generative AI training.”
Filing, business, 10-K periodic report, 2024-10-10
“As you know, we are investing $3 billion in AI over three years.”
CEO, prepared remarks, earnings call, 2023-12-19
“We have nearly reached our goal of 80,000 AI and data professionals, and our people participated in approximately eight million training hours this quarter, with a significant focus on building advanced AI technology and industry skills.”
CEO, prepared remarks, earnings call, 2025-12-18

Figures

  • Training hours, quarter ended 2025-11-30, approximate · 2025-CQ4
  • Training hours completed in the quarter · 2026-CQ1
  • People who completed the Agentic AI Fundamentals program · as-of 2026-02-28
  • AI and data professionals, lower bound · as-of 2026-02-28

What else could explain it

  • line composition: Training hours cover every subject; the share that is AI training is the ledger's assumption.
  • other: No hours, course completions or headcount of AI and data professionals are given this quarter; the size carries the prior quarter's stated hours.

Quotes

“In closing, we remain focused on executing our business and capturing new opportunities for growth while continuing to invest to strengthen our relevance in the age of AI for long-term market leadership.”
c32 · CFO, prepared remarks, earnings call, 2026-06-18
“Cost of services and the related gross margin may be impacted by several factors, including contract profitability, which includes the pricing on the work that we sell, as well as by the investments we make in our business, such as research and development to build assets, platforms and industry and functional solutions and strategic acquisitions, as well as in our people, such as total rewards and learning and professional development.”
c16 · Filing, mdna, 10-Q periodic report, 2026-06-18

By quarter

  • Q1 2026quantified · our inference · narrative-defensive · $65mn to $585mn
  • Q2 2026withdrawn · our inference · narrative-defensive · $41mn to $585mn

vendor bill

AI tools, models and tokens bought for the company's own use

0.01% to 0.9% of the quarter’s revenue

Incremental total: counts in full.

our inferencedisclosure: described· motive: narrative-defensive· before LLMs: new

The channel is spoken of directly for the first time. Asked about clients' token spending, the CEO says the company does a good job itself of knowing how tokens are used and which models suit which problems, and that it built an internal platform to optimize tokens which it now takes to clients (claims c51, c57). No cost is given. The 10-Q names higher facility and technology costs in Asia Pacific without attributing them (claim c15). The size is the ledger's own, , the seat decomposition of the prior quarter on a workforce of with a higher ceiling for metered use. The counterparty is mixed: model developers, software vendors and cloud providers, which no source splits.

Evidence: 3 quotes, 1 figure, 2 confounds, 2 from before coverage

What the company pays model developers, software vendors and cloud providers for the AI tools its people use and for the models and tokens consumed in its own work. Management describes company-wide use of AI tools and, later, managing token use across models; no cost is given. The bill would sit in non-payroll technology costs, which the filing does not break out. The payees are model providers, software vendors and hyperscalers together, so the counterparty is mixed.

Why this motive

Carried from Q1, where the usage mandate (Q1 claim c30) read narrative-defensive. Management now describes the bill as something it manages, choosing models by problem and optimizing tokens with an internal platform (claims c51, c57), without a cost or a purpose beyond doing it well; that adds no tell of its own and contradicts nothing.

Before LLMs: new

The anchor reports no model, token or AI licence bill. The annual report says AI was being applied to client delivery and internal operations, and the line that would hold the cost is non-cancelable commitments of for cloud hosting, software subscriptions and information technology services, with no AI share given.

“We are increasingly applying AI-based technologies, including generative AI, to our services and solutions, to how we deliver work to our clients, and to our own internal operations.”
Filing, risk factors, 10-K periodic report, 2024-10-10
“Our proprietary switchboard allows a user to select the combination of models to address business context or factors like cost or accuracy.”
CEO, prepared remarks, earnings call, 2023-12-19

Figures

  • Workforce at quarter end, approximate · as-of 2026-05-31

What else could explain it

  • bundling: Enterprise agreements with model developers and hyperscalers usually cover internal use and client delivery together, and part of the cost may be recovered in client fees.
  • line composition: The cost would sit in non-payroll costs, in the quarter against ; the filing names subcontractors as the main cause of the increase and technology costs only in Asia Pacific (claim c15).

Quotes

“they're coming to us because we're doing a really good job ourselves of being able to know how you use the tokens, which models you use for which problems. That's something we've been focused on since the very beginning.”
c51 · CEO, qa, earnings call, 2026-06-18
“In addition to looking at it for acquisitions, we're also going to be building more and more, and we've already started with our ecosystem partners, where we're basically going to have IP together that creates solutions that also drive then our services. We're building them ourselves. For example, our tokenomics platform internally, we're now taking to clients, and that's a platform we built to optimize tokens.”
c57 · CEO, qa, earnings call, 2026-06-18
“Operating income decreased as revenue growth was offset by higher non-payroll costs, including an increase in facility and technology costs.”
c15 · Filing, mdna, 10-Q periodic report, 2026-06-18

By quarter

  • Q1 2026described · our inference · narrative-defensive · $2.4mn to $83mn
  • Q2 2026described · our inference · narrative-defensive · $2.4mn to $168mn

other

Amortization and deal costs of acquired AI companies

0% to 0.13% of the quarter’s revenue

Incremental total: counts in full.

Matched line moved : the whole line, not this channel.

our inferencedisclosure: described· motive: exploratory· before LLMs: expanded

Faculty, which closed early in the quarter by the prior call's account, is not mentioned on this call, in the release or in the 10-Q. Management's acquisition commentary is about the security platform: the year's spending plan rises to about from , the CEO says product acquisitions are skewing to areas triggered by AI, and the CFO expects to raise long-term debt (claims c19, c62, c65). The 10-Q shows of consideration for acquisitions completed in the quarter (claim c17). The size is the ledger's own, : amortization of Faculty on a price capped by the quarter's acquisition consideration, plus an allowance for deal costs. Funding is carried as operating cash flow; no debt was raised in the quarter.

Evidence: 6 quotes, 7 figures, 3 confounds, 2 from before coverage

The income-statement cost of buying companies that are themselves AI companies whose product sits in an AI channel: so far Faculty, an AI-native services firm with a decision intelligence product, which closed in March 2026. The cost is amortization of its acquired intangibles and its share of acquisition-related costs; the purchase price is outside the income statement. Every other acquisition management groups under AI (data center engineering, cybersecurity, network data, capital projects, partner implementation firms) is read as an acquisition confound with nothing credited to AI.

Why this motive

Carried: the CEO again frames acquisitions as a way into demand triggered by AI and into new commercial models, with valuations rising (claims c62, c64), and the CFO plans long-term debt for the spending (claim c65): investing for the long term.

Before LLMs: expanded

Buying companies, AI firms among them, predates coverage: went into acquisitions in fiscal 2024, and on the call for the first quarter of fiscal 2024 the CEO said the company was investing in AI acquisitions and named Ammagamma. No cost of Faculty was in this income statement before the deal closed in March 2026. The size is the change AI made, not the whole line.

“As you know, we are investing $3 billion in AI over three years.”
CEO, prepared remarks, earnings call, 2023-12-19
“We are also investing in AI acquisitions. For example, we recently announced our intent to acquire Ammagamma, an Italy-based firm that helps companies advance their uses of AI and generative AI technologies.”
CEO, prepared remarks, earnings call, 2023-12-19

Figures

  • Capital expected to be deployed in acquisitions in fiscal 2026, as of June 2026, approximate · FY2026
  • Consideration for acquisitions completed in the nine months ended 2026-05-31, net of cash acquired · as-of 2026-05-31
  • Consideration for acquisitions completed in the quarter, net of cash acquired · 2026-CQ2
  • Amortization of intangible assets, all acquisitions · 2026-CQ2
  • Amortization of intangible assets, all acquisitions, prior-year quarter · 2025-CQ2
  • Amortization of intangible assets, year-over-year increase · 2026-CQ2
  • Annual recurring revenue of the operational technology security companies being acquired · as-of 2026-06-18

Reported line it is matched to

Amortization of intangible assets for all acquisitions was against , an increase of , and lower than the of the prior quarter. The filing attributes nothing in the line to any one acquisition; Faculty's part of it is not visible.

2026-CQ2: 2025-CQ2:

What else could explain it

  • acquisition: Only Faculty is read as an AI company. The operational technology security companies announced with the results, with of recurring revenue, are security software and have not closed; nothing of their cost is credited to AI.
  • other: Faculty is not named in any of this quarter's sources, and the 10-Q calls the period's acquisitions individually immaterial (claim c17).
  • other: The company did not disclose the purchase price; the estimate bounds it by the 10-Q's consideration for all acquisitions completed in the quarter, , which is net of cash acquired and covers deals the stored sources do not name, and places Faculty's share using a dated press report of the deal value.

