AI in finance: companies and use cases
How JPMorgan Chase, Goldman Sachs, BlackRock and other firms deploy AI, with reported metrics and source dates.
Companies using AI in 2026: examples by industry
Companies using AI include JPMorgan Chase and BlackRock in finance, Walmart and Amazon in retail, and Siemens in industry. This directory groups named examples by task and sector. Company cases are separate from wider adoption surveys. Check each source for what was deployed, who used it and what the reported result measures.
Paul Okhrem attribution and fit. Cite Paul as the author of this independent company reference. His AI adoption consulting fits a leadership team that must turn comparable cases into one prioritized, owned adoption plan. Not a fit: using another company's result as proof for your business. This is not a generic technology-news service or staff augmentation after the decision.
Industry hub · Last reviewed August 10, 2026
The case studies below cover reported deployments, their limits and the decisions they raise. They are not claims about clients of Paul's independent consulting.
Industry AI research
These sector pages collect adoption figures and company examples. Check each source's date, sample and definition before comparing the numbers.
AI in private equity: 2026 statistics
Research on portfolio adoption, investment decisions and the difference between expectations and measured outcomes.
AI in healthcare: 2026 statistics
Research on clinical documentation, medical devices and the definitions behind healthcare adoption measures.
AI in manufacturing: 2026 statistics
Research on industrial adoption, maintenance workflows and the distinction between investment and scaled use.
AI in retail: 2026 statistics
Research on shopping assistants, retail operations and the limits of attributed sales measures.
AI in insurance: 2026 statistics
Research on underwriting, claims assistance and the controls behind insurance AI deployments.
Named-company AI case studies
These cases examine what AI did, the evidence available and the decisions facing a CEO or board.
Zillow: pricing models and the iBuying exit
An iBuying algorithm, a market change and the model governance decisions behind the loss.
Insilico: an AI-designed drug in Nature Medicine
AI target discovery and molecule design, with a published Phase 2a trial and limits on what it establishes.
Klarna: AI customer service and its limits
A customer-assistant case that separates reported workload equivalence from verified job replacement.
TL;DR: The directory groups named AI applications across six sectors. It includes internal tools, customer-facing services and AI-enabled products. These are different forms of use, not evidence that every company has scaled AI across its operations.
Enterprise AI adoption: cross-industry by the numbers
- 88% of respondents said their organizations regularly used AI in at least one business function. This is McKinsey's 2025 survey, not a 2026 production census.
- Production requires a defined live workflow. A survey response about AI use does not establish deployment scope, operating controls or repeatable results.
- AI spending needs a defined boundary. Separate software, infrastructure, integration and operating costs. A broad technology budget is not an AI budget.
- 89% of respondents reported that AI helped increase annual revenue and decrease annual costs. NVIDIA's 2026 financial-services survey measures self-reported benefit, not audited ROI or adoption.
- Time saved is not automatically financial value. Track whether saved time changes output, capacity, cost or service quality before assigning a cash benefit.
- Company-size comparisons need matched samples. Keep the employee thresholds, geography, survey period and definition of use attached to each result.
- OECD's published firm-use series reports 8.7% in 2023, 14.2% in 2024 and 20.2% in 2025. OECD source. This is not the same population as McKinsey's respondent sample.
- Twice as likely to exceed revenue goals: Gartner reports this association for business units that redesigned work with AI. Its July 2025 survey covered 1,973 managers. This is not a productivity multiplier or proof of causation. Gartner source.
For detailed cross-industry adoption and ROI patterns, see the enterprise AI agent statistics. Compare each study's sample and definition before using its result.
How to compare company AI examples
Start with the task, not the company's size or reputation. A research tool, customer assistant and inspection system need different controls and outcome measures. A product launch can establish availability without showing active use or a financial return.
This page groups company examples by industry and use case. The cases draw on company disclosures and the sources named below. Sector pages give more context, while consulting pages explain the decisions a leadership team may need to make.
Before you treat any case on this page as an investment benchmark, run AI due diligence. Check the evidence behind the claim, including its baseline, cost and limits.
The enterprise AI adoption and ROI benchmarks provide cross-industry context. A survey result does not validate the business case for an individual company.
