In Q2 2026 management spoke about AI at greater length and gave fewer dollars. The level of fiscal 2026 AI investment stated in Q1 was not repeated; the head of technology again said about of the year’s investments go into turning internal use cases into products sold outside the group, and the CFO expects the operating cost ratio at the higher end of its range partly because of AI investment. Under the ledger’s rules the level, given on the Q4 2025 call and in Q1, is withdrawn; the ledger keeps it as the basis and reads the expensed part of the quarter as and , now as its own inference, with the quarter’s share of the year from the fixed annual-plan band.
Claims processing, Optum Insight’s own service delivery and AI-enhanced fraud work open as channels: claims processing, which management says AI now automates even for complex claims; Optum Insight’s own service delivery, where the head of technology credits part of the segment’s quarter to AI efficiency gains while the 10-Q credits business services growth; and AI-enhanced fraud, waste and abuse work in the commercial business, tagged relabelled because the anchor already shows machine-learning fraud work. Both the delivery and fraud channels credit AI beside other causes with nothing to separate its part, so neither is sized. Optum Insight’s costs held flat while its revenues excluding investment income rose, against the prior-year rate, the only line that moved the way a saving would. Consolidated operating costs ran above their prior-year share of revenues, about the size of the AI spend; with it and the portfolio loss taken out remains, so the line does not show which way savings moved. No saving is measured, so every savings range starts at zero, and the UnitedHealthcare-side high ends stay under the nearly of 2026 cost reductions, many AI-enabled, given on the Q4 2025 call.
Management gave more rates and fewer costs: ambient listening available to of employed providers, case summaries faster for care managers, a real-time prior authorization target for 2027 and a prior authorization tool credited with administrative hours saved, which moves that channel’s motive from exploratory to efficiency, and contact centers getting far more efficient with savings partly reinvested. Optum Rx contact centers, measured in Q1, are not mentioned. The Optum Insight products widen to autonomous coding and Value Connect and remain unsized by management; the ledger puts their outside revenue at as outside revenue fell .
Sized channels against the income statement, Q2 2026
6 of 11 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.
Total revenuesreported line
$112.03bn
Total operating costsreported line
$104.04bn
Medical costsreported line
$75.36bn
Operating costsreported line
$14.27bn
Cost of products soldreported line
$13.38bn
Investment in AI across the company’s own processes and functionsspend · our inference
Investment in AI products and platforms sold through Optum Insightspend · our inference
Revenue from AI-first products and AI consulting sold by Optum Insightrevenue in · our inference
Claims processing cost displaced by AI at UnitedHealthcarecost displaced · our inference
Clinician and care-manager time saved by AI (ambient documentation, case summaries, scheduling)cost displaced · our inference
AI channel, dollars for the quarter low to high of an estimate reported lineLog scale: each gridline is ten times the one before.
New money and old money, Q2 2026
10 expanded1 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.
Flow, sized total
Split by novelty
Incremental total
Paid for AI$162mn to $450mn sized
expanded $162mn to $450mn
Incremental total $0 to $450mnpoint $0$313mn in 2 channels has no traced baseline
Cost displaced by AI$0 to $143mn sized
expanded $0 to $143mnrelabelled not sized
Incremental total $0 to $143mnpoint $24mn
Revenue arriving through AI$7.8mn to $62mn sized
expanded $7.8mn to $62mn
Incremental total $0 to $62mnpoint $0$31mn in 1 channel has no traced baseline
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 AI2 channels · $162mn to $450mn sized · $0 to $450mn incremental
engineering
Investment in AI products and platforms sold through Optum Insight
0.03% to 0.13% 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: exploratory· before LLMs: expanded
Withdrawn: the level, nearly , was given on the Q4 2025 call and again in Q1 and is not given in Q2. The Q2 call repeats that about of the year’s AI investment goes into products sold outside the group, a share with no restated base, and the CFO still names AI investment as a reason for the higher-end operating cost ratio. The quarter is the ledger’s inference from the earlier level, expensed.
Evidence: 7 quotes, 2 figures, 2 confounds, 6 from before coverage
The part of the company’s stated AI investment for 2026 that goes into software products and platforms: rebuilding Optum Insight’s products as AI-first software and services and turning internal AI use cases into products sold outside the group. Management gives the whole-year amount and the share, not the split between staff, vendors and capitalized software; the part that is capitalized reaches the income statement later as amortization, which the ledger cannot separate, so the channel is read as operating spend with that caveat.
Why this motive
Management describes a multi-year path of reinvestment and new AI products with early positive momentum from customers (claims c2 and c15): the separately priced product line is still described as early, so the Q1 reading is kept.
Before LLMs: expanded
At the anchor the FY2024 10-K already described AI and machine learning in its customer-facing technology products, expected generative AI to play a growing role in them, and said keeping pace would need significant development resources, with no amount; Optum Insight then had total revenues of , of which from customers outside the group. At the anchor the Q1 2024 call, held in the quarter of the Change Healthcare cyberattack, carried no AI passage. The plan to invest nearly in 2026 was first given on the Q4 2025 call (reference quarter). The earlier size of product development on AI is not traced, so the level has no baseline. The size is the whole of an activity that existed before.
“In addition, increasing connectivity among technologies and recent trends toward greater consumer engagement in health care require new and enhanced technologies, including more sophisticated applications for mobile devices and new tools and products that leverage AI/ML to improve the customer experience.”