Quotes

“On acquisitions, because of the exciting OT cybersecurity acquisitions we announced today, which I'll talk about in a moment, we now expect to deploy approximately $9 billion of capital this year based on the anticipated closing dates of the acquisitions.”
c19 · CEO, prepared remarks, earnings call, 2026-06-18
“We are going to continue to look at those opportunities because particularly as the technology and AI changes so much, clients are looking for more and more opportunities to not have to build things, to not have to try to figure it out themselves.”
c56 · CEO, qa, earnings call, 2026-06-18
“We are skewing, however, toward where we see the demand, which is in areas where we uniquely have the domain expertise the clients needed. It's usually triggered by AI. That's what we're seeing. We're skewing our acquisitions to where we see the biggest growth opportunities right now, and where we see the biggest growth opportunities on the product side are in these areas that are being triggered by AI.”
c62 · CEO, qa, earnings call, 2026-06-18
“The valuations on the services we're seeing are ticking up because they're in demand, and we have different valuations for software and products”
c64 · CEO, qa, earnings call, 2026-06-18
“we expect to access the long-term debt market to increase our liquidity for M&A spend and general corporate purposes”
c65 · CFO, prepared remarks, earnings call, 2026-06-18
“During the nine months ended May 31, 2026, we completed individually immaterial acquisitions for total consideration of $2,815,388, net of cash acquired.”
c17 · Filing, notes, 10-Q periodic report, 2026-06-18

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026described · our inference · exploratory · $521k to $24mn

engineering

Building proprietary AI platforms and assets

0.16% to 1.3% of the quarter’s revenue

Incremental total: counts at zero at the point and in full at the high end, with no traced baseline.

our inferencedisclosure: described· motive: exploratory· before LLMs: expanded

The CEO describes a build agenda beside the acquisitions: joint intellectual property with ecosystem partners, products the company builds itself, and a tokenomics platform built for internal use and now taken to clients (claims c57, c56). Proprietary platforms are again what managed services clients are said to lean on (claim c36). No amount is given. The size is the ledger's own, , on the prior quarter's assumptions.

Evidence: 6 quotes, 2 confounds, 3 from before coverage

What the company spends building its own AI platforms, agents and reusable assets: the platforms behind managed services and fixed-price work, an agentic marketing console, a platform for managing token use, and solutions built jointly with partners. The cost is research and development inside Cost of services, which the filing names as a factor in gross margin and does not quantify in a quarter.

Why this motive

Carried: the CEO says the company will build more and more itself and with partners, and names an internal token platform now offered to clients (claim c57), with no revenue tied to it; the CFO frames spending as investment for long-term leadership (claim c32).

Before LLMs: expanded

Research and development was in fiscal 2024, about a quarter, spent on assets, platforms and industry solutions that already included automation platforms and one named for AI. The share directed at AI is not given at the anchor or since, so no baseline figure is registered. The size is the whole of an activity that existed before.

“As you know, we are investing $3 billion in AI over three years.”
CEO, prepared remarks, earnings call, 2023-12-19
“together with our assets and platforms including myWizard, myNav, SynOps and AI Navigator for Enterprise to deliver tangible value for our clients.”
Filing, business, 10-K periodic report, 2024-10-10
“Our proprietary switchboard allows a user to select the combination of models to address business context or factors like cost or accuracy.”
CEO, prepared remarks, earnings call, 2023-12-19

What else could explain it

  • line composition: Research and development sits inside cost of services and is disclosed once a year; the AI share is an assumption.
  • other: The base is the fiscal 2024 figure from the anchor; later spending is not in the stored sources.

Quotes

“Cost of services and the related gross margin may be impacted by several factors, including contract profitability, which includes the pricing on the work that we sell, as well as by the investments we make in our business, such as research and development to build assets, platforms and industry and functional solutions and strategic acquisitions, as well as in our people, such as total rewards and learning and professional development.”
c16 · Filing, mdna, 10-Q periodic report, 2026-06-18
“Clients continue to be focused on transforming their operations through technology, AI and data, and leveraging our proprietary assets and platforms and talent to drive productivity and cost savings.”
c5 · Filing, mdna, 10-Q periodic report, 2026-06-18
“In closing, we remain focused on executing our business and capturing new opportunities for growth while continuing to invest to strengthen our relevance in the age of AI for long-term market leadership.”
c32 · CFO, prepared remarks, earnings call, 2026-06-18
“Second, clients continue to look to reinvent faster, leverage our proprietary platforms and expertise, and achieve greater efficiencies and growth, including through managed services across the enterprise. We're seeing the nature of these programs with managed services evolve, with clients asking for more consulting and AI expertise within them.”
c36 · CEO, prepared remarks, earnings call, 2026-06-18
“We are going to continue to look at those opportunities because particularly as the technology and AI changes so much, clients are looking for more and more opportunities to not have to build things, to not have to try to figure it out themselves.”
c56 · CEO, qa, earnings call, 2026-06-18
“In addition to looking at it for acquisitions, we're also going to be building more and more, and we've already started with our ecosystem partners, where we're basically going to have IP together that creates solutions that also drive then our services. We're building them ourselves. For example, our tokenomics platform internally, we're now taking to clients, and that's a platform we built to optimize tokens.”
c57 · CEO, qa, earnings call, 2026-06-18

By quarter

  • Q1 2026described · our inference · exploratory · $30mn to $234mn
  • Q2 2026described · our inference · exploratory · $30mn to $234mn

Cost displaced by AI2 channels · 2 not sized

cost of-revenue · cheap to verify

Delivery cost avoided through AI in client work

Not sized

No source this quarter makes a statement about AI in the company's own delivery: the one passage is a client example in which the savings go to the client to reinvest (claim c37), and a client's saving is not the company's. The fixed-price remark (claims c67, c68) names no AI saving. With nothing from the company to size, the reading gets no ballpark (applied in the rules sweep of 2026-10-06). The former estimate, acn-2026-cq2-f79, carried the prior quarter's construction.

Matched line moved : the whole line, not this channel.

described, no sizedisclosure: described· motive: narrative-defensive· before LLMs: expanded

The direct statement of a quarter earlier, that AI is applied in delivery and efficiencies keep improving, is not repeated. What the call offers is a client example, a managed services program with agentic AI embedded throughout in which automation replaces manual effort and the savings go to the client to reinvest (claim c37), and the CFO's remark that fixed-price work is over and rising with margins little different from other contracts (claims c67, c68). The filing's shape changed: gross margin slipped to from , payroll costs fell as a share of revenues and subcontractor costs rose, and the workforce grew by in the quarter. The client's savings are the client's, so the reading is unsized. The counterparty is mixed: the company’s own delivery staff and the subcontractors in cost of services.

Evidence: 12 quotes, 23 figures, 4 confounds, 5 from before coverage

Cost of services the company avoids by using AI in the work it delivers to clients: code conversion and software delivery, managed services operations, and the proprietary platforms behind fixed-price work. The displaced cost is its own payroll and subcontractors in Cost of services, which the filing explains by payroll and non-payroll costs without naming AI. Under fixed-price contracts a saving stays with the company; under time-based contracts it leaves as fewer billed hours and is read in the toll channels.

Why this motive

Carried, with the conflict unchanged. The filing again shows an efficiency tell, lower payroll costs as a share of revenues (claims c8, c9); management says nothing this quarter about AI in its own delivery beyond a client example, quantifies no saving, and the workforce grew from a year earlier. The less durable motive is kept: narrative-defensive.

Before LLMs: expanded

Cost of services was in fiscal 2024, of revenues, with utilization at and a workforce of . The annual report already listed delivery efficiencies from people mix and technology and said AI, including generative AI, was being applied to client delivery; the CFO said automation and value-based projects had loosened the link between headcount and revenue. No saving was quantified. The size is the change AI made, not the whole line.

“We are increasingly applying AI-based technologies, including generative AI, to our services and solutions, to how we deliver work to our clients, and to our own internal operations.”
Filing, risk factors, 10-K periodic report, 2024-10-10
“We strive to adjust pricing as well as drive cost and delivery efficiencies, such as changing the mix of people and utilizing technology, to reduce the impact of compensation increases on our margin and contract profitability.”
Filing, mdna, 10-K periodic report, 2024-10-10
“As it relates to the revenue per head and the non-linearity, I mean, we do have automation. We do have value-based projects. While there still is, obviously, a connection to the amount of people that we have, we have been able to break that.”
CFO, qa, earnings call, 2023-12-19
“together with our assets and platforms including myWizard, myNav, SynOps and AI Navigator for Enterprise to deliver tangible value for our clients.”
Filing, business, 10-K periodic report, 2024-10-10
“And then with respect to pricing and passing it along to our clients, remember, this has been a business model for the industry, really all the way going back to the introduction of technology and RPA, right, where we're signing contracts that depend on our use of more technology over time to provide productivity. And so that's still the commercial model of the industry right now.”
CEO, qa, earnings call, 2025-12-18