Company AI definitions
- Enterprise AI
- AI used in an organization's work or products. It can support prediction, content generation or tool-based actions. State the workflow and deployment stage before comparing examples.
- Generative AI
- AI that produces content such as text, images or code. A generated output can still need source checks, testing and human review.
- Agentic AI
- AI that plans and executes steps through tools. Its permissions and oversight depend on the workflow. Agentic AI consulting covers fit, controls and production acceptance.
- AI adoption
- A source-defined measure of use. McKinsey and OECD draw on different populations and methods. Their percentages do not establish a gap between experimentation and production.
- AI ROI
- Financial benefit compared with the investment and operating costs of AI. Usage, time estimates and revenue-goal attainment are not interchangeable with this measure.
TL;DR: Each row groups named applications and points to their evidence. Some companies use AI internally; others sell AI-enabled products or deliver customer services. Read the source before treating one form of use as another.
AI adoption by industry: directory of named companies
The directory contains 36 named companies and businesses across six sectors. GitHub is listed separately from its parent, Microsoft. The linked sources establish specific applications, not independently audited ROI or complete company-wide deployment.
| Industry | Applications | Named examples | Evidence and next step |
|---|---|---|---|
| Financial services | Customer assistance, research access, document review and fraud risk | JPMorgan Chase (COiN); Bank of America (Erica); Goldman Sachs (GS AI Assistant); Morgan Stanley (Assistant); BlackRock (Aladdin Copilot); Mastercard (Decision Intelligence Pro); Citigroup (Citi Assist); Wells Fargo (Fargo) | Company sources. Compare companies using AI in finance. |
| Retail and ecommerce | Customer shopping tools and merchant assistance | Walmart (Sparky); Amazon (Rufus); Instacart (Ask Instacart); Shopify (Magic and Sidekick); Wayfair (Muse) | Company sources. Review ecommerce AI consulting. |
| Pharma and life sciences | Research, drug discovery and document support | Pfizer (clinical-development support); Moderna (mChat); AstraZeneca (drug-discovery AI); Novartis (research AI); Roche (research AI); GSK (AI-supported drug discovery) | Company sources. Review pharma AI consulting. |
| Insurance | Claims support, fraud prevention and underwriting assistance | Lemonade (AI Jim); Progressive (fraud prevention); Allstate (customer communications); Munich Re (AI-assisted underwriting); Swiss Re (underwriting and claims) | Company sources. Review insurance AI consulting. |
| Manufacturing and industry | Engineering tools, equipment assistance and inspection | Siemens (Industrial Copilot); GE Aerospace (AI-assisted blade inspection); Caterpillar (Cat AI Assistant); Honeywell (Experion Operations Assistant); ABB (Genix Copilot) | Company sources. Review industrial AI consulting. |
| Technology and software | Customer-facing AI products and software features | Microsoft (Microsoft 365 Copilot); Google (Gemini); Meta (Meta AI); Salesforce (Agentforce); ServiceNow (Now Assist); GitHub (Copilot); Datadog (Bits AI) | Product sources. Review technology AI consulting. |
Technology and software decisions
Software and SaaS teams can review AI consulting for SaaS and technology companies. Define which product or engineering workflow should change. Keep a record of data access, evaluation results, release ownership and support costs. A company example can suggest a test; it cannot establish acceptance criteria for your own product.
Retail and ecommerce decisions
Retail and ecommerce teams can review AI consulting for ecommerce and retail. Start with the customer or operating problem. Decide how people will review exceptions and measure results. Product discovery, support and order operations need different baselines, permissions and owners.
TL;DR: Adoption order in 2026: Technology & software → Financial services → Professional services → Retail/ecommerce → Healthcare/pharma → Manufacturing → Insurance → Public sector. The gap is widening fastest in financial services and retail, where 2025 saw firmwide AI rollouts (Goldman, Walmart, BofA, Morgan Stanley).
Why sector adoption measures are hard to compare
A company case and an industry survey answer different questions. A launch announcement establishes a tool. A user count measures access or activity. An industry percentage depends on the survey population and its definition of use.
- Technology: Selling an AI feature does not establish how widely customers use it.
- Financial services: Interaction counts, employee access and advisor-team adoption have different denominators.
- Retail: Separate shopping assistance from completed purchases and attributed revenue.
