AI Absorption Ledger / ELV

Elevance Health

ELV · Q2 2026 · reported 2026-07-15 · revenue $50.47bn

Assessment

AI disclosure thinned: the Q2 call has fewer AI passages than Q1, the release still has none, and the 10-Q repeats only the Transformation Program’s aims. The member assistant, provider matching, Health OS prior authorization, Carelon risk identification and associate tools are described this quarter, where at all, without AI, so they read as not mentioned. A new channel opened: AI-powered member engagement in the CAHPS survey infrastructure behind Star Ratings, unsized because it is named with other programmes and pays only in later years.

The investment door stays described. The CFO ties targeted investment spending of approximately of EPS to AI adoption, workforce enablement and Carelon scaling, and adjusted operating expense ran above its prior-year share of operating revenue, which the filings credit to investments in the workforce and technology adoption, regulatory matters and premium taxes without naming AI. AI is one of several named causes and nothing separates its part, so the line’s movement is quoted as a ceiling and not read as a shape. Medical cost management is again framed as analytics and AI-enabled tools whose effect is expected next year.

Q2 carries no size. Administrative automation, ballparked by the ledger in Q1, reads described and unsized, since this quarter’s only words on administrative cost name analytics and AI-enabled tools together and place the effect next year. Management sizes nothing, and the one amount it gave in Q1 is not repeated.

Paid for AI1 channel · 1 not sized

engineering

Investment in digital and AI-enabled capabilities

Not sized

AI adoption is named beside workforce enablement and Carelon scaling and nothing separates its part (the methodology rule for AI named beside another cause); the amounts are per share and cover all targeted investments, so they are a ceiling, not a size. The rise in adjusted operating expense is kept in the metrics as the ceiling line’s movement and is not read as a shape, since the filings credit it to several causes and never to AI.

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

The CFO says the 2026 outlook already held approximately of EPS of targeted investment spending that supports AI adoption, workforce enablement and Carelon scaling, and that approximately more will go to one-time investments in the second half. Adjusted operating expense was against , and ran above its prior-year share of operating revenue; the adjusted ratio rose from to . The 10-Q explains the rise by investments in the workforce and technology adoption, regulatory matters and premium taxes, never by AI. The Q1 amount, more than for digital and AI-enabled capabilities together, is not repeated.

Evidence: 8 quotes, 11 figures, 3 confounds, 3 from before coverage

What the company spends building and rolling out AI: the stated investment in digital and AI-enabled capabilities, the targeted investments the CFO says support AI adoption and workforce enablement beside Carelon scaling, and the one-time investments of the second half of 2026 in medical cost management, member engagement and provider connectivity. Management always names AI together with digital capabilities, technology adoption, workforce and Carelon, and never separates the AI part: the Q1 amount is a floor on digital and AI together, and the per-share investment amounts cover all targeted investments, so none of them sizes the AI part. Staff, vendor and model bills are not split, and part of the spend may be capitalized. The vendor bill is not registered apart because no vendor or model is named.

Why this motive

The CFO places AI adoption among the areas the targeted investments support (claim c8) and the CEO says the medical cost investments pay off next year (claim c3): investing for later, with no displaced line shown shrinking.

Before LLMs: expanded

At the anchor the FY2024 10-K already described investments in AI administrative tools and solutions (claim elv-anchor-c1) and an ongoing programme of digital technology and internal operating improvements (claim elv-anchor-c9), with no amount, against operating expense of for the year. The first call of 2024 carries no AI passage. The earlier size of AI spending is not traced, so the level has no baseline. The size is the whole of an activity that existed before.

“As part of our operations, we are making investments in certain AI administrative tools and solutions to enhance our operations and positively impact the experience of our members, and we continue to explore further innovation using AI.”
Filing, risk factors, 10-K periodic report, 2025-02-20
“In addition, connectivity amongst technologies is becoming increasingly important, with recent trends bringing greater consumer engagement in healthcare; therefore, the pace at which our customers will need enhanced technologies with sophisticated applications for mobile interfaces, including tools and products that leverage AI to improve the customer experience, will quicken.”
Filing, risk factors, 10-K periodic report, 2025-02-20
“We continue to enhance interactions with customers, providers, brokers, agents, employees and other stakeholders through digital technology and improvements to internal operations.”
Filing, business, 10-K periodic report, 2025-02-20