“We anticipate that fast-evolving AI/ML technologies, including generative AI, will play an increasingly important role in our information systems and customer-facing technology products.”
“Our ability to protect and enhance existing systems and develop new systems to keep pace with changes in information processing technology (including AI/ML), regulatory standards and changing customer preferences will require an ongoing commitment of significant development and operational resources.”
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“Optum Insight connects the health care system with services, analytics and platforms that make clinical, administrative and financial processes simpler and more efficient for all participants in the health care system.”
“These efforts can help address long-standing needs in healthcare for simplicity, speed, certainty, insight, consumer empowerment, and convenience. We're hoping to invest nearly $1.5 billion in 2026 and expect at least as much to follow in 2027.”
Planned AI investment for fiscal 2026, as first given on the Q4 2025 call (nearly; reference quarter) · FY2026
Share of the year’s AI investment going into commercializing internal use cases as external products · FY2026
What else could explain it
line composition: The investment is not split between staff, vendors and capitalized software; the operating cost ratio it lifts also carries infrastructure, care delivery, customer experience and community spending (claim c31).
relabel: Older Optum Insight products are being moved to AI-based versions (Q1 claim c22), so part of the spend may be the existing product budget under a new name.
Quotes
“Optum Insight, also on plan, remains on a multi-year path of reinvestment and innovation as we bring modern intelligent technologies and services to the areas of greatest need in the health system. We believe Optum Insight is exceptionally well positioned to help modernize and simplify the health system as it brings AI-enabled tools and services to market.”
“The operating cost ratio was 12.7% for the quarter, compared to 12.3% a year ago, as we continue to focus on operating discipline while making targeted investments across technology, AI, care delivery enhancements, customer experience, and advancing healthier communities through the United Health Foundation.”
“We expect the operating cost ratio to come in at the higher end of our previously discussed range as a result of investments in our people, communities, and AI.”
“Q2 performance was driven by strong operational execution. Some of that actually due to the early AI investments we made in AI efficiency gains. There's also some client transaction volume that moved earlier into the year from H2 to H1, more than we expected. We continue to invest, even in the back half of this year, into new AI products and services where we see early positive momentum from customers.”
“As you are suggesting in your question, this is the beginning, but it will have compounding effects as we make these investments. We continue to get this change driven into our business. It's also a real catalyst and a real opportunity in terms of the Optum Insight business to take everything that we're doing here and bringing commercial versions to the marketplace.”
“Really one quick example that probably strings through everything that Tim Noel, and Patrick described is that about a third of our investments this year are going into commercializing all these internal use cases to external products. The prior auth example that Tim described, that Patrick described, has been converted into a digital prior auth product, which we launched about a quarter ago under the Optum Real family of products.”
“The operating cost ratio of 12.7% included targeted investments in infrastructure, artificial intelligence, care delivery enhancements, consumer experience and community support.”
Investment in AI across the company’s own processes and functions
0.11% to 0.27% 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: exploratory· before LLMs: expanded
Withdrawn: the level, nearly , was given on the Q4 2025 call and again in Q1 and is not given in Q2; the share was restated without a base. Management describes AI as the operating infrastructure of the future and the CFO still expects the operating cost ratio at the higher end of its range partly because of AI investment. Read from the earlier level, the quarter is expensed; the ratio was against .
Evidence: 12 quotes, 4 figures, 2 confounds, 4 from before coverage
The rest of the stated AI investment for 2026: AI deployed in the company’s own end-to-end processes (member service, prior authorization, claims, clinical workflows) and functions (HR, finance, legal, marketing). Management says every internal AI use case is routed through Optum Insight. Staff, vendor bills and model costs are not split out, and part may be capitalized.
Why this motive
The CEO calls the work the beginning, with compounding effects to come (claim c17), and the head of UnitedHealthcare places the efficiencies in 2027 and 2028 (claim c18): investing for later, with no displaced line shown shrinking.
Before LLMs: expanded
At the anchor the FY2024 10-K said the company increasingly relied on AI and machine learning in its internal operations and that keeping its systems current would take significant development and operational resources, with no amount, against operating costs of for the year. At the anchor the Q1 2024 call, held in the quarter of the Change Healthcare cyberattack, carried no AI passage. The plan to invest nearly in 2026 was first given on the Q4 2025 call (reference quarter). No earlier AI share of the spend is traced, so the level has no baseline. The size is the whole of an activity that existed before.
“We anticipate that fast-evolving AI/ML technologies, including generative AI, will play an increasingly important role in our information systems and customer-facing technology products.”
“Our ability to protect and enhance existing systems and develop new systems to keep pace with changes in information processing technology (including AI/ML), regulatory standards and changing customer preferences will require an ongoing commitment of significant development and operational resources.”
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“These efforts can help address long-standing needs in healthcare for simplicity, speed, certainty, insight, consumer empowerment, and convenience. We're hoping to invest nearly $1.5 billion in 2026 and expect at least as much to follow in 2027.”
Planned AI investment for fiscal 2026, as first given on the Q4 2025 call (nearly; reference quarter) · FY2026
Share of the AI investment spent across the company’s own processes and functions · FY2026
Operating cost ratio (operating costs as a share of total revenues) · 2026-CQ2
Operating cost ratio, prior year · 2025-CQ2
What else could explain it
line composition: Staff, vendor bills and any capitalized part are not separated; the head of UnitedHealthcare also mentions large-scale capital projects (claim c18).
transformation program: The investment runs beside the restructuring and portfolio actions begun in 2025, which are not AI.