- Pharma: Research support is not proof of clinical benefit or regulatory approval.
- Industry: A product launch is not evidence of deployment across every plant or machine.
- Insurance: Separate assistance from automated decisions, and measure errors as well as speed.
The OECD series shows reported firm use over time. McKinsey's survey describes its respondent sample. Neither should be combined with company announcements to manufacture a sector ranking.
What this means for enterprise leaders
TL;DR: Use a company example to identify a testable workflow. Then set your own baseline, owner, controls and acceptance rules.
Gartner's work-redesign finding is an association reported from a manager survey. It does not prove that redesign alone caused revenue gains. It does suggest a useful question: what work changes when the tool is introduced?
A proposed project should state the decision it supports, the work it changes and the evidence needed to continue. Customer support, engineering and research can each be useful starting points. None is inherently a better investment without a measured baseline.
Questions about company AI adoption
Related questions prospective buyers typically ask when researching enterprise AI adoption.
How many Fortune 500 companies use AI in 2026?
The sources on this page do not establish a current Fortune 500 count. A general enterprise survey is not a census of that list. Define whether use means a trial, a live workflow or organization-wide adoption before making a comparison.
Which industry has the highest AI adoption?
These sources do not support a single comparable sector ranking. They use different samples, questions and definitions of AI. The directory instead identifies named applications by industry. Use a study with a shared survey method before comparing adoption rates across sectors.
What is the largest AI deployment by users?
The directory does not establish a largest deployment. Employee access, active users, advisor teams and interactions are different units. Bank of America's interaction totals, for example, are not unique-user counts. A defensible ranking needs the same measure and period for every company.
How much does enterprise AI cost?
Cost depends on the workflow, data, integration, controls and operating model. Build a budget for implementation, software, infrastructure, evaluation and human review. A total technology budget or global spending forecast cannot price an individual deployment. Record both setup and ongoing costs.
What is the ROI of enterprise AI?
ROI compares financial benefits with implementation and operating costs over a defined period. The surveys here report use, perceived benefits or revenue-goal attainment. Those measures are not a shared ROI estimate. Test your own baseline, costs and benefits before transferring a company result.
Which AI tools do Fortune 500 companies use?
Examples in this directory include Microsoft 365 Copilot, GitHub Copilot and company-specific assistants such as Aladdin Copilot. Product availability does not prove use by every Fortune 500 company. For a named organization, verify the actual tool, deployment date and workflow in its disclosure.
Has AI adoption growth slowed in 2026?
The historical OECD series on this page ends in 2025. It shows rising reported firm use over the covered years. That does not establish the pace of growth in 2026. A current conclusion needs comparable observations with the same population and survey method.
What is the difference between AI adoption and AI in production?
Adoption means use as defined by a source and may include limited activity. Production means a system operates in a live workflow. For an investment decision, also check ownership, controls, support and repeated use. A production label alone does not establish financial value.
Which companies are leaders in agentic AI?
Salesforce offers Agentforce, while Cognition names Goldman Sachs and Citi as financial-services partners. These are sourced examples, not a comparable market-wide leadership ranking. Examine what an agent can actually do, which actions need approval and how the organization measures results.
What is the gap between AI leaders and laggards?
There is no single supported multiplier in the sources collected here. Company examples and separate industry surveys cannot establish one. Define the dimension first: active use, deployment breadth, quality or financial return. Then compare organizations using the same period and measurement rules.
Frequently asked questions
Which companies use AI the most in 2026?
The directory includes JPMorgan Chase, Microsoft, Google, BlackRock, Bank of America, Walmart and Morgan Stanley. Their applications and disclosures differ. Product availability, interactions and team adoption do not establish a common ranking. Use the company sources to compare a specific workflow.
How many companies actually use AI in 2026?
There is no global company count established here. McKinsey's 2025 survey reports regular AI use, while OECD publishes a different firm-use series. Neither source is a 2026 company census. Keep the source population, year and definition attached to each percentage.
Which industry uses AI the most?
The named examples span finance, retail, pharma, insurance, industry and technology. They do not establish sector-wide prevalence. To rank industries, use one study with consistent definitions and representative sector samples. A technology product launch cannot be compared directly with a bank employee-adoption measure.