Figures

  • Targeted investment spending in the 2026 outlook, per diluted share (approximately; spans AI adoption, workforce enablement and Carelon scaling) · FY2026
  • One-time accelerated investments planned for the second half of 2026, per diluted share (approximately) · 2026-07-01..2026-12-31
  • Adjusted operating expense (release reconciliation) · 2026-CQ2
  • Adjusted operating expense, prior-year quarter · 2025-CQ2
  • Total operating revenue · 2026-CQ2
  • Total operating revenue, prior-year quarter · 2025-CQ2
  • Adjusted operating expense at the prior-year share of operating revenue · 2026-CQ2
  • Adjusted operating expense above the prior-year share of operating revenue · 2026-CQ2
  • Adjusted operating expense ratio · 2026-CQ2
  • Adjusted operating expense ratio, prior-year quarter · 2025-CQ2
  • Investment in digital and AI-enabled capabilities together, no period stated (more than; the AI part is not separated) · as-of 2026-04-22

What else could explain it

  • line composition: The targeted investments support AI adoption beside workforce enablement and Carelon scaling (claim c8), and the 10-Q names workforce and technology adoption, regulatory matters and premium taxes (claim c13); no AI part is separated.
  • one time item: The second-half investments funded by investment income are one-time (claims c9 and c10) and start after the quarter.
  • other: The investment amounts are given per diluted share, not in dollars.

Quotes

“Medical cost management is really about investments in analytics and AI-enabled tools that, again, allow us to identify those pressures earlier, but more importantly, allow us to implement clinical oversight, payment integrity, changes in our network, those kind of interventions. Those take time to mature, but we need the data, we need the infrastructure in place this year, and we expect to see that next year.”
c3 · CEO, qa, earnings call, 2026-07-15
“On the investment spend here, the 2026 outlook from the first quarter already included approximately $0.75 of EPS tied to those targeted investment spending that we spoke about at the beginning of the year.”
c7 · CFO, qa, earnings call, 2026-07-15
“Those investments should be viewed as part of the ongoing run rate of the business, and that will support areas like AI adoption and workforce enablement, Carelon scaling, etc.”
c8 · CFO, qa, earnings call, 2026-07-15
“In addition, as we've spoken about this morning, we now expect to deploy approximately $0.80 of net below the line favorability from the second quarter into one-time accelerated investments in the second half.”
c9 · CFO, qa, earnings call, 2026-07-15
“We are looking at these as long-term, durable capabilities. Some are technology, but honestly, they're all driven based on improving the capabilities that we have inside of the business.”
c10 · CEO, qa, earnings call, 2026-07-15
“Operating expense increased primarily due to increases in targeted investments to support and strengthen our workforce and accelerate technology adoption, an increase in expenses to address regulatory matters and an increase in premium tax expenses and assessments.”
c13 · Filing, mdna, 10-Q periodic report, 2026-07-15
“The adjusted operating expense ratio was 11.0 percent, an increase of 100 basis points, driven primarily by targeted investments in our workforce and in capabilities that support our long-term operating model.”
c15 · Filing, press release, 8-K earnings release, 2026-07-15
“Beyond our EPS outlook, the principal operating elements of our full-year framework remain unchanged, though we now expect our adjusted operating expense ratio to be in the upper half of our full-year guidance range.”
c17 · CFO, prepared remarks, earnings call, 2026-07-15

By quarter

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

Cost displaced by AI7 channels · 7 not sized

customer support · cheap to verify

Member service cost displaced by the AI-enabled virtual assistant

Not sized

No source this quarter names AI in member service; nothing permits a size.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: expanded

The CEO says Sydney Health touches about the same number of members (claim c5) but does not call it AI. The Q1 count was for the AI-enabled virtual assistant; the Q2 count is for the whole Sydney Health app. The channel is read as not mentioned and no size is carried.

Evidence: 0 quotes, 1 confound, 3 from before coverage

The cost of answering members (call centre staff, case handling) that is lower than it would have been because the AI-enabled virtual assistant answers questions for commercial members. The CEO gives a count of members using it and says consumer effort scores improved, with no cost, contact or handling-time measure. The assistant appears to sit in the Sydney Health app, which the anchor already shows; the Q2 2026 call repeats the member count under the Sydney Health name without calling it AI. Comparable with the member service doors at UnitedHealth and CVS Health.