Quotes
“AI technology is helping us move faster. We're using it to improve service interactions, reduce administrative burden, and support better decision-making, always in service of improved experiences and outcomes for both patients and care providers.”
“That evolution today includes a tech-forward view actively and appropriately embracing an AI paradigm for our businesses, a management team with skills and vision to help in building a more advanced health system, and an ever-evolving organizational structure and culture aligned to that system.”
“Optum Health will build upon this foundation with additional investments in clinical workflow improvements and network performance, more deeply embedding AI and automation to further improve operational performance and clinician experience.”
“The operating cost ratio was 12.7% for the quarter, compared to 12.3% a year ago, as we continue to focus on operating discipline while making targeted investments across technology, AI, care delivery enhancements, customer experience, and advancing healthier communities through the United Health Foundation.”
“We expect the operating cost ratio to come in at the higher end of our previously discussed range as a result of investments in our people, communities, and AI.”
“Well, we'll take this as a team activity, because actually AI is a very core initiative for us. We are really thinking of it in terms of reimagining our entire enterprise, virtually everything that we do, we see it basically as the operating infrastructure of the future.”
“As you are suggesting in your question, this is the beginning, but it will have compounding effects as we make these investments. We continue to get this change driven into our business. It's also a real catalyst and a real opportunity in terms of the Optum Insight business to take everything that we're doing here and bringing commercial versions to the marketplace.”
“I agree with your kind of commentary on the pacing, where we're going to see that accelerate into 2027 and even go deeper in terms of the efficiencies provided kind of across the board into 2028. As we're focused on this, you would imagine that we obviously have large-scale capital projects that we're deploying, but perhaps more importantly and significantly, we're infusing AI across all of our administrative activities.”
“Really one quick example that probably strings through everything that Tim Noel, and Patrick described is that about a third of our investments this year are going into commercializing all these internal use cases to external products. The prior auth example that Tim described, that Patrick described, has been converted into a digital prior auth product, which we launched about a quarter ago under the Optum Real family of products.”
“The operating cost ratio of 12.7% included targeted investments in infrastructure, artificial intelligence, care delivery enhancements, consumer experience and community support.”
“The operating cost ratio increased primarily due to investments in people, process and technology to drive improved consumer and care provider experiences and greater operating efficiencies, and business mix; partially offset by operating cost management and the revenue impacts of government programs.”
“The only thing I’ll say at the group level is every function, HR, finance, legal, everything, is really being reimagined in an AI context, which we think will drive much greater precision, greater efficiencies, responsiveness.”
Cost displaced by AI8 channels · $0 to $143mn sized · $0 to $143mn incremental · 5 not sized
customer support · cheap to verify
Member and provider service cost displaced by AI at UnitedHealthcare
Not sized
The quarter's only measure is a share by count: virtually every provider and consumer interaction uses AI (claim c19); the contact centers getting far more efficient carries no measure (claim c23). A share of interactions by count sizes nothing (the methodology rule for counts).
described, no sizedisclosure: direction only· motive: exploratory· before LLMs: expanded
Management says virtually every provider and consumer interaction uses AI and that contact centers are getting far more efficient, with no measure. A share of interactions by count reads directional and sizes nothing, so the channel is left unsized.
Evidence: 6 quotes, 2 confounds, 6 from before coverage
The cost of answering members and care providers (contact center advocates, digital self-service) at UnitedHealthcare that is lower than it would have been because AI answers questions (the Avery generative AI chatbot), supports advocates in every interaction, and moves contacts to digital channels. Management says the savings are partly reinvested in concierge-style service.
Why this motive
The head of UnitedHealthcare says contact centers are getting far more efficient and that part of the savings is reinvested (claim c23); no cost, headcount or rate is given, so the efficiency tell is not met and the reading stays exploratory.
Before LLMs: expanded
At the anchor the 10-K listed service and advanced technology, including digital consumer engagement, among UnitedHealthcare’s strengths and described tools that leverage AI and machine learning to improve the customer experience; on the Q4 2025 call (reference quarter) management said over of member calls already used AI tools. Service cost sat inside operating costs of for fiscal 2024, not split out. The size is the change AI made, not the whole line.
“In addition, increasing connectivity among technologies and recent trends toward greater consumer engagement in health care require new and enhanced technologies, including more sophisticated applications for mobile devices and new tools and products that leverage AI/ML to improve the customer experience.”
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“We depend on the integrity of the data in our information systems to implement new and innovative services, automate and deploy new technologies to simplify administrative processes and clinical decision making, price our products and services adequately, provide effective service to our customers and consumers in an efficient and uninterrupted fashion, provide timely payments to care providers, and accurately report our results of operations.”
“We anticipate operating cost reductions of nearly $1 billion in 2026, many AI-enabled and importantly resulting in higher customer experience and satisfaction at a lower cost. Over 80% of calls from members leverage AI tools to help answer members' questions faster and more accurately.”
other: Savings are said to be partly reinvested in concierge-style service (claim c23), so a gain may not lower the line.
line composition: Member service cost sits inside operating costs with every other administrative cost.
Quotes
“AI technology is helping us move faster. We're using it to improve service interactions, reduce administrative burden, and support better decision-making, always in service of improved experiences and outcomes for both patients and care providers.”
“AI is both an enabler and accelerant to this effort. We're early in this work, but clearly on the path to improve the healthcare experience and strengthen relationships with our stakeholders, starting with consumers and care providers.”