Figures

  • Utilization · 2026-CQ2
  • Utilization, prior-year quarter · 2025-CQ2
  • Workforce at quarter end, approximate · as-of 2026-05-31
  • Workforce a year earlier, approximate · as-of 2025-05-31
  • Workforce, growth from a year earlier · as-of 2026-05-31
  • Workforce added since the prior quarter end · 2026-CQ2
  • Revenues per member of the quarter-end workforce · 2026-CQ2
  • Revenues per member of the quarter-end workforce, prior-year quarter · 2025-CQ2
  • Revenues per member of the workforce, year-over-year growth in U.S. dollars · 2026-CQ2
  • Gross margin · 2026-CQ2
  • Gross margin, prior-year quarter · 2025-CQ2
  • Payroll costs, all functions · 2026-CQ2
  • Payroll costs, all functions, prior-year quarter · 2025-CQ2
  • Payroll costs as a share of revenues · 2026-CQ2
  • Payroll costs as a share of revenues, prior-year quarter · 2025-CQ2
  • Payroll costs against the prior-year share of revenues (negative is below the prior-year rate) · 2026-CQ2
  • Non-payroll costs including subcontractor costs · 2026-CQ2
  • Non-payroll costs including subcontractor costs, prior-year quarter · 2025-CQ2
  • Non-payroll costs as a share of revenues · 2026-CQ2
  • Non-payroll costs as a share of revenues, prior-year quarter · 2025-CQ2
  • Non-payroll costs against the prior-year share of revenues (positive is above the prior-year rate) · 2026-CQ2
  • Share of work that is fixed price, as management states it (given a quarter earlier as a share of bookings), lower bound · 2026-CQ2
  • Business optimization costs under the program completed in the fiscal first quarter, employee severance part · as-of 2025-11-30

Reported line it is matched to

Cost of services was , of revenues against ; relative to the prior-year share the line is (a positive figure is above it), where a quarter earlier it was . Inside it the composition moved: payroll costs across all functions are relative to their prior-year share of revenues and non-payroll costs, which include subcontractors, . The filing gives lower payroll costs offset by higher subcontractor costs as the explanation (claim c8) and does not mention AI.

2026-CQ2: 2025-CQ2: 2026-CQ2: 2025-CQ2: 2026-CQ2: 2025-CQ2: 2026-CQ2: 2025-CQ2: 2026-CQ2: 2025-CQ2:

What else could explain it

  • mix shift: Payroll fell and subcontractor costs rose by about the same share of revenues; work moved between the company's own staff and subcontractors, which changes the payroll line without any tool.
  • fx: In EMEA the filing ties lower payroll costs as a share of revenues to currency translation (claim c14).
  • transformation program: The business optimization program completed in the fiscal first quarter, of severance in all (claim c18), lowers payroll for reasons management does not attribute to AI.
  • operating leverage: Utilization was against .

Quotes

“We're expanding our partnership to make that possible, consolidating fragmented operations across critical business functions into a unified managed services model with agentic AI embedded throughout and humans in the lead. The result is automation replacing manual effort, faster speed to market, and significant cost savings and productivity gains that Bath & Body Works can reinvest directly into growth.”
c37 · CEO, prepared remarks, earnings call, 2026-06-18
“our goal is always to continue to drive improved gross margins, improved SG&A, while we invest significantly in our business.”
c48 · CFO, qa, earnings call, 2026-06-18
“We continue to see our fixed price work be over 60% and continuing to increase.”
c67 · CFO, qa, earnings call, 2026-06-18
“Margin's not a big difference that I would call out relative to fixed price versus the other commercial constructs, but it is embedded in our 20 basis points of exp ansion for the year.”
c68 · CFO, qa, earnings call, 2026-06-18
“The decrease in gross margin was primarily due to higher non-payroll costs, including higher subcontractor costs, largely offset by lower payroll costs.”
c8 · Filing, mdna, 10-Q periodic report, 2026-06-18
“The increase in gross margin was primarily due to lower payroll costs, partially offset by an increase in non-payroll costs.”
c9 · Filing, mdna, 10-Q periodic report, 2026-06-18
“Utilization for the third quarter of fiscal 2026 was 93%, compared to 92% in the third quarter of fiscal 2025.”
c10 · Filing, mdna, 10-Q periodic report, 2026-06-18
“Our workforce, the majority of which serves our clients, was approximately 799,000 as of May 31, 2026, compared to approximately 779,000 as of August 31, 2025 and 791,000 as of May 31, 2025.”
c11 · Filing, mdna, 10-Q periodic report, 2026-06-18
“We strive to adjust pricing as well as drive cost and delivery efficiencies, such as changing the mix of people and utilizing technology, to reduce the impact of compensation increases on our margin and contract profitability.”
c12 · Filing, mdna, 10-Q periodic report, 2026-06-18
“Additionally, payroll costs for our geographic markets increased in line with revenues, except as described below.”
c13 · Filing, mdna, 10-Q periodic report, 2026-06-18
“Operating income increased due to revenue growth in local currency and the positive impact of foreign currency exchange rates, which resulted in an increase in U.S. dollar revenues and lower payroll costs as a percentage of revenues, partially offset by higher non-payroll costs, including an increase in sub-contractor costs.”
c14 · Filing, mdna, 10-Q periodic report, 2026-06-18
“We recorded a total of $923 million under the program, including $628 million of employee severance and $295 million primarily related to the divestiture of two acquisitions in the Americas.”
c18 · Filing, notes, 10-Q periodic report, 2026-06-18

By quarter

  • Q1 2026direction only · our inference · narrative-defensive · $7.6mn to $483mn
  • Q2 2026described · described, no size · narrative-defensive

back office · cheap to verify

Corporate function cost avoided through the company's own AI

Not sized

No source this quarter mentions AI in the company's own corporate functions, and a not-mentioned reading carries no new size (the methodology's reading of a silent quarter, applied in the rules sweep of 2026-10-06). The former estimate, acn-2026-cq2-f87, carried the prior quarter's reference-class share onto this quarter's corporate lines.

Matched line moved : the whole line, not this channel.

inscrutabledisclosure: not mentioned· motive: narrative-defensive· before LLMs: expanded

Neither the call, the release nor the 10-Q says anything this quarter about AI in the company's own corporate functions, and the claim of a quarter earlier, that running its own services internally shows in operating income, is not repeated. The lines moved the way it predicted without anyone saying why: sales and marketing was of revenues against , general and administrative costs against , and the two together relative to their prior-year share (a negative figure is below it), with operating margin at against . State steps from described to not-mentioned, and the reading is unsized: a silent quarter carries no new size.

Evidence: 0 quotes, 11 figures, 2 confounds, 2 from before coverage

Cost the company says it avoids by running its own services and AI tools in how it operates: sales support, finance, HR, IT and other non-client-facing work. The displaced cost sits in Sales and marketing and in General and administrative costs, which the filing explains by selling costs, payroll and non-payroll costs without naming AI.

Why this motive

Carried from Q2 FY2026, where the tells conflicted and the less durable motive was taken. Silence is not evidence about motive.

Before LLMs: expanded

Sales and marketing was and General and administrative costs in fiscal 2024, together of revenues. A cost program to transform non-billable corporate functions ran from fiscal 2023 to fiscal 2024 at a total of , mostly severance, with no reference to AI, and the annual report said AI was being applied to internal operations. No saving was quantified. The size is the change AI made, not the whole line.

“We are increasingly applying AI-based technologies, including generative AI, to our services and solutions, to how we deliver work to our clients, and to our own internal operations.”
Filing, risk factors, 10-K periodic report, 2024-10-10
“During the second quarter of fiscal 2023, we initiated actions to streamline our operations, transform our non-billable corporate functions and consolidate our office space to reduce costs.”
Filing, notes, 10-K periodic report, 2024-10-10

Figures

  • Sales and marketing · 2026-CQ2
  • Sales and marketing, prior-year quarter · 2025-CQ2
  • General and administrative costs · 2026-CQ2
  • General and administrative costs, prior-year quarter · 2025-CQ2
  • Sales and marketing as a share of revenues · 2026-CQ2
  • Sales and marketing as a share of revenues, prior-year quarter · 2025-CQ2
  • General and administrative costs as a share of revenues · 2026-CQ2
  • General and administrative costs as a share of revenues, prior-year quarter · 2025-CQ2
  • Sales and marketing plus general and administrative costs against the prior-year share of revenues (negative is below the prior-year rate) · 2026-CQ2
  • Operating margin · 2026-CQ2
  • Operating margin, prior-year quarter · 2025-CQ2

Reported line it is matched to

Sales and marketing was of revenues against and general and administrative costs against ; together the two lines are relative to their prior-year share of revenues (a negative figure is below it). The filing explains the lower sales and marketing ratio by selling costs and does not mention AI.

2026-CQ2: 2025-CQ2: 2026-CQ2: 2025-CQ2: 2026-CQ2: 2025-CQ2: 2025-CQ2:

What else could explain it

  • other: The CFO's one remark on overhead this quarter is a general goal of improving the ratio, with no reference to AI.
  • transformation program: The business optimization program completed in the fiscal first quarter, of severance in all, lowers corporate payroll for reasons management does not attribute to AI.

By quarter

  • Q1 2026described · our inference · narrative-defensive · $3.0mn to $59mn
  • Q2 2026not mentioned · inscrutable · narrative-defensive

Revenue arriving through AI5 channels · $704mn to $1.58bn sized · $704mn to $1.58bn incremental · 3 not sized

product revenue

Advanced AI services: generative and agentic AI projects

3.8% to 8.5% of the quarter’s revenue

Incremental total: counts in full.

our inferencedisclosure: withdrawn· motive: narrative-defensive· before LLMs: new

A second covered quarter without the advanced AI bookings and revenue figures, whose withdrawal was announced on the last call before coverage (anchor claim acn-anchor-c32); the state stays withdrawn and no step is recorded. No other dollar figure for AI is given. The count is repeated, another clients starting advanced AI projects, and the CEO adds that AI projects are still small with average size rising steadily (claims c24, c60); a list of large wins and a client result are offered as evidence (claims c22, c43). The framing moved from the present to the future: a quarter earlier AI was a tailwind helping the company win today, and now it will be a tailwind as it scales (claims c21, c61). Consulting revenues grew in local currency. The size is the ledger's own, , the last reported level of carried forward two quarters under assumed growth, against for the prior quarter on the same basis.