Why this motive

Carried from Q1, where it rested on a member count with no measured result (claim c8); silence is not evidence about motive.

Before LLMs: expanded

At the anchor Sydney Health was already the digital engagement platform for members (claim elv-anchor-c3), the 10-K expected customers to need tools that use AI to improve the customer experience (claim elv-anchor-c2) and described investment in AI administrative tools for the member experience (claim elv-anchor-c1); member service cost sat inside operating expense of for fiscal 2024, not split out. The first call of 2024 carries no AI passage. The size is the change AI made, not the whole line.

“As part of our operations, we are making investments in certain AI administrative tools and solutions to enhance our operations and positively impact the experience of our members, and we continue to explore further innovation using AI.”
Filing, risk factors, 10-K periodic report, 2025-02-20
“In addition, connectivity amongst technologies is becoming increasingly important, with recent trends bringing greater consumer engagement in healthcare; therefore, the pace at which our customers will need enhanced technologies with sophisticated applications for mobile interfaces, including tools and products that leverage AI to improve the customer experience, will quicken.”
Filing, risk factors, 10-K periodic report, 2025-02-20
“Our digital engagement platform, Sydney Health, is designed to give our members access to personalized health and wellness resources, medical, pharmacy, dental and vision benefits details, and virtual care services, all in one place.”
Filing, business, 10-K periodic report, 2025-02-20

What else could explain it

  • line composition: Member service cost sits inside operating expense and is not split out.

By quarter

  • Q1 2026direction only · described, no size · exploratory
  • Q2 2026not mentioned · inscrutable · exploratory

search discovery · expensive to verify

Medical cost lowered by matching members to high-performing providers through Sydney

Not sized

No source this quarter names AI in provider matching.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: relabelled

Sydney Health is described this quarter as data, digital tools and care teams helping members navigate (claim c5), without AI.

Evidence: 0 quotes, 1 confound, 2 from before coverage

Benefit expense that is lower because members who use Sydney, the personalized provider-matching tool the CEO describes as a deployment of this technology in an answer about AI, choose high-performing care providers. The only measure is a share of members who have connected, a count share that separates no dollars. The care providers paid less are the counterparty, as at CVS Health. Medical choices are expensive to verify.

Why this motive

Carried from Q1, where it rested on a count share named in an answer about AI (claim c9); silence is not evidence about motive.

Before LLMs: relabelled

At the anchor Sydney Health already gave members personalized health resources and benefits details (claim elv-anchor-c3) and the 10-K expected tools that use AI to improve the customer experience (claim elv-anchor-c2); benefit expense was for fiscal 2024. No movement in cost or volume is attributed to AI matching in coverage, so the activity is read as renamed.

“In addition, connectivity amongst technologies is becoming increasingly important, with recent trends bringing greater consumer engagement in healthcare; therefore, the pace at which our customers will need enhanced technologies with sophisticated applications for mobile interfaces, including tools and products that leverage AI to improve the customer experience, will quicken.”
Filing, risk factors, 10-K periodic report, 2025-02-20
“Our digital engagement platform, Sydney Health, is designed to give our members access to personalized health and wellness resources, medical, pharmacy, dental and vision benefits details, and virtual care services, all in one place.”
Filing, business, 10-K periodic report, 2025-02-20

What else could explain it

  • other: Benefit expense moves with membership, mix, rates, utilization and reserve development.

By quarter

  • Q1 2026direction only · described, no size · exploratory
  • Q2 2026not mentioned · inscrutable · exploratory

back office · expensive to verify

Prior authorization handling cost displaced by Health OS and AI

Not sized

No source this quarter names AI in prior authorization.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: relabelled

The CEO describes Health OS improving prior authorization toward a real-time share (claim c6) without naming AI, so the channel is read as not mentioned.

Evidence: 0 quotes, 1 confound, 4 from before coverage

The cost of handling prior authorization requests, follow-ups and denials with care providers that is lower because Health OS, the data-sharing platform, and AI reduce missing-information denials. The CEO credits the reduction to the technology and AI together, and the Q2 2026 call describes Health OS prior authorization without naming AI. Medical necessity decisions are regulated and costly to get wrong, so the work is expensive to verify; AI in utilization management is a regulatory subject as well as a cost door, and only the company’s words are quoted.