“I agree with your kind of commentary on the pacing, where we're going to see that accelerate into 2027 and even go deeper in terms of the efficiencies provided kind of across the board into 2028. As we're focused on this, you would imagine that we obviously have large-scale capital projects that we're deploying, but perhaps more importantly and significantly, we're infusing AI across all of our administrative activities.”
“First, virtually every provider and consumer interaction uses AI. AI gives us the ability to empower our advocates with predictive insights, real-time data, which all leads to more productive interactions, and importantly, also lends itself to translating some of those experiences into digital experiences that oftentimes, most times, come with better provider and consumer satisfaction as well.”
“The second thing to note and remind folks of is that our contact centers are getting far more efficient, but they will never be fully automated. Engaging with the healthcare system that we play an important role in will always be deeply personal. We see great opportunity to invest some of the savings that we’re getting from these efficiencies in actually enhanced and more concierge service-like models.”
“The operating cost ratio increased primarily due to investments in people, process and technology to drive improved consumer and care provider experiences and greater operating efficiencies, and business mix; partially offset by operating cost management and the revenue impacts of government programs.”
Q1 2026direction only · described, no size · exploratory
Q2 2026direction only · described, no size · exploratory
customer support · cheap to verify
Pharmacy member contact-center cost displaced by AI at Optum Rx
Not sized
The Q2 call, release and 10-Q say nothing about Optum Rx contact centers or AI self-service, and the contact-center cost is not split out in any filing.
inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: expanded
Not mentioned in Q2. Optum Rx adjusted scripts fell , which lowers contact volume on its own.
Evidence: 0 quotes, 3 figures, 6 from before coverage
Contact-center cost at Optum Rx, the pharmacy benefit manager, that is lower because digital and AI-enabled self-service take calls out of the queue. Split from UnitedHealthcare member service because it is a different business, function and contact base.
Why this motive
Carried from Q1, where the call-volume cut was credited to digital and AI-enabled self-service together (Q1 claim c13), which reads exploratory; silence this quarter is not evidence about motive.
Before LLMs: expanded
At the anchor Optum Rx already offered digital tools and services to its clients and their members, fulfilled adjusted scripts in fiscal 2024, and credited part of its earnings to operating cost efficiencies without naming AI; on the Q4 2025 call (reference quarter) it described significant AI automation-enabled operating efficiencies. Its contact-center cost is inside operating costs and is not split out. The size is the change AI made, not the whole line.
“In addition, increasing connectivity among technologies and recent trends toward greater consumer engagement in health care require new and enhanced technologies, including more sophisticated applications for mobile devices and new tools and products that leverage AI/ML to improve the customer experience.”
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“Optum Rx offers multiple clinical programs, digital tools and services to help clients manage overall pharmacy and health care costs in a clinically appropriate manner which are designed to deliver improved consumer experiences, better health outcomes and a lower total cost of care.”
“While new customer wins were offset by membership contraction at UnitedHealthcare, we are capitalizing on significant AI automation-enabled operating efficiencies to support our expanded margin outlook in 2026.”
Optum Rx adjusted scripts, change year over year · 2026-CQ2
By quarter
Q1 2026described · described, no size · exploratory
Q2 2026not mentioned · inscrutable · exploratory
back office · expensive to verify
Prior authorization handling cost displaced by AI
0% to 0.03% of the quarter’s revenue
Incremental total: counts in full.
our inferencedisclosure: described· motive: efficiency· before LLMs: expanded
UnitedHealthcare commits to processing of prior authorizations in real time by the end of 2027 and to cutting volume by policy this year; a clinician still makes every denial. The size shown is the ledger’s own estimate, .
Evidence: 7 quotes, 3 figures, 2 confounds, 4 from before coverage
The cost of receiving, reviewing and deciding prior authorization requests at UnitedHealthcare and Optum Rx that is lower because AI automates intake and real-time decisions. A denial stays with a clinician, and errors are costly and regulated, so the work is expensive to verify. Reductions in the number of authorizations required are a policy change, not an AI saving, and are read as a confound. The version sold to outside payers and providers belongs to the Optum Insight product channel.
Why this motive
An implemented tool with a measured operating result attributed to it (motive table, efficiency row as ruled 2026-10-06): the digital prior authorization tool built from the internal use case is in production and is credited with about authorizations processed and administrative hours saved year to date (claim c29). No cost line is shown to move. The "early" remark (claim c5) refers to the wider simplification effort, not this tool.
Before LLMs: expanded
At the anchor the 10-K described automating and deploying technology to simplify administrative processes and clinical decision making, and Optum Rx already ran utilization management programs; authorization handling sat inside operating costs of for fiscal 2024, not split out. The size is the change AI made, not the whole line.
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“We depend on the integrity of the data in our information systems to implement new and innovative services, automate and deploy new technologies to simplify administrative processes and clinical decision making, price our products and services adequately, provide effective service to our customers and consumers in an efficient and uninterrupted fashion, provide timely payments to care providers, and accurately report our results of operations.”
“Optum Rx provides various utilization management, medication management, quality assurance, adherence and counseling programs to complement each client’s plan design and clinical strategies.”
“We anticipate operating cost reductions of nearly $1 billion in 2026, many AI-enabled and importantly resulting in higher customer experience and satisfaction at a lower cost. Over 80% of calls from members leverage AI tools to help answer members' questions faster and more accurately.”