Evidence: 28 quotes, 8 figures, 4 confounds, 15 from before coverage

Revenue from consulting and managed services work whose subject is advanced AI: generative and agentic AI strategy, build and deployment for clients, dedicated AI hubs run on their behalf, and agentic commerce work in Song. Management once reported this stream in dollars, as generative AI sales and later as advanced AI bookings and revenue; in the covered quarters it gives client counts and examples and no dollar figure. The work being automated is the client's.

Why this motive

Carried, with the conflict unchanged. Management again counts more clients starting advanced AI work and names large wins (claims c24, c22), which would read offensive; a second quarter passes with no dollar of AI bookings or revenue, the CEO describes the projects as still small (claim c60), and the release and the CFO each mention AI demand without a figure (claims c1, c45): AI talk heavy relative to any quantified dollar. The reason given before coverage for stopping the metric (anchor claim acn-anchor-c33) is management's account of the disclosure, not evidence about the money. The less durable motive is kept: narrative-defensive.

Before LLMs: new

The work is generative and agentic AI projects, which cannot exist without a language model, so the money is new whatever practice delivers it (the first step of the novelty test; the metric as defined leaves out data, classical AI and RPA). At the anchor this money was generative AI sales, a bookings figure the CEO called pure because it left out data work and pull-through: in all of fiscal 2023 and over in the first quarter of fiscal 2024, sold from a data and AI practice that already had people. Consulting revenues were in fiscal 2024. The metric, later named advanced AI and defined to leave out data, classical AI and RPA, was reported every quarter up to the one before coverage, when bookings were and revenue about , of revenues; on that call management said it was the last quarter the figures would be shared.

“Last quarter, we shared that we had sold approximately 300 projects with $300 million in sales in all of FY 2023. Demand continued to accelerate in Q1, with over $450 million in Gen AI sales.”
CEO, prepared remarks, earnings call, 2023-12-19
“I'll just remind you, that's not the pull-through, that's not data. We are very pure because we really want to be sharing with all of you, where is GenAI in the market?”
CEO, qa, earnings call, 2023-12-19
“We are progressing towards our goal of doubling our deeply skilled data and AI practitioners from 40,000 to 80,000, with an additional 5,000 practitioners as of Q1.”
CEO, prepared remarks, earnings call, 2023-12-19
“At the same time, we see AI as the new digital. Like digital, AI is both a technology and a new way of working, and its full value will only come from strategies built on both productivity and growth. And we believe it will be used in every part of the enterprise. We also believe the introduction of generative AI signifies a transformative era that is set to drive growth for us and our clients.”
Filing, business, 10-K periodic report, 2024-10-10
“Our consulting revenue continues to be driven by helping our clients accelerate their reinvention, in particular technology, data, and AI led digital transformations.”
Filing, mdna, 10-K periodic report, 2024-10-10
“Our advanced AI bookings this quarter were $2.2 billion, nearly doubling from Q1 last year and also up from Q4. Revenue reached another milestone this quarter at approximately $1.1 billion.”
CEO, prepared remarks, earnings call, 2025-12-18
“Advanced AI new bookings of $2.2 billion”
Filing, press release, 8-K earnings release, 2025-12-18
“as you know, we were the first in our industry to share our bookings and revenue from advanced AI, which we define as Gen AI, Agentic AI, and Physical AI, and does not include data, classical AI, or RPA.”
CEO, prepared remarks, earnings call, 2025-12-18
“We introduced the metrics in Q3 FY23, just months after Gen AI burst onto the scene, initially to size the reality of the opportunity and to demonstrate our early leadership. At that time, bookings were about $100 million across roughly 100 projects, and revenue was immaterial. We have measured it consistently since that time.”
CEO, prepared remarks, earnings call, 2025-12-18
“To date, we have now delivered approximately $11.5 billion in bookings across 11,000 projects with revenue of $4.8 billion.”
CEO, prepared remarks, earnings call, 2025-12-18
“This will be the last quarter in which we share these specific metrics.”
CEO, prepared remarks, earnings call, 2025-12-18
“We're shifting to more scaled end-to-end solutions that integrate multiple forms of AI, and it has become less meaningful to isolate the data specifically for advanced AI as it does not reflect how the demand is evolving on the ground, the full scope of our AI work, or the value we're creating.”
CEO, prepared remarks, earnings call, 2025-12-18
“Over the last nine quarters, we've seen about 100 incremental clients initiate advanced AI projects with us each quarter.”
CEO, prepared remarks, earnings call, 2025-12-18
“The way I would really just think about it, so it's not meant to be some new metric in that way. What it's really showing is just how rapidly it's moving: 100 clients initiating a quarter, and that it's at the same time really, really early, right, when you think of our whole client base. But I wouldn't start now creating kind of a new metric in that.”
CEO, qa, earnings call, 2025-12-18
“We've now reached a point where advanced AI is being embedded in some way across nearly everything we do, and many of our clients are focusing on moving beyond standalone proofs of concept or initiatives.”
CEO, prepared remarks, earnings call, 2025-12-18

Figures

  • Advanced AI revenue, quarter ended 2025-11-30 (the last quarter reported), approximate · 2025-CQ4
  • Clients initiating advanced AI projects in the quarter · 2026-CQ2
  • Productivity gain of a client's marketing content teams from an AI engine built with the company · 2026-CQ2
  • Consulting revenues · 2026-CQ2
  • Consulting revenues, year-over-year growth in local currency · 2026-CQ2
  • Consulting new bookings, year-over-year growth in local currency · 2026-CQ2
  • Revenues · 2026-CQ2
  • Revenues, year-over-year growth in local currency · 2026-CQ2

What else could explain it

  • relabel: Advanced AI was defined on the last call before coverage (anchor claim acn-anchor-c29) and is not redefined in the covered sources; the CEO now speaks of clients running on AI inside larger programs (claim c23), which widens what could be counted.
  • mix shift: The CEO says client budgets are not increasing and are being spent differently (claim c54); AI work that replaces other work inside a flat budget is not additional revenue for the company.
  • acquisition: Faculty, an AI-native services firm, closed early in the quarter by the prior call's account; its revenue is in this quarter and is not disclosed.
  • other: The size carries forward the last level the company reported, ; the growth since is the ledger's assumption, and management said on stopping the metric that it no longer reflected the full scope of its AI work (anchor claim acn-anchor-c33).