Why this motive

Carried from Q1, where it rested on a denial reduction credited to Health OS and AI together (claim c13); silence is not evidence about motive.

Before LLMs: relabelled

At the anchor clinical data exchange through the HealthOS platform already existed in Carelon Insights (claim elv-anchor-c4), and utilization management was a named medical management programme (claims elv-anchor-c5 and elv-anchor-c7); handling cost sat inside operating expense of for fiscal 2024, not split out. The first call of 2024 carries no AI passage. The coverage measure is credited to the platform and AI together, so the activity is read as renamed.

“Carelon Insights capabilities include payment integrity, subrogation, clinical data exchange through our HealthOS platform, research and data services, reporting and clinical analytics, and information technology services and global business process support.”
Filing, business, 10-K periodic report, 2025-02-20
“We work to mitigate these trends through various medical management programs such as care and condition management, program integrity and specialty pharmacy management and utilization management, as well as benefit design changes.”
Filing, mdna, 10-K periodic report, 2025-02-20
“Our approach to cost management relies on capabilities including provider enablement, value-based networks, member engagement, and utilization management.”
Filing, business, 10-K periodic report, 2025-02-20
“This data can allow us to more efficiently capture information regarding the risk of our membership and the overall adherence to evidence-based medicine, as well as information to more efficiently perform utilization management administration.”
Filing, business, 10-K periodic report, 2025-02-20

What else could explain it

  • other: Health OS is a data-sharing platform that predates LLMs at the company (anchor claim elv-anchor-c4).

By quarter

  • Q1 2026direction only · described, no size · exploratory
  • Q2 2026not mentioned · inscrutable · exploratory

cost of-revenue · expensive to verify

Medical cost avoided by AI in trend detection, payment integrity and clinical review

Not sized

AI-enabled tools are named together with analytics and the effect is placed next year (claim c3): a joint cause nothing separates, and money that has not started (the methodology rule for unsized channels). No ballpark is built; benefit expense of is the ceiling line.

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

The CEO says medical cost management rests on investments in analytics and AI-enabled tools that surface pressures earlier and support clinical oversight, payment integrity and network changes, and that the effect should show next year. Benefit expense was against .

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

Benefit expense the company does not pay because AI and analytics find outliers, emerging cost drivers and improper payments earlier and speed clinical, network and payment integrity actions. In Q1 2026 the CEO says the cost effect comes over time and that for now the gain is information arriving faster; in Q2 2026 she names analytics and AI-enabled tools together and places the effect on medical cost next year. The providers whose claims are not paid are the counterparty, which the schema has no member for; the channel is tagged internal as at UnitedHealth. Errors fall on patients and providers, so the work is expensive to verify.

Why this motive

Management again places the effect next year (claim c3): investing for later, with no line moving.

Before LLMs: relabelled

At the anchor payment integrity and clinical analytics already existed in Carelon Insights (claim elv-anchor-c4) and program integrity was a named medical management programme (claim elv-anchor-c5); benefit expense was for fiscal 2024. The first call of 2024 carries no AI passage. No measure of medical cost moved by AI is given in coverage, so the activity is read as renamed.

“Carelon Insights capabilities include payment integrity, subrogation, clinical data exchange through our HealthOS platform, research and data services, reporting and clinical analytics, and information technology services and global business process support.”
Filing, business, 10-K periodic report, 2025-02-20
“We work to mitigate these trends through various medical management programs such as care and condition management, program integrity and specialty pharmacy management and utilization management, as well as benefit design changes.”
Filing, mdna, 10-K periodic report, 2025-02-20

Figures

  • Benefit expense · 2026-CQ2
  • Benefit expense, prior-year quarter · 2025-CQ2

What else could explain it

  • other: Analytics, clinical oversight, payment integrity and network changes are named together with AI-enabled tools (claim c3); these programmes predate LLMs (anchor claims elv-anchor-c4 and elv-anchor-c5).
  • other: Benefit expense moves with membership, mix, rates, utilization and reserve development; the CFO credits the quarter to Medicare Advantage portfolio actions and ACA seasonality.