Share of prior authorizations to be processed in real time by the end of 2027 (target) · as-of 2027-12-31
Prior authorizations processed by the digital prior auth product year to date, about, apparently including the group’s own use (the external count is unclear in the transcript) · 2026-01-01..2026-07-16
Administrative hours saved by the digital prior auth product, year to date · 2026-01-01..2026-07-16
What else could explain it
other: Eliminating authorization volume and requirements by policy (claim c4) lowers handling cost with no AI.
line composition: Authorization handling sits inside operating costs and is not split out.
Quotes
“AI technology is helping us move faster. We're using it to improve service interactions, reduce administrative burden, and support better decision-making, always in service of improved experiences and outcomes for both patients and care providers.”
“For example, in the quarter, we committed to eliminating, by the end of this year, 30% of prior authorization volume and nearly two-thirds of prior authorization requirements for pediatric care.”
“AI is both an enabler and accelerant to this effort. We're early in this work, but clearly on the path to improve the healthcare experience and strengthen relationships with our stakeholders, starting with consumers and care providers.”
“I agree with your kind of commentary on the pacing, where we're going to see that accelerate into 2027 and even go deeper in terms of the efficiencies provided kind of across the board into 2028. As we're focused on this, you would imagine that we obviously have large-scale capital projects that we're deploying, but perhaps more importantly and significantly, we're infusing AI across all of our administrative activities.”
“That all comes to light in our commitment to process 80% of our prior authorizations in real time by the end of 2027. In doing so, it creates kind of a touchless environment, which eliminates a lot of the back and forth between health systems, care providers, and health plans, which not only improves experiences, but you can imagine the elimination of the abrasion unlocks a lot of operational efficiencies for both health systems and health plans.”
“Year to date, it has, for external entities to UHG, processed about 69,000 prior auths, so it's processed about half a million prior auths and saved 69,000 administrative hours. This is in real life, in real time, an internal use case which we are investing in giving us internal efficiencies that is being converted into a commercial external product, helping improve the system and making it better for everyone.”
Claims processing cost displaced by AI at UnitedHealthcare
0% to 0.05% of the quarter’s revenue
Incremental total: counts in full.
our inferencedisclosure: described· motive: exploratory· before LLMs: expanded
Opened this quarter: the head of UnitedHealthcare says AI now automates complex claims that were never expected to be automated, with higher accuracy, eliminating cost. No figure is given. The size shown is the ledger’s own estimate, .
Evidence: 2 quotes, 1 confound, 4 from before coverage
The cost of processing claims at UnitedHealthcare, including complex claims that management says it never expected to automate, that is lower because AI processes them with higher accuracy. Payment errors are costly, so the work is expensive to verify. Opened in Q2 2026, when management first named claims.
Why this motive
Management says complex claims are now automated and that this eliminates cost (claim c20), with no measure, and places the efficiencies as accelerating into 2027 and 2028 (claim c18). With no line shown shrinking, exploratory.
Before LLMs: expanded
At the anchor the 10-K described automating administrative processes and providing timely payments to care providers as uses of its information systems; claims processing sat inside operating costs of for fiscal 2024, not split out. The size is the change AI made, not the whole line.
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“We depend on the integrity of the data in our information systems to implement new and innovative services, automate and deploy new technologies to simplify administrative processes and clinical decision making, price our products and services adequately, provide effective service to our customers and consumers in an efficient and uninterrupted fashion, provide timely payments to care providers, and accurately report our results of operations.”
“We anticipate operating cost reductions of nearly $1 billion in 2026, many AI-enabled and importantly resulting in higher customer experience and satisfaction at a lower cost. Over 80% of calls from members leverage AI tools to help answer members' questions faster and more accurately.”
line composition: Claims processing cost sits inside operating costs and may partly be work done by Optum Insight for the group.
Quotes
“I agree with your kind of commentary on the pacing, where we're going to see that accelerate into 2027 and even go deeper in terms of the efficiencies provided kind of across the board into 2028. As we're focused on this, you would imagine that we obviously have large-scale capital projects that we're deploying, but perhaps more importantly and significantly, we're infusing AI across all of our administrative activities.”
“Second area is it's providing step function enhancements in our core operations. Things like very complex claims that we never before thought we would be able to automate. We're able to automate those and process those with higher accuracy than we have been able to before, which eliminates cost and also increases turnaround times, which is very important.”
Clinician and care-manager time saved by AI (ambient documentation, case summaries, scheduling)
0% to 0.05% of the quarter’s revenue
Incremental total: counts in full.
our inferencedisclosure: described· motive: exploratory· before LLMs: expanded
Ambient listening is available to of employed Optum Health providers, on track for over by year end, and AI case summaries make nurse care managers faster; AI scheduling is described without a measure. The size shown is the ledger’s own estimate, .
Evidence: 5 quotes, 3 figures, 2 confounds, 4 from before coverage
Clinical staff time at Optum Health and UnitedHealthcare that AI frees: ambient listening that writes clinical documentation, case summaries for nurse care managers and clinical reviews, and AI-enabled scheduling that routes patients to the right appointment. Management describes the freed time as more patient-facing hours and less burnout, so much of it may show as capacity rather than lower cost. Clinical work is expensive to verify. Scheduling is held here until management separates it.
Why this motive
Rollout figures (ambient listening available to of employed providers) and a task speed-up (claim c24) with the time framed as more time with patients: capacity, not a shrinking line, so the reading stays exploratory.