Quotes

“Demand for large-scale reinvention remains strong — 104 quarterly client bookings of $100 million or more year-to-date, up 13% — and we are seeing more large-scale AI transformation programs, while executing our strategy to capture new areas of growth.”
c1 · Filing, press release, 8-K earnings release, 2026-06-18
“driven by helping our clients accelerate their reinvention, leveraging cloud, enterprise platforms, security, AI and data, including advanced AI”
c4 · Filing, mdna, 10-Q periodic report, 2026-06-18
“Clients continue to be focused on transforming their operations through technology, AI and data, and leveraging our proprietary assets and platforms and talent to drive productivity and cost savings.”
c5 · Filing, mdna, 10-Q periodic report, 2026-06-18
“I want to provide a few examples of how we're capturing new areas of demand in the age of AI, what we're doing to expand our total addressable market, and the areas where we are shifting to more non-FTE commercial models over time.”
c20 · CEO, prepared remarks, earnings call, 2026-06-18
“We believe that AI will be a tailwind for us and our industry as it scales, because it is a catalyst for reinvention and is creating new opportunities for growth and efficiency for our clients and for us.”
c21 · CEO, prepared remarks, earnings call, 2026-06-18
“We are starting to see clients who have more advanced digital cores move to larger AI transformation programs. You can see this demand in several significant AI-focused wins across multiple industries and markets, which we publicly announced with companies like BT Group, Mitsubishi Chemical, NSK, Piraeus, Stellantis, TEPCO, Vodafone, and the Women's Tennis Association.”
c22 · CEO, prepared remarks, earnings call, 2026-06-18
“The major theme of all of these programs is that we are moving clients from using AI to running on AI.”
c23 · CEO, prepared remarks, earnings call, 2026-06-18
“We are also seeing more clients move from pilots to production, and all of this is happening even as AI is still in the early innings. This quarter, we saw another 100 clients initiate advanced AI projects with us.”
c24 · CEO, prepared remarks, earnings call, 2026-06-18
“Together, these actions show how we are building a strong foundation for us to win in AI, expanding our addressable market across new growth areas and client segments, and evolving towards more non-FTE revenue over time.”
c31 · CEO, prepared remarks, earnings call, 2026-06-18
“The reason is that we are truly the only company that can cover, at scale, everything from the AI and technology foundation to reinventing nearly every part of the enterprise.”
c33 · CEO, prepared remarks, earnings call, 2026-06-18
“Second, clients continue to look to reinvent faster, leverage our proprietary platforms and expertise, and achieve greater efficiencies and growth, including through managed services across the enterprise. We're seeing the nature of these programs with managed services evolve, with clients asking for more consulting and AI expertise within them.”
c36 · CEO, prepared remarks, earnings call, 2026-06-18
“We're expanding our partnership to make that possible, consolidating fragmented operations across critical business functions into a unified managed services model with agentic AI embedded throughout and humans in the lead. The result is automation replacing manual effort, faster speed to market, and significant cost savings and productivity gains that Bath & Body Works can reinvest directly into growth.”
c37 · CEO, prepared remarks, earnings call, 2026-06-18
“Finally, clients with more advanced digital cores are starting to take on larger AI programs. Exciting green shoots. These large-scale AI programs are complex, to make advanced AI work, deep industry and functional knowledge is needed in addition to technology and AI expertise.”
c39 · CEO, prepared remarks, earnings call, 2026-06-18
“We're embedding AI directly into the core of how they operate, building on their existing network intelligence, customer data, and service management platform.”
c40 · CEO, prepared remarks, earnings call, 2026-06-18
“AIOps capabilities with autonomous agents will detect, route, and resolve incidents with self-healing that accelerates how quickly issues are resolved.”
c41 · CEO, prepared remarks, earnings call, 2026-06-18
“Together with a leading large language model provider and hyperscaler, we built an AI engine that validates and enriches leads, generates personalized campaign content, and automates brand and legal validations.”
c42 · CEO, prepared remarks, earnings call, 2026-06-18
“In B2B sales and marketing, conversion rates increased and drove net new revenue. Lead accuracy jumped from 13%-97%. Campaign speed to market improved by 55%, and marketing content teams are 40% more productive, with capacity freed to drive further growth. This is what AI ROI looks like in practice.”
c43 · CEO, prepared remarks, earnings call, 2026-06-18
“At the same time, we are executing in new areas, including demand in AI and expanding our TAM.”
c45 · CFO, qa, earnings call, 2026-06-18
“We've seen a three-quarter trend now of more consulting work in those large programs for managed services because our clients are asking us to help them use AI and change the processes to do more change management to really embed new ways of working.”
c46 · CEO, qa, earnings call, 2026-06-18
“one of the things we're clearly seeing, in fact, we have a whole practice that we're starting to grow now is on how to help clients optimize their use of tokens.”
c50 · CEO, qa, earnings call, 2026-06-18
“It's also helping because we have delivered real ROI, and our clients are seeing the spend, but they're struggling with the ROI.”
c52 · CEO, qa, earnings call, 2026-06-18
“Because the budgets haven't been, even with AI, they're spending it differently, but they haven't been increasing.”
c54 · CEO, qa, earnings call, 2026-06-18
“We are really focus ed on expanding our TAM while we're capturing more of the AI spend.”
c55 · CEO, qa, earnings call, 2026-06-18
“In addition to looking at it for acquisitions, we're also going to be building more and more, and we've already started with our ecosystem partners, where we're basically going to have IP together that creates solutions that also drive then our services. We're building them ourselves. For example, our tokenomics platform internally, we're now taking to clients, and that's a platform we built to optimize tokens.”
c57 · CEO, qa, earnings call, 2026-06-18
“We called out that we're starting to see the AI large en terprise programs where they're not just use cases, but really embedding it.”
c59 · CEO, qa, earnings call, 2026-06-18
“When we look at the AI projects themselves, while still small, there's been a steady increase in the average size.”
c60 · CEO, qa, earnings call, 2026-06-18
“That fundamental building of every quarter is continuing. The demand is the same. Getting ready for AI and then deploying AI. We're really optimistic because we believe AI is going to be a tailwind as it scales for us in the industry.”
c61 · CEO, qa, earnings call, 2026-06-18
“Clients are very focused on whatever the kind of work, whether it's AI or not, in tangible results.”
c66 · CEO, qa, earnings call, 2026-06-18

By quarter

  • Q1 2026withdrawn · our inference · narrative-defensive · $880mn to $1.32bn
  • Q2 2026withdrawn · our inference · narrative-defensive · $704mn to $1.58bn

customer cohort

Data projects that follow advanced AI projects

Not sized

The only measure is the same floor on a share by count, at least of advanced AI projects leading to a data project (claim c35), with no number of projects and no dollars. A share by count separates nothing in dollars, so the channel reads directional and gets no ballpark (the methodology's count-only rule, rules sweep of 2026-10-06). The former estimate, acn-2026-cq2-f75, applied that floor to the advanced AI estimate with an assumed relative size and an assumed share added.

described, no sizedisclosure: direction only· motive: offensive· before LLMs: expanded

The one stated rate is repeated in the same words: at least of advanced AI projects lead to a data project (claim c35). It remains a floor on a count with no dollars; the attach rate is quoted and the reading is unsized, as in the prior quarter.

Evidence: 1 quote, 1 figure, 3 confounds, 5 from before coverage

Data modernization work that clients commission after starting an advanced AI project. Management gives an attach rate, the share of advanced AI projects that lead to a data project, and no dollar figure. The dollars sit inside consulting and managed services revenues.

Why this motive

Carried: the attach rate is restated, at least of advanced AI projects continuing to lead to a data project (claim c35), the offensive tell.

Before LLMs: expanded

Data and AI modernization was already sold as part of building the digital core at the anchor, and the CEO said the generative AI sales figure of left data work out. The fiscal 2024 anchor gives no attach rate and no revenue for data work; the practice had data and AI practitioners at the end of fiscal 2024. The attach rate of at least one advanced AI project in two was already being stated on the last call before coverage. The size is the change AI made, not the whole line.

“I'll just remind you, that's not the pull-through, that's not data. We are very pure because we really want to be sharing with all of you, where is GenAI in the market?”
CEO, qa, earnings call, 2023-12-19
“Turning to data and AI, we estimate that less than 10% of companies have mature data and AI capabilities. This is a critical part of building the digital core, and we see this embedded in our larger transformations, in work focused on data and AI modernization, and in the opportunities of generative AI.”
CEO, prepared remarks, earnings call, 2023-12-19
“For lots of companies, it's also more spending, mostly on building that digital core, because many companies don't have the data estates in order. They're not in the cloud. They don't have the data in order to use the GenAI.”
CEO, qa, earnings call, 2023-12-19
“as you know, we were the first in our industry to share our bookings and revenue from advanced AI, which we define as Gen AI, Agentic AI, and Physical AI, and does not include data, classical AI, or RPA.”
CEO, prepared remarks, earnings call, 2025-12-18
“That's why we continue to see at least one out of every two advanced AI projects lead to a data project”
CEO, prepared remarks, earnings call, 2025-12-18

Figures

  • Share of advanced AI projects that lead to a data project, lower bound · 2026-CQ2

What else could explain it

  • other: The attach rate counts projects, not dollars; the size of the data work beside the AI work is the ledger's assumption.
  • relabel: Data modernization was sold before AI was the stated reason (anchor claims acn-anchor-c5, acn-anchor-c6); a data project that would have happened anyway counts as following an AI project. The estimate keeps only an assumed share of the data work as added by AI.
  • other: The base, advanced AI revenue, is itself the ledger's estimate.

Quotes

“A lot of our reinvention work today is helping clients get ready for AI, data remains a critical enabler with at least one out of every two advanced AI projects continuing to lead to a data project.”
c35 · CEO, prepared remarks, earnings call, 2026-06-18

By quarter

  • Q1 2026direction only · described, no size · offensive
  • Q2 2026direction only · described, no size · offensive

customer cohort

Cloud, platform and operating-model work sold as getting ready for AI

Not sized

The filing gives becoming AI-ready as one part of large-scale transformations that also cover cloud, platforms, security and data (claim c3), and management attaches no measure to AI’s part, so the channel is not sized (the methodology rule for AI named beside another cause).

Matched line moved : the whole line, not this channel.

described, no sizedisclosure: described· motive: exploratory· before LLMs: relabelled

The CEO says clients kept investing in the foundations needed to scale AI and that a lot of reinvention work is getting them ready for it (claims c34, c35); the 10-Q again says transformations include becoming AI-ready (claim c3). She also says client budgets are not increasing even with AI, only being spent differently (claim c54), which fits a relabelled channel. Revenues grew in local currency and new bookings changed . AI readiness is again one part of broader transformations with nothing to separate it, so the channel is left unsized, as in Q1; it overlaps the data-projects channel and stays out of the revenue total.

Evidence: 9 quotes, 10 figures, 6 confounds, 4 from before coverage

The part of demand for established work (cloud migration, ERP and SaaS implementation, mainframe and application modernization, operating model change) that management attributes to clients preparing for AI. The filing describes it as large-scale transformations that include becoming AI-ready. No figure separates it from the rest of revenues.

Why this motive

Carried from Q1: management again names AI as the reason clients are modernizing (claims c34, c35) and attaches no price, attach rate or figure to it, and the CEO says budgets are not increasing (claim c54). AI named as the cause with no measure is the exploratory tell; the sources do not contradict each other.