Quotes

“First, we are managing medical cost trend with greater speed and precision. In a dynamic environment, we're improving our ability to detect cost pressures earlier and respond quickly with targeted action plans across our clinical, network, payment integrity, and operating teams. In many cases, we've compressed months of work into days.”
c2 · CEO, prepared remarks, earnings call, 2026-07-15
“Medical cost management is really about investments in analytics and AI-enabled tools that, again, allow us to identify those pressures earlier, but more importantly, allow us to implement clinical oversight, payment integrity, changes in our network, those kind of interventions. Those take time to mature, but we need the data, we need the infrastructure in place this year, and we expect to see that next year.”
c3 · CEO, qa, earnings call, 2026-07-15
“It's not just going to be in our cost structure, our expense cost structure. We expect to see it in our medical cost structure, that's what these investments in medical management are about.”
c4 · CEO, qa, earnings call, 2026-07-15

By quarter

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

other · expensive to verify

Medical cost in Carelon risk-based programmes lowered by AI identification of high-risk members

Not sized

No source this quarter names AI in Carelon’s risk identification.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: relabelled

Carelon programmes are described with savings shares this quarter, none attributed to AI, so the channel is read as not mentioned.

Evidence: 0 quotes, 1 confound, 3 from before coverage

Medical cost in Carelon’s risk-based solutions that is lower because AI and advanced analytics identify high-risk members earlier for coordinated care, which the CEO credits with higher medication adherence, fewer emergency room visits and fewer readmissions. AI is named together with advanced analytics and the care programmes themselves, and no part is separated. Split from the health plan’s medical cost door because it is a different business and function.

Why this motive

Carried from Q1, where it rested on outcomes credited to AI and advanced analytics together (claim c2); silence is not evidence about motive.

Before LLMs: relabelled

At the anchor Carelon Services already managed cost through member engagement, value-based networks and utilization management (claim elv-anchor-c6), and provider data was used to capture member risk (claim elv-anchor-c7); benefit expense was for fiscal 2024. The first call of 2024 carries no AI passage.

“We work to mitigate these trends through various medical management programs such as care and condition management, program integrity and specialty pharmacy management and utilization management, as well as benefit design changes.”
Filing, mdna, 10-K periodic report, 2025-02-20
“Our approach to cost management relies on capabilities including provider enablement, value-based networks, member engagement, and utilization management.”
Filing, business, 10-K periodic report, 2025-02-20
“This data can allow us to more efficiently capture information regarding the risk of our membership and the overall adherence to evidence-based medicine, as well as information to more efficiently perform utilization management administration.”
Filing, business, 10-K periodic report, 2025-02-20

What else could explain it

  • other: Carelon results are credited to CareBridge, behavioral health and risk-based programme scaling, not to AI.

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026not mentioned · inscrutable · exploratory

back office · expensive to verify

Administrative expense displaced by AI automation in operational workflows

Not sized

The only Q2 words on administrative cost name analytics and AI-enabled tools together and place the effect next year (claims c3 and c4): a joint cause nothing separates, and money that has not started (the methodology rule for unsized channels). No ballpark is built; net operating expense of is the ceiling line.

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

The CEO expects the investments in analytics and AI-enabled tools to show in the expense cost structure as well as medical cost, next year, with no figure, and the 10-Q repeats the Transformation Program’s aims. Operating expense net of transaction costs and litigation was , above its prior-year share of revenue. The Q1 ballpark is not carried: this quarter’s words are joint-cause and forward-looking, so the channel carries no size.

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

Operating expense that is lower than it would have been because AI embedded in operational and administrative workflows automates work (claims operations, enrollment, service back office), which the CEO describes as reducing administrative expense through automation. Member service, prior authorization and associates’ personal productivity tools are separate channels. The Operating Model Transformation Program, which names enhancing the use of AI among its aims beside fewer organizational layers, workforce reductions and platform modernization, is a confound and is never credited to AI. Much of this work is claims and enrollment handling, regulated and costly to get wrong, so it is read as expensive to verify.

Why this motive

The CEO expects investments in analytics and AI-enabled tools to show in the expense cost structure as well as medical cost, next year (claims c3 and c4); no cost line is shown to fall.