Before LLMs: expanded
At the anchor the company had nearly employees, more than of them clinical professionals, and Optum Health already served patients through digital health technologies such as telehealth and remote monitoring; the 10-K described technology to simplify clinical decision making. None of it was AI documentation. The size is the change AI made, not the whole line.
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“We depend on the integrity of the data in our information systems to implement new and innovative services, automate and deploy new technologies to simplify administrative processes and clinical decision making, price our products and services adequately, provide effective service to our customers and consumers in an efficient and uninterrupted fashion, provide timely payments to care providers, and accurately report our results of operations.”
“Our nearly 400,000 employees, as of December 31, 2024, including more than 140,000 clinical professionals, are guided by our mission to help people live healthier lives and help make the health system work better for everyone.”
Share of Optum Health employed providers with AI-based ambient listening available · as-of 2026-07-16
Share of employed providers planned to have ambient listening by the end of 2026 (a floor) · as-of 2026-12-31
Speed-up for nurse care managers from AI case summaries · 2026-CQ2
What else could explain it
transformation program: Patient-facing hours and patient experience are credited to the operating changes at Optum Health as a whole, not to AI alone.
line composition: Clinical staff cost is split between operating costs and medical costs and is not reported.
Quotes
“There is real progress on the rollout of AI-based ambient listening capabilities, available to 70% of our employed providers today and on track to exceed 90% by year-end.”
“Optum Health will build upon this foundation with additional investments in clinical workflow improvements and network performance, more deeply embedding AI and automation to further improve operational performance and clinician experience.”
“First area I'll call out is administrative efficiency, where we're using AI to summarize cases, for example, for care managers. Makes it 40% faster for that nurse care manager, more efficient, also allows him or her to spend more time with patients, more time delivering care in the home or whatever the setting may be.”
Corporate function cost displaced by AI (HR, finance, legal, marketing)
Not sized
Withdrawn from the ledger: the only support was a one-clause aim (functions being reimagined around AI) with no tool, deployment or measure, which is a context claim and not a channel (the methodology rule for AI named beside another cause). The former estimate is kept in the figures, unreferenced.
inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: expanded
Withdrawn from the ledger as not an AI channel under the one-clause-aim rule. The passages on reimagining corporate functions around AI are kept as context claims with no channel.
Evidence: 0 quotes, 4 from before coverage
Withdrawn from the ledger as not an AI channel under the rule for a one-clause aim with no tool, deployment or measure (the methodology rule settled in wave C, applied in the rules sweep of 2026-10-06). The only support in either quarter was a clause saying HR, finance, legal, marketing and every other function are being reimagined around AI, with no tool named, no deployment described and no measure; those passages are kept as context claims. Registered at first as the cost of the corporate functions lower because management is reimagining them around AI; the id is kept because a published entry cites its former size.
Why this motive
Carried; the channel is withdrawn from the ledger and no source is read for it.
Before LLMs: expanded
At the anchor the company had nearly employees and operating costs of for fiscal 2024, and relied on AI and machine learning in its internal operations; corporate function cost is not split out. The size is the change AI made, not the whole line.
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“We depend on the integrity of the data in our information systems to implement new and innovative services, automate and deploy new technologies to simplify administrative processes and clinical decision making, price our products and services adequately, provide effective service to our customers and consumers in an efficient and uninterrupted fashion, provide timely payments to care providers, and accurately report our results of operations.”
“Our nearly 400,000 employees, as of December 31, 2024, including more than 140,000 clinical professionals, are guided by our mission to help people live healthier lives and help make the health system work better for everyone.”
“We anticipate operating cost reductions of nearly $1 billion in 2026, many AI-enabled and importantly resulting in higher customer experience and satisfaction at a lower cost. Over 80% of calls from members leverage AI tools to help answer members' questions faster and more accurately.”
Optum Insight service-delivery cost displaced by AI
Not sized
The head of technology credits some of the segment’s quarter to AI efficiency gains beside client volume moved into the first half (claim c15), and the filings credit business services growth and contract timing; nothing separates AI’s part (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: expanded
Opened this quarter. The head of technology says some of Optum Insight’s better quarter came from early AI efficiency gains, beside client volume that moved earlier into the year. The segment’s costs held flat while revenues excluding investment income rose, against the prior-year rate. Nothing separates AI’s part, so the channel is left unsized; it would in any case stay out of totals, because part of the segment’s work is the group’s own claims and authorization processing.
Evidence: 3 quotes, 2 figures, 2 confounds, 3 from before coverage
The cost of delivering Optum Insight’s business and technology services (revenue cycle management, coding, administrative operations for the group and outside clients) that is lower because of AI efficiency gains, which management credits for part of the segment’s Q2 2026 earnings. Optum Insight also runs services for UnitedHealthcare, so part of this saving may be the same cost pool as the group’s claims and prior authorization savings; the reading is kept out of totals as an overlap. Opened in Q2 2026.
Why this motive
The head of technology credits part of the quarter to AI efficiency gains (claim c15) and the segment’s costs held flat as revenue grew, close to the efficiency tell; but the 10-Q and the release credit business services growth, operational improvements and contract timing (claims c35 and c33), so with the tells in conflict the less durable reading is taken.
Before LLMs: expanded
At the anchor Optum Insight already sold services that make clinical, administrative and financial processes more efficient and revenue cycle management to care providers, on total revenues of for fiscal 2024; its delivery cost was not split out and no AI share was named. The size is the change AI made, not the whole line.