Before LLMs: relabelled

The same work was the stated driver of revenue at the anchor: the fiscal 2024 annual report attributes consulting revenues of to technology, data and AI led transformations, and on the call for the first quarter of fiscal 2024 the CEO said much of clients' added spending went to building the digital core because they lacked the cloud and data foundations to use generative AI. Revenues were in fiscal 2024.

“Turning to data and AI, we estimate that less than 10% of companies have mature data and AI capabilities. This is a critical part of building the digital core, and we see this embedded in our larger transformations, in work focused on data and AI modernization, and in the opportunities of generative AI.”
CEO, prepared remarks, earnings call, 2023-12-19
“For lots of companies, it's also more spending, mostly on building that digital core, because many companies don't have the data estates in order. They're not in the cloud. They don't have the data in order to use the GenAI.”
CEO, qa, earnings call, 2023-12-19
“We are continuing to see significant demand in areas like cloud migration and modernization, modern ERP and data and AI, including GenAI, platforms and security, all of which represent areas of great opportunity”
CEO, prepared remarks, earnings call, 2023-12-19
“Our consulting revenue continues to be driven by helping our clients accelerate their reinvention, in particular technology, data, and AI led digital transformations.”
Filing, mdna, 10-K periodic report, 2024-10-10

Figures

  • Revenues, year-over-year growth in local currency · 2026-CQ2
  • Revenues, year-over-year increase at the local-currency growth rate · 2026-CQ2
  • Consulting revenues, year-over-year growth in local currency · 2026-CQ2
  • Managed services revenues, year-over-year growth in local currency · 2026-CQ2
  • Currency translation effect on U.S. dollar revenue growth · 2026-CQ2
  • New bookings · 2026-CQ2
  • New bookings, year-over-year change in local currency · 2026-CQ2
  • Managed services new bookings, year-over-year change in local currency · 2026-CQ2
  • Revenue shortfall against expectations attributed to the conflict in the Middle East, approximate · 2026-CQ2
  • Expected contribution of acquisitions to fiscal 2026 revenue growth, approximate · FY2026

Reported line it is matched to

Revenues rose to , with currency translation adding ; in local currency the growth is , about . The filing attributes consulting growth to reinvention work across cloud, platforms, security, AI and data and says the discretionary environment is unchanged (claims c4, c3); it does not say how much of the growth is AI readiness.

2026-CQ2: 2025-CQ2: 2026-CQ2: 2026-CQ2: 2026-CQ2:

What else could explain it

  • relabel: Cloud, platform and modernization work was the stated driver of revenue before AI was the stated reason; the CEO says the demand is the same, getting ready for AI and then deploying it (claim c61).
  • fx: Currency translation added to reported revenue growth of .
  • line composition: Data modernization done to get ready for AI is read in the data-projects channel, and the CEO's statement here is tagged to both (claim c35); this reading overlaps that channel and is left out of flow totals so that work is not counted twice.
  • acquisition: Management expects acquisitions to add about to the fiscal year's growth; that part of the increase is bought.
  • one time item: Management attributes a shortfall of about in consulting revenues to the conflict in the Middle East, and says a couple of large managed services deals moved into the next fiscal year.
  • other: Cloud, platform, security and data work are named with becoming AI-ready as parts of the same large-scale transformations (claim c3).

Quotes

“While the discretionary environment is unchanged, clients continue to prioritize large-scale transformations, which include becoming AI-ready.”
c3 · Filing, mdna, 10-Q periodic report, 2026-06-18
“driven by helping our clients accelerate their reinvention, leveraging cloud, enterprise platforms, security, AI and data, including advanced AI”
c4 · Filing, mdna, 10-Q periodic report, 2026-06-18
“We continue to assist clients with reinvented operations, application development and maintenance, and infrastructure management including cloud and security.”
c6 · Filing, mdna, 10-Q periodic report, 2026-06-18
“We see that companies in this segment face many of the same technology, data, AI, cybersecurity, and productivity challenges as large enterprises, but they often need solutions that are faster to deploy, more repeatable, and right-sized for their scale.”
c30 · CEO, prepared remarks, earnings call, 2026-06-18
“Clients continued to invest in the foundations needed to scale AI. This includes strengthening their digital core through cloud, data security, and operating model transformation.”
c34 · CEO, prepared remarks, earnings call, 2026-06-18
“A lot of our reinvention work today is helping clients get ready for AI, data remains a critical enabler with at least one out of every two advanced AI projects continuing to lead to a data project.”
c35 · CEO, prepared remarks, earnings call, 2026-06-18
“We've seen a three-quarter trend now of more consulting work in those large programs for managed services because our clients are asking us to help them use AI and change the processes to do more change management to really embed new ways of working.”
c46 · CEO, qa, earnings call, 2026-06-18
“Because the budgets haven't been, even with AI, they're spending it differently, but they haven't been increasing.”
c54 · CEO, qa, earnings call, 2026-06-18
“That fundamental building of every quarter is continuing. The demand is the same. Getting ready for AI and then deploying AI. We're really optimistic because we believe AI is going to be a tailwind as it scales for us in the industry.”
c61 · CEO, qa, earnings call, 2026-06-18

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026described · described, no size · exploratory

distribution

Work sold with AI and data ecosystem partners

0.69% to 4.1% of the quarter’s revenue; overlaps another channel, not added into totals

our inferencedisclosure: direction only· motive: offensive· before LLMs: expanded

The emerging AI and data partners are named for the first time: Anthropic, Databricks, Gemini, Mistral AI, Nvidia, OpenAI, Palantir and Snowflake (claim c25). The measure is unchanged, bookings on track to more than double in the fiscal year, with no base. The size here is the ledger's own, , an assumed share of the advanced AI estimate; it sits inside that figure and is left out of totals.

Evidence: 2 quotes, 2 confounds, 2 from before coverage

Bookings and revenue from work sold together with technology partners, which management reports as two measures: revenue from its largest ecosystem partners growing faster than the company, and bookings with a named group of emerging AI and data partners (model developers, data platforms and a chip maker) on track to more than double in the fiscal year. The client pays; the partner is the route to the sale. Most of the emerging-partner work is advanced AI or data work, so it sits inside those channels.

Why this motive

Carried: the same two measures are restated, revenue from the largest partners outpacing the company and bookings with the emerging AI and data partners on track to more than double (claim c25).

Before LLMs: expanded

Selling through partners predates LLMs: the fiscal 2024 annual report says a very significant portion of revenue rests on technology from a few major ecosystem partners and lists Databricks, NVIDIA and Snowflake among them, with no model developer on the list. No partner revenue or bookings figure is in the anchor, so the level before AI cannot be traced and no baseline figure is registered. The size is the whole of an activity that existed before.

“A very significant portion of our revenue and services and solutions are based on technology or software provided by a few major ecosystem partners.”
Filing, risk factors, 10-K periodic report, 2024-10-10
“Our strong ecosystem relationships provide a significant competitive advantage, and we are a key partner of a broad range of technology providers, including Adobe, Alibaba, Amazon Web Services, Blue Yonder, Cisco, Databricks, Dell, Google, HPE, IBM RedHat, Microsoft, NVIDIA, Oracle, Palo Alto Networks, Pegasystems, Salesforce, SAP, ServiceNow, Snowflake, VMware, Workday and many others.”
Filing, business, 10-K periodic report, 2024-10-10

What else could explain it

  • line composition: The emerging group, now named, mixes model developers (Anthropic, Gemini, Mistral AI, OpenAI) with data platforms and a chip maker; the measure is not an AI figure.
  • other: Bookings are not revenue, and a growth multiple without a base gives no level.

Quotes

“We have announced a number of expansions of partnerships with our top 10 ecosystem partners in AI and data, and our revenue growth from these partners continues to outpace our overall growth. We are also on track to more than double our bookings from our key emerging AI and data partners compared with FY 2025, including Anthropic, Databricks, Gemini, Mistral AI, Nvidia, OpenAI, Palantir, and Snowflake.”
c25 · CEO, prepared remarks, earnings call, 2026-06-18
“Together with a leading large language model provider and hyperscaler, we built an AI engine that validates and enriches leads, generates personalized campaign content, and automates brand and legal validations.”
c42 · CEO, prepared remarks, earnings call, 2026-06-18

By quarter

  • Q1 2026direction only · our inference · offensive · $119mn to $713mn
  • Q2 2026direction only · our inference · offensive · $128mn to $770mn

product revenue

Cybersecurity, data center, capital projects and education work sold as AI enablers

Not sized

Management names AI beside geopolitical risk as the driver of demand for security work (claims c28, c29) and gives no measure of AI’s part; the stated cybersecurity figure is the whole business, so the channel is not sized (the methodology rule for AI named beside another cause).

described, no sizedisclosure: described· motive: exploratory· before LLMs: relabelled

For the first time management gives a size for one of the enabler businesses: cybersecurity services of in fiscal 2025, a compound growth rate of from fiscal 2016 (claim c29), offered as the foundation for acquiring a group of operational technology security companies. The CEO calls cybersecurity one of the largest AI enablers and says demand is strong across capital projects, data centers and education (claims c38, c27). The stated figure is the whole business, not the part AI adds, so the state stays described. AI is named beside geopolitical risk as the cause of that demand and nothing separates AI's part, so the channel is now left unsized; as a relabelled channel it would add nothing to the incremental total in any case.