Before LLMs: expanded

At the anchor the 10-K described investment in AI administrative tools (claim elv-anchor-c1) and advanced capabilities that improve service while optimizing administrative costs (claim elv-anchor-c10); operating expense was for fiscal 2024. The first call of 2024 carries no AI passage. The size is the change AI made, not the whole line.

“As part of our operations, we are making investments in certain AI administrative tools and solutions to enhance our operations and positively impact the experience of our members, and we continue to explore further innovation using AI.”
Filing, risk factors, 10-K periodic report, 2025-02-20
“We continue to enhance interactions with customers, providers, brokers, agents, employees and other stakeholders through digital technology and improvements to internal operations.”
Filing, business, 10-K periodic report, 2025-02-20
“Our approach includes not only the sales and distribution of health benefits products through digital technology, but also implementing advanced capabilities that improve services benefiting customers, agents, brokers and providers while optimizing administrative costs.”
Filing, business, 10-K periodic report, 2025-02-20

Figures

  • Operating expense net of the special items the release names · 2026-CQ2
  • Adjusted operating expense (release reconciliation) · 2026-CQ2
  • Adjusted operating expense, prior-year quarter · 2025-CQ2
  • Adjusted operating expense ratio · 2026-CQ2
  • Adjusted operating expense ratio, prior-year quarter · 2025-CQ2

What else could explain it

  • transformation program: The Operating Model Transformation Program names enhancing the use of AI among its aims beside fewer organizational layers, workforce reductions and platform modernization (claim c14; claim c19); it took no charges this quarter and nothing it saves is credited to AI.
  • line composition: Operating expense carries selling expense, premium taxes, investment spending and all administration.
  • other: Targeted investments in the workforce and technology adoption raised the line (claim c13), so any saving is netted against spend.

Quotes

“Medical cost management is really about investments in analytics and AI-enabled tools that, again, allow us to identify those pressures earlier, but more importantly, allow us to implement clinical oversight, payment integrity, changes in our network, those kind of interventions. Those take time to mature, but we need the data, we need the infrastructure in place this year, and we expect to see that next year.”
c3 · CEO, qa, earnings call, 2026-07-15
“It's not just going to be in our cost structure, our expense cost structure. We expect to see it in our medical cost structure, that's what these investments in medical management are about.”
c4 · CEO, qa, earnings call, 2026-07-15
“Operating expense increased primarily due to increases in targeted investments to support and strengthen our workforce and accelerate technology adoption, an increase in expenses to address regulatory matters and an increase in premium tax expenses and assessments.”
c13 · Filing, mdna, 10-Q periodic report, 2026-07-15
“In the first quarter of 2026, based on a strategic review of our operations, assets and investments, management implemented the 2026 - 2027 Operating Model Transformation Program (the “Transformation Program”) to streamline decision-making, simplify organizational structures, and enhance the use of advanced technologies, including artificial intelligence, across the enterprise.”
c14 · Filing, notes, 10-Q periodic report, 2026-07-15

By quarter

  • Q1 2026described · our inference · exploratory · $0 to $52mn
  • Q2 2026described · described, no size · exploratory

other · cheap to verify

Associate time saved by AI productivity tools

Not sized

No source this quarter names AI productivity tools for associates; the CFO names AI adoption only as an area of investment.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: expanded

The CFO names AI adoption and workforce enablement as areas the targeted investments support, which is read on the investment channel; no use, access count or saving is given for associates.

Evidence: 0 quotes, 1 confound, 2 from before coverage

Staff time freed by AI in the productivity tools associates use, which the CEO says a count of associates already have access to. No time saving, headcount or cost is measured, and freed time may show as capacity rather than lower cost. The tools’ licence cost sits inside the investment channel, since no vendor is named. Overlaps in cost pool with the administrative automation channel; the two are kept apart because one is personal tools and the other workflow automation.

Why this motive

Carried from Q1, where it rested on a count of associates with access (claim c14); silence is not evidence about motive.

Before LLMs: expanded

At the anchor the company had approximately employees and described investment in AI administrative tools (claim elv-anchor-c1) and capabilities that optimize administrative costs (claim elv-anchor-c10); no AI productivity tool was named. The first call of 2024 carries no AI passage. The size is the change AI made, not the whole line.