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“Optum Insight connects the health care system with services, analytics and platforms that make clinical, administrative and financial processes simpler and more efficient for all participants in the health care system.”
“Serves hospitals, physicians and other care providers to improve operating performance, better coordinate care and reduce administrative costs through technology and services to improve population health management, patient engagement, revenue cycle management and strategic growth plans.”
Optum Insight operating costs as a share of its revenues excluding investment and other income · 2026-CQ2
Optum Insight operating costs as a share of its revenues excluding investment and other income, prior year · 2025-CQ2
Reported line it is matched to
Optum Insight operating costs were against while its revenues excluding investment and other income, which has no cost behind it, were against : of revenues against , against the prior-year rate. The line moved the way the claim predicts; the filing credits business services growth and higher investment income.
seasonality: Client transaction volume moved from the second half into the first (claim c15), raising revenue against a fixed cost base.
acquisition: Business sales in the first half change the segment’s cost and revenue base for reasons unrelated to AI.
Quotes
“Q2 performance was driven by strong operational execution. Some of that actually due to the early AI investments we made in AI efficiency gains. There's also some client transaction volume that moved earlier into the year from H2 to H1, more than we expected. We continue to invest, even in the back half of this year, into new AI products and services where we see early positive momentum from customers.”
“Second quarter 2026 earnings from operations were $1.4 billion compared to $1.2 billion in the second quarter 2025. The year-over-year increase was driven by operational improvements and timing of contracts.”
“Revenues at Optum Insight increased due to elevated investment and other income and growth in business and technology services. Earnings from operations for the three months ended June 30, 2026, increased due to growth in business services.”
Medical cost avoided by AI-enhanced fraud, waste and abuse work
Not sized
AI-enhanced fraud, waste and abuse work is named as one of several bases of the commercial margin recovery, beside administrative cost efficiency and medical cost affordability (claim c12), with no measure of its part (the methodology rule for AI named beside another cause).
described, no sizedisclosure: described· motive: exploratory· before LLMs: relabelled
Opened this quarter and tagged relabelled: the head of the commercial business names AI-enhanced fraud, waste and abuse efforts among the bases of margin recovery, with no figure and no generative tool named; the anchor already ties AI and machine learning to fraud detection. Medical costs were against . Nothing separates AI’s part of the recovery, so the channel is left unsized.
Evidence: 3 quotes, 2 figures, 3 confounds, 3 from before coverage
Medical costs that UnitedHealthcare does not pay because AI-enhanced fraud, waste and abuse work finds improper claims. The counterparty is the provider or member whose claim is not paid, which the schema has no member for; the channel is tagged internal as the work is the company’s own. Management names it as part of the commercial margin recovery. Errors fall on patients and providers, so the work is expensive to verify. Opened in Q2 2026.
Why this motive
Named among the bases of a commercial margin recovery that now runs past 2027 (claims c12 and c13): a plan, with no measure, so exploratory.
Before LLMs: relabelled
At the anchor Optum Insight already sold a comprehensive payment integrity portfolio to health plans, the 10-K tied its AI and machine learning systems to preventing, detecting and controlling fraud, and medical costs were for fiscal 2024. The coverage evidence is a single phrase, AI-enhanced fraud, waste and abuse efforts, with no generative tool named and no line, rate or volume shown to move, so the channel reads as machine-learning fraud work the anchor already shows under the AI name; a named generative tool or a measure would be a step to expanded.
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“Serves health plans by improving financial performance and enhancing outcomes through proactive analytics, a comprehensive payment integrity portfolio and technology-enabled and staff-supported risk and quality services.”
“including systems powered by or incorporating artificial intelligence and machine learning (AI/ML), we could experience failures in our health, wellness and information technology products; lose existing customers; have difficulty attracting new customers; experience problems in determining medical cost estimates and establishing appropriate pricing; have difficulty preventing, detecting and controlling fraud”
other: Payment integrity and fraud work predate AI at the anchor (anchor claims unh-anchor-c7 and unh-anchor-c13); medical costs move far more with benefit design, membership, reserve development and trend than with this work.
other: Provider coding intensity is named as a driver of commercial cost trend (claim c14); management does not attribute it to AI, so no toll channel is registered for it.
other: Administrative cost efficiency and medical cost affordability are named beside AI-enhanced fraud work as bases of the same margin recovery (claim c12).
Quotes
“That performance is pacing, that recovery is built on a few things. Improved administrative cost efficiency, AI enhanced fraud, waste, and abuse efforts, and diligent focus, as always, on medical cost affordability.”
“The sticky nature of the persistent and elevated trend is extending the timeframe for full margin recovery past 2027, as we’ve previously discussed and you highlighted.”
“Additionally, provider coding intensity with office visits, emergency departments, and selective other care sites, being the primary drivers, is also contributing incremental trend to last year and to our expectations.”
Revenue arriving through AI1 channel · $7.8mn to $62mn sized · $0 to $62mn incremental
product revenue · expensive to verify
Revenue from AI-first products and AI consulting sold by Optum Insight
0.01% to 0.06% 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
Management names more products (AI-enabled and autonomous coding, Value Connect, digital prior authorization within Optum Real) and usage, about prior authorizations processed year to date, apparently including the group’s own, but no revenue. Outside revenue was against . The size shown is the ledger’s own estimate, .