Evidence: 11 quotes, 5 figures, 3 confounds, 4 from before coverage

Revenue in businesses management groups as AI enablers: cybersecurity, data center engineering, energy and capital projects, and the LearnVantage education business. Management says AI is a catalyst for them. Several were bought, and by the ledger's acquisitions rule the purchase of a company that is not an AI company is an acquisition confound with nothing credited to AI.

Why this motive

Carried from Q1 and confirmed: management calls cybersecurity a key enabler for AI and now sizes that business, but the figure is the whole of cybersecurity services with a decade of growth and acquisitions in it, and AI is named beside geopolitical risk as the driver (claims c29, c28). AI named as a cause with no measure of its own is the exploratory tell; the sources do not contradict each other.

Before LLMs: relabelled

These businesses were in place at the anchor under other descriptions: security was called essential to the digital core and was growing at a very strong double-digit rate in the first quarter of fiscal 2024, and capital projects was a growth area entered by acquisition in 2023. The anchor ties neither to AI and gives no revenue for them; went into acquisitions in fiscal 2024.

“We are continuing to see significant demand in areas like cloud migration and modernization, modern ERP and data and AI, including GenAI, platforms and security, all of which represent areas of great opportunity”
CEO, prepared remarks, earnings call, 2023-12-19
“Security is also essential to a digital core, and we continue to see very strong double-digit growth in our security business this quarter.”
CEO, prepared remarks, earnings call, 2023-12-19
“In North America, we are continuing to build out our new growth area of capital projects, an $88 billion addressable market in North America, which we entered in August with the acquisition of Anser Advisory.”
CEO, prepared remarks, earnings call, 2023-12-19
“Security remains one of our fastest-growing businesses, growing very strong double digits this quarter.”
CEO, prepared remarks, earnings call, 2025-12-18

Figures

  • Cybersecurity services revenue, fiscal 2025 · FY2025
  • Cybersecurity services revenue, compound annual growth from fiscal 2016 to fiscal 2025 · FY2025
  • Annual recurring revenue of the operational technology security companies being acquired · as-of 2026-06-18
  • Capital expected to be deployed in acquisitions in fiscal 2026, as of June 2026, approximate · FY2026
  • Consideration for acquisitions completed in the quarter, net of cash acquired · 2026-CQ2

What else could explain it

  • acquisition: Cybersecurity services were built organically and by acquisition (claim c29), and the companies being bought bring of recurring revenue that is not yet in the income statement; none of it is credited to AI.
  • relabel: Security and capital projects were growth areas before they were called AI enablers; the growth rate cited starts in fiscal 2016.
  • other: Geopolitical risk is named with AI as the cause of demand for operational technology security.

Quotes

“Our agreement to acquire a majority stake in Dragos and all of runZero and NetRise, leaders in OT Security, is the type of move that defines our strategy”
c2 · Filing, press release, 8-K earnings release, 2026-06-18
“On acquisitions, because of the exciting OT cybersecurity acquisitions we announced today, which I'll talk about in a moment, we now expect to deploy approximately $9 billion of capital this year based on the anticipated closing dates of the acquisitions.”
c19 · CEO, prepared remarks, earnings call, 2026-06-18
“Now, let's talk about our big move in OT security to create a platform-led growth business with a non-FTE commercial model.”
c26 · CEO, prepared remarks, earnings call, 2026-06-18
“Cyber is a key enabler for AI. We cannot have an AI revolution without critical infrastructure, and you cannot have those without OT security, which is where today the world is most vulnerable.”
c27 · CEO, prepared remarks, earnings call, 2026-06-18
“AI and geopolitical risk are accelerating the need for cybersecurity adoption for the operational technology that underpins critical infrastructure and industrial operations”
c28 · CEO, prepared remarks, earnings call, 2026-06-18
“We have grown our services organically and inorganically over the last decade from roughly $700 million in FY 2016 to $10 billion in fiscal 2025, a 35% CAGR over the period, four times that of Accenture's over the same period.”
c29 · CEO, prepared remarks, earnings call, 2026-06-18
“Another area of strong demand are the AI enablers we've been investing in from capital projects to data centers to LearnVantage to cybersecurity, one of our largest AI enablers. As I mentioned earlier, OT security is one of the hottest areas driven by AI cyber threats and geopolitical risk.”
c38 · CEO, prepared remarks, earnings call, 2026-06-18
“You can't have an AI revolution unless you have critical infrastructure and unless you secure when you start moving into physical AI, and you can't have that without OT security.”
c44 · CEO, qa, earnings call, 2026-06-18
“As well as AI enablers, right? Cybersecurity, capital projects, data centers.”
c58 · CEO, qa, earnings call, 2026-06-18
“We are skewing, however, toward where we see the demand, which is in areas where we uniquely have the domain expertise the clients needed. It's usually triggered by AI. That's what we're seeing. We're skewing our acquisitions to where we see the biggest growth opportunities right now, and where we see the biggest growth opportunities on the product side are in these areas that are being triggered by AI.”
c62 · CEO, qa, earnings call, 2026-06-18
“If you think about what we did with DLB Associates, that acquisition, they're growing very high double digits, right?”
c63 · CEO, qa, earnings call, 2026-06-18

By quarter

  • Q1 2026described · our inference · exploratory · $0 to $318mn
  • Q2 2026described · described, no size · exploratory

Cost imposed, or revenue lost, by others’ AI3 channels · $0 to $561mn sized · $0 to $373mn incremental

other · cheap to verify

Billable work shortened or replaced by AI

0% to 1.5% of the quarter’s revenue

Incremental total: counts in full.

our inferencedisclosure: described· motive: imposed· before LLMs: expanded

No one asks about compressed timelines this quarter and management does not raise them. Its position on the net effect is restated in the future tense: AI will be a tailwind for the company and its industry as it scales (claims c21, c61), where a quarter earlier it was a tailwind today. Consulting revenues grew in local currency while consulting bookings grew , a gap an analyst asked about and the CFO put down to the Middle East (claim c47); managed services bookings changed , and the 10-Q says the company continues to assist clients with managed services where a quarter earlier it said demand for them was growing (claim c6). None of this is attributed to AI. The ledger's range, , is the prior quarter's, from nothing to a small share of consulting revenues.

Evidence: 4 quotes, 7 figures, 2 confounds, 3 from before coverage

Revenue the company would have billed at the prior scope of work had AI not shortened it: project timelines compressed by AI tools, technical work such as code conversion done in less time, and work clients take in-house or leave to their software vendors' AI. Management says faster technical work leads to more work and is a net benefit; no figure for work lost is given. It would show in consulting revenues and in revenue against headcount.

Why this motive

Carried: a toll channel, imposed by construction.

Before LLMs: expanded

Automation replacing services predates LLMs: the fiscal 2024 annual report says technological developments had reduced and replaced some historical services and expects AI-enabled automation to replace tasks its people perform, and the CFO said automation had already loosened the link between headcount and revenue. Consulting revenues were in fiscal 2024, a change of in local currency that the filing attributed to slower client spending on smaller contracts and not to automation. The size is the change AI made, not the whole line.

“As it relates to the revenue per head and the non-linearity, I mean, we do have automation. We do have value-based projects. While there still is, obviously, a connection to the amount of people that we have, we have been able to break that.”
CFO, qa, earnings call, 2023-12-19
“Some of these technological developments have reduced and replaced, in whole or in part, some of our historical services and solutions and will continue to do so in the future.”
Filing, risk factors, 10-K periodic report, 2024-10-10
“As these technologies evolve, some services and tasks currently performed by our people will be replaced by automation, including AI-enabled solutions, which will lead to reduced demand for our services and/or adversely affect the utilization rate of our professionals, if demand for those services is not replaced by demand for new services.”
Filing, risk factors, 10-K periodic report, 2024-10-10

Figures

  • Consulting revenues · 2026-CQ2
  • Consulting revenues, prior-year quarter · 2025-CQ2
  • Consulting revenues, year-over-year growth in local currency · 2026-CQ2
  • Consulting new bookings, year-over-year growth in local currency · 2026-CQ2
  • Managed services revenues, year-over-year growth in local currency · 2026-CQ2
  • Managed services new bookings, year-over-year change in local currency · 2026-CQ2
  • Revenue shortfall against expectations attributed to the conflict in the Middle East, approximate · 2026-CQ2

What else could explain it

  • one time item: Management attributes the weak consulting quarter to the conflict in the Middle East, about of revenue, and the fall in managed services bookings to a couple of large deals moving into the next fiscal year.
  • other: Under fixed-price contracts, over of work, a shorter delivery does not reduce revenue; it shows as margin.