“As part of our operations, we are making investments in certain AI administrative tools and solutions to enhance our operations and positively impact the experience of our members, and we continue to explore further innovation using AI.”
Filing, risk factors, 10-K periodic report, 2025-02-20
“Our approach includes not only the sales and distribution of health benefits products through digital technology, but also implementing advanced capabilities that improve services benefiting customers, agents, brokers and providers while optimizing administrative costs.”
Filing, business, 10-K periodic report, 2025-02-20

What else could explain it

  • other: Freed time may show as capacity rather than lower cost.

By quarter

  • Q1 2026direction only · described, no size · exploratory
  • Q2 2026not mentioned · inscrutable · exploratory

Revenue arriving through AI1 channel · 1 not sized

customer cohort · cheap to verify

Medicare Advantage Star Ratings bonus revenue from AI-powered member engagement

Not sized

AI is named beside other engagement programmes and nothing separates its part; ratings set now pay in later years and management declines to predict payment year 2028 (claim c12), so the money has not started (the methodology rule for unsized channels). No ballpark is built.

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

Opened this quarter. The head of the government business says the CAHPS survey infrastructure was enhanced with AI-powered personalized member engagement beside outreach and rewards programmes. The 10-Q gives approximately of Medicare Advantage members in plans rated four stars or higher for 2027 payments, against approximately for 2026 payments; neither is attributed to AI.

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

Quality bonus payments from CMS that may rise if member experience survey (CAHPS) scores and Star Ratings improve, which the head of the government business ties to AI-powered personalized member engagement beside omni-channel outreach and rewards programmes. Ratings set now pay in later years, and management declines to predict payment year 2028, so no money has started. The payer is CMS, which the schema has no counterparty member for, so the tag is unknown. Outreach output is cheap to check.

Why this motive

An AI tool deployed toward a revenue that pays in later years, with management declining to predict it (claims c11 and c12): investing for the long term.

Before LLMs: relabelled

At the anchor Star Ratings bonus payments from CMS already existed (claim elv-anchor-c8) and Sydney Health already carried member engagement (claim elv-anchor-c3); total revenues were for fiscal 2024. No measure of AI-attributed ratings or revenue is given, so the activity is read as renamed.

“The Star Ratings are used by CMS to award quality-based bonus payments to plans that receive a rating of 4.0 or higher.”
Filing, business, 10-K periodic report, 2025-02-20

Figures

  • Share of Medicare Advantage members in plans rated four stars or higher under the 2026 Star Ratings, which set bonus payments in 2027 (approximately) · FY2027
  • Share of Medicare Advantage members in plans rated four stars or higher under the 2025 Star Ratings, used for payment year 2026 (approximately) · FY2026

What else could explain it

  • other: AI-powered engagement is named beside omni-channel outreach and rewards programmes (claim c11), and the same answer names clinical data interoperability, provider engagement and gap-closure programmes (claim c18).
  • other: Star Ratings move with plan mix, CMS methodology and clinical measures, not only survey scores.

Quotes

“As you know, stars remains one of our core enterprise priorities. We've made significant investments that we believe will continue to improve performance over time. We've enhanced our CAHPS infrastructure with AI-powered personalized member engagement, omni-channel outreach, and rewards programs.”
c11 · Executive, qa, earnings call, 2026-07-15
“While we're not prepared or should not be making predictions about payment year 2028 in the future, we feel really good about the trajectory of the business.”
c12 · Executive, qa, earnings call, 2026-07-15
“Our 2026 Star Ratings reflect that approximately 59% of our Medicare Advantage members are enrolled in plans rated at least 4.0 Stars or higher, or the equivalent.”
c16 · Filing, mdna, 10-Q periodic report, 2026-07-15
“We've also invested in clinical data interoperability, strengthening provider engagement, and expanding programs focusing on closing gaps in care, which is obviously very important.”
c18 · Executive, qa, earnings call, 2026-07-15

Reported lines, year-over-year growth

Revenue +1.4%

Q2 2026. Growing slower than revenue: benefit expense (+0.9%). 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.

Benefit expenseCost of products soldOperating expenseTotal expensesRevenue
-10%0%10%20%30%40%Q1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026Operating expenseCost of products soldTotal expensesRevenueBenefit expense
Reported values and filings