Evidence: 10 quotes, 8 figures, 2 confounds, 6 from before coverage
Revenue from customers outside the group for Optum Insight’s AI-first products and services: the Optum Real platform for claim adjudication, coverage validation and digital prior authorization, AI-enabled coding, the Value Connect insights platform, and the Optum AI consulting arm. The products automate payment and clinical-administrative work, so their output is expensive to verify. Revenue from the group’s own businesses is eliminated in consolidation and is not part of the channel. The CFO describes the products as replacing older products that were not AI-based, so the channel is a level of an existing line, not new money.
Why this motive
Priced AI products with client results (claims c8, c9 and c32) meet the offensive tell, but management calls the momentum early and the segment’s outside revenue fell; with the tells in conflict the Q1 reading is kept.
Before LLMs: expanded
At the anchor Optum Insight already sold services, analytics and platforms that make clinical, administrative and financial processes more efficient, a payment integrity portfolio to health plans and revenue cycle management to care providers, with revenues from customers outside the group of for fiscal 2024, a figure that predates the 2026 realignment that moved Optum Financial (Optum Bank) into the segment, so it is not like for like with the covered quarters; the 10-K described products that leverage AI and machine learning. No AI-first product revenue was given then or since. The size is the whole of an activity that existed before.
“In addition, increasing connectivity among technologies and recent trends toward greater consumer engagement in health care require new and enhanced technologies, including more sophisticated applications for mobile devices and new tools and products that leverage AI/ML to improve the customer experience.”
“We anticipate that fast-evolving AI/ML technologies, including generative AI, will play an increasingly important role in our information systems and customer-facing technology products.”
“As an enterprise, we increasingly rely on new and evolving technologies, including those powered by or incorporating AI/ML, as part of our internal operations and in the delivery of our products and services.”
“Optum Insight connects the health care system with services, analytics and platforms that make clinical, administrative and financial processes simpler and more efficient for all participants in the health care system.”
“Serves health plans by improving financial performance and enhancing outcomes through proactive analytics, a comprehensive payment integrity portfolio and technology-enabled and staff-supported risk and quality services.”
“Serves hospitals, physicians and other care providers to improve operating performance, better coordinate care and reduce administrative costs through technology and services to improve population health management, patient engagement, revenue cycle management and strategic growth plans.”
Optum Insight total revenues (affiliated and unaffiliated) · 2026-CQ2
Optum Insight total revenues, prior year · 2025-CQ2
Optum Insight revenues from unaffiliated customers · 2026-CQ2
Optum Insight revenues from unaffiliated customers, prior year · 2025-CQ2
Optum Insight revenues from unaffiliated customers, growth year over year · 2026-CQ2
First-pass approval rate for digital prior authorization powered by AI · 2026-CQ2
Administrative hours saved by the digital prior auth product, year to date · 2026-01-01..2026-07-16
Pharmacy cost reduction in early Value Connect client results (management claim) · 2026-CQ2
What else could explain it
relabel: AI-first products replace older Optum Insight products (Q1 claim c22); part of their revenue is revenue the anchor already shows.
other: From 2026 the segment includes Optum Financial (Optum Bank), which sells no AI product, and the anchor’s fiscal 2024 figure predates that realignment. Optum Insight revenue from outside the group fell after business sales; the client results quoted (a pharmacy cost reduction, approval rates, hours saved) are management claims about customers, not revenue.
Quotes
“Optum Insight, also on plan, remains on a multi-year path of reinvestment and innovation as we bring modern intelligent technologies and services to the areas of greatest need in the health system. We believe Optum Insight is exceptionally well positioned to help modernize and simplify the health system as it brings AI-enabled tools and services to market.”
“Moving to Optum Insight. AI-enabled approaches continue to gain traction as more payer and provider customers seek differentiated capabilities to drive better performance. The emerging suite of products includes solutions such as AI-enabled coding, real-time payer and provider interfaces, and clinical quality and safety support.”
“For example, Value Connect is an AI-driven insights platform integrated into provider workflows and electronic health records to improve value-based care performance. Early client results include a 17% reduction in pharmacy costs.”
“Q2 performance was driven by strong operational execution. Some of that actually due to the early AI investments we made in AI efficiency gains. There's also some client transaction volume that moved earlier into the year from H2 to H1, more than we expected. We continue to invest, even in the back half of this year, into new AI products and services where we see early positive momentum from customers.”
“As you are suggesting in your question, this is the beginning, but it will have compounding effects as we make these investments. We continue to get this change driven into our business. It's also a real catalyst and a real opportunity in terms of the Optum Insight business to take everything that we're doing here and bringing commercial versions to the marketplace.”
“Really one quick example that probably strings through everything that Tim Noel, and Patrick described is that about a third of our investments this year are going into commercializing all these internal use cases to external products. The prior auth example that Tim described, that Patrick described, has been converted into a digital prior auth product, which we launched about a quarter ago under the Optum Real family of products.”
“Year to date, it has, for external entities to UHG, processed about 69,000 prior auths, so it's processed about half a million prior auths and saved 69,000 administrative hours. This is in real life, in real time, an internal use case which we are investing in giving us internal efficiencies that is being converted into a commercial external product, helping improve the system and making it better for everyone.”
“Revenues at Optum Insight increased due to elevated investment and other income and growth in business and technology services. Earnings from operations for the three months ended June 30, 2026, increased due to growth in business services.”
Q2 2026. Growing slower than revenue: medical costs (−4.1%), total operating costs (−2.3%). 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.
Medical costsOperating costsCost of products soldTotal operating costsRevenue