Quotes

“We believe that AI will be a tailwind for us and our industry as it scales, because it is a catalyst for reinvention and is creating new opportunities for growth and efficiency for our clients and for us.”
c21 · CEO, prepared remarks, earnings call, 2026-06-18
“That fundamental building of every quarter is continuing. The demand is the same. Getting ready for AI and then deploying AI. We're really optimistic because we believe AI is going to be a tailwind as it scales for us in the industry.”
c61 · CEO, qa, earnings call, 2026-06-18
“In terms of consulting type of work, we did see it tick down, and it was the result of the indirect and the direct impact of the Middle East.”
c47 · CFO, qa, earnings call, 2026-06-18
“We continue to assist clients with reinvented operations, application development and maintenance, and infrastructure management including cloud and security.”
c6 · Filing, mdna, 10-Q periodic report, 2026-06-18

By quarter

  • Q1 2026described · our inference · imposed · $0 to $266mn
  • Q2 2026described · our inference · imposed · $0 to $280mn

pricing packaging

Price conceded to clients for AI productivity

0% to 1% of the quarter’s revenue

Incremental total: counts at zero.

our inferencedisclosure: described· motive: imposed· before LLMs: relabelled

The 10-Q says pricing was relatively stable, where a quarter earlier it had improved in some areas (claim c7); neither is attributed to AI. The CFO says fixed-price work is over and rising, with margins little different from other contracts (claims c67, c68). The one delivery example on the call has the savings from agentic AI in a managed services program going to the client (claim c37), and the CEO says the shift away from headcount-based pricing starts in new categories because buying patterns for long-established services change slowly (claim c49). The ledger's range, , runs from nothing to a small share of revenues.

Evidence: 9 quotes, 3 figures, 2 confounds, 5 from before coverage

Revenue given up in price because AI makes delivery cheaper: lower rates on legacy services, productivity commitments written into contracts, and savings in managed services that go to the client. Management reports pricing, which it defines as contract profitability, as improved in some areas or stable and attributes nothing to AI. The rising share of fixed-price work decides who keeps a productivity gain.

Why this motive

Carried: a toll channel, imposed by construction.

Before LLMs: relabelled

Price pressure predates AI as a named cause: the fiscal 2024 annual report says the company continued to experience lower pricing across the business in a competitive environment, and lists generative AI only as something that could reduce pricing. On the last call before coverage the CEO described contracts that depend on using more technology over time to provide productivity as the industry's commercial model since RPA. No price movement is attributed to AI at the anchor or since.

“As it relates to the revenue per head and the non-linearity, I mean, we do have automation. We do have value-based projects. While there still is, obviously, a connection to the amount of people that we have, we have been able to break that.”
CFO, qa, earnings call, 2023-12-19
“The business environment is competitive, and we continue to experience lower pricing across the business.”
Filing, mdna, 10-K periodic report, 2024-10-10
“the introduction of new technologies (such as generative AI), services or products by competitors, which could reduce our ability to obtain favorable pricing”
Filing, risk factors, 10-K periodic report, 2024-10-10
“And then with respect to pricing and passing it along to our clients, remember, this has been a business model for the industry, really all the way going back to the introduction of technology and RPA, right, where we're signing contracts that depend on our use of more technology over time to provide productivity. And so that's still the commercial model of the industry right now.”
CEO, qa, earnings call, 2025-12-18
“And we are starting to see more focus on trying to get to outcome-based.”
CEO, qa, earnings call, 2025-12-18

Figures

  • Share of work that is fixed price, as management states it (given a quarter earlier as a share of bookings), lower bound · 2026-CQ2
  • Gross margin · 2026-CQ2
  • Gross margin, prior-year quarter · 2025-CQ2

What else could explain it

  • other: Pricing as the company defines it is contract profitability, so a price concession matched by a delivery saving does not show.
  • mix shift: The move to commercial models not tied to headcount is being made in new categories and by acquisition; the CEO says buying patterns for established services will take a while to change (claim c49).

Quotes

“While the business environment remained competitive, pricing was relatively stable.”
c7 · Filing, mdna, 10-Q periodic report, 2026-06-18
“We strive to adjust pricing as well as drive cost and delivery efficiencies, such as changing the mix of people and utilizing technology, to reduce the impact of compensation increases on our margin and contract profitability.”
c12 · Filing, mdna, 10-Q periodic report, 2026-06-18
“I want to provide a few examples of how we're capturing new areas of demand in the age of AI, what we're doing to expand our total addressable market, and the areas where we are shifting to more non-FTE commercial models over time.”
c20 · CEO, prepared remarks, earnings call, 2026-06-18
“Together, these actions show how we are building a strong foundation for us to win in AI, expanding our addressable market across new growth areas and client segments, and evolving towards more non-FTE revenue over time.”
c31 · CEO, prepared remarks, earnings call, 2026-06-18
“Now, let's talk about our big move in OT security to create a platform-led growth business with a non-FTE commercial model.”
c26 · CEO, prepared remarks, earnings call, 2026-06-18
“We're expanding our partnership to make that possible, consolidating fragmented operations across critical business functions into a unified managed services model with agentic AI embedded throughout and humans in the lead. The result is automation replacing manual effort, faster speed to market, and significant cost savings and productivity gains that Bath & Body Works can reinvest directly into growth.”
c37 · CEO, prepared remarks, earnings call, 2026-06-18
“One of the things that I've said consistently is that in things that our clients have been bu ying and services for a long time, it's going to take a while to change the buying patterns. It's much easier to go into new categories or to provide new kinds of value, and switch to non-FTE models.”
c49 · CEO, qa, earnings call, 2026-06-18
“We continue to see our fixed price work be over 60% and continuing to increase.”
c67 · CFO, qa, earnings call, 2026-06-18
“Margin's not a big difference that I would call out relative to fixed price versus the other commercial constructs, but it is embedded in our 20 basis points of exp ansion for the year.”
c68 · CFO, qa, earnings call, 2026-06-18

By quarter

  • Q1 2026described · our inference · imposed · $0 to $180mn
  • Q2 2026described · our inference · imposed · $0 to $187mn

other

Client budget diverted to AI infrastructure and tokens

0% to 0.5% of the quarter’s revenue

Incremental total: counts in full.

our inferencedisclosure: bounded· motive: imposed· before LLMs: new

An analyst asked whether clients' spending on AI infrastructure and tokens is pressing on their budgets. The CEO bounded the effect: not material to spending on services today, and if anything a reason to use more services (claim c53). In the same answer she said clients see the spending and struggle with the return, and that budgets have not been increasing even with AI (claims c52, c54). The channel opens bounded near nothing by management; the ledger's range, , keeps a small allowance.

Evidence: 4 quotes, 2 figures, 1 confound, 1 from before coverage

Services revenue lost because clients spend on AI infrastructure and model tokens out of technology budgets that are not growing. Management says the effect on services spending is not material and sells a practice to manage token cost; no figure is given.

Why this motive

A toll channel: how much of a flat budget clients move to infrastructure and tokens is their decision. Imposed by construction.

Before LLMs: new

The anchor describes no client spending on tokens or AI infrastructure. It does record the budget constraint: on the call for the first quarter of fiscal 2024 the CEO said clients were reprioritizing within technology budgets that were growing more slowly.

“Right now, we're seeing a lot of reprioritization, right? Because, I mean, obviously the market's growing. Like, we're growing, the market's growing, so spending in technology is increasing. It's not increasing as fast as it was increasing a couple of years ago, right?”
CEO, qa, earnings call, 2023-12-19

Figures

  • Revenues · 2026-CQ2
  • Revenues, year-over-year growth in local currency · 2026-CQ2

What else could explain it

  • other: The same spending is a source of work: the company is growing a practice to help clients optimize token use (claim c50).

Quotes

“one of the things we're clearly seeing, in fact, we have a whole practice that we're starting to grow now is on how to help clients optimize their use of tokens.”
c50 · CEO, qa, earnings call, 2026-06-18
“It's also helping because we have delivered real ROI, and our clients are seeing the spend, but they're struggling with the ROI.”
c52 · CEO, qa, earnings call, 2026-06-18
“At the same time, there's a certain amount of spending that's going to happen, so we're not seeing it be material to impact the spend on services today. If anything, we think it's going to drive more to use services, and that's how we're seeing it develop.”
c53 · CEO, qa, earnings call, 2026-06-18
“Because the budgets haven't been, even with AI, they're spending it differently, but they haven't been increasing.”
c54 · CEO, qa, earnings call, 2026-06-18

Reported lines, year-over-year growth

Revenue +5.6%

Q2 2026. Growing slower than revenue: sales and marketing (+2.8%), total operating expenses (+5.4%). A displaced cost shows up as a line that stays under the dashed revenue line. These are the audited lines, as first reported; nothing here is attributed to AI by the filing. Not drawn: business optimization costs, which one-time items move by more than 60% in a quarter; the values are in the table below.

Cost of servicesSales and marketingGeneral and administrative costsTotal operating expensesRevenue
-5%0%5%10%15%20%Q1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026General and administrative costsCost of servicesRevenueTotal operating expensesSales and marketing
Reported values and filings
LineQ1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026
Revenues$16.66bn$17.73bn$17.60bn$18.74bn$18.04bn$18.72bn
Cost of services$11.68bn$11.90bn$11.99bn$12.55bn$12.58bn$12.58bn
Sales and marketing$1.68bn$1.76bn$1.79bn$1.87bn$1.75bn$1.81bn
General and administrative costs$1.05bn$1.08bn$1.15bn$1.14bn$1.22bn$1.15bn
Business optimization costs$0$0$615mn$308mn$0$0
Total operating expenses$14.41bn$14.75bn$15.55bn$15.87bn$15.55bn$15.54bn