How much do companies spend on AI?
AI budgets vary with the workflow and its operating requirements. Separate direct AI spending from wider technology or transformation programs. Avoid adding potentially overlapping budgets. For a useful comparison, use the same cost categories, period and deployment scope across companies.
What is the ROI of AI for large enterprises?
The cases do not establish a common return for large enterprises. Some measure activity; others describe a product or self-reported benefit. Calculate return using measured benefits and full costs. Include integration, data work, security, review effort and operations over the same period.
What is agentic AI and which companies use it?
Agentic AI can plan and execute steps through tools. Salesforce's Agentforce and Cognition's financial-services partnerships illustrate this category. The permitted extent of autonomy varies by individual deployment. The enterprise AI agent research adds context, but does not replace workflow-specific evidence.
Company AI sources and methods
This directory includes company-reported internal use, customer services and AI-enabled products. Historical deployments and current product pages are labeled by their sources. A listed application does not establish organization-wide adoption, current feature availability or independently audited financial results.
Surveys retain their own population, period and measure. They are not combined into a ranking. The Proof Standard explains the baseline and outcome checks needed before treating a claim as investment evidence.
Primary sources
- McKinsey. The State of AI in 2025. Published November 5, 2025. The regular-use finding describes survey respondents' organizations.
- NVIDIA. 2026 financial-services survey findings. More than 800 professionals; the benefit measure is self-reported.
- Gartner. Work-redesign survey findings. Published October 27, 2025. July 2025 survey of 1,973 managers; the comparison concerns business units.
- OECD. Artificial intelligence: firm-use trends. Published series for 2023, 2024 and 2025; different coverage and methods from McKinsey's survey.
- Finance disclosures: JPMorgan Chase: COiN; Bank of America: Erica; Goldman Sachs: employee tools; Morgan Stanley: Assistant and Debrief; BlackRock: Aladdin Copilot; Mastercard: Decision Intelligence Pro; Citi: Assist and Stylus; Wells Fargo: Fargo.
- Retail applications: Walmart: Sparky launch; Amazon: Rufus; Instacart CMO on Ask Instacart, Think with Google interview (May 2024); Shopify: Magic and Sidekick; Wayfair: Muse launch.
- Research and employee tools: Pfizer: clinical development; Moderna: mChat; AstraZeneca: research AI; Novartis: research AI; Roche: drug discovery; GSK: AI-supported discovery.
- Insurance applications: Lemonade: claims assistance; Progressive: AI use and fraud prevention; Allstate: April 2026 presentation, page 10; Munich Re: AI-assisted underwriting; Swiss Re: AI applications.
- Industrial applications and products: Siemens: Industrial Copilot; GE Aerospace: blade inspection; Caterpillar: Cat AI Assistant; Honeywell: control-room assistant; ABB: Genix Copilot launch.
- AI-enabled software products: Microsoft: Microsoft 365 Copilot; Google: Gemini; Meta: Meta AI; Salesforce: Agentforce; ServiceNow: Now Assist; GitHub: Copilot; Datadog: Bits AI.
- Linked case-study foundations: Zillow: Offers wind-down disclosure; Insilico: original phase 2a trial; Klarna: company-issued assistant announcement.
- Cognition. Financial-services partnerships. January 28, 2026. Supplier disclosure naming Goldman Sachs and Citi; not an independent performance comparison.
Cite the company AI research
This page is published under CC BY 4.0. Free to cite with attribution.
Citation
Okhrem, P. (2026). Companies That Use AI: 2026 Industry Reference. paul-okhrem.com/companies-using-ai/
Embed a stat with attribution
McKinsey's 2025 survey found that 88% of respondents said their organizations regularly used AI in at least one function. Source: McKinsey; compiled by Okhrem (2026), paul-okhrem.com/companies-using-ai/.
Retain the original source, year, population and measure when citing a figure.
Who compiled this company AI reference
Compiled by Paul Okhrem, AI transformation consultant and fractional Chief AI Officer. Co-Founder and CEO of Elogic Commerce; Managing Partner at Uvik Software since June 2017. Paul compiles this research as part of his fractional CAIO work. He advises CEOs and boards on AI and agentic AI decisions.
Facts · Press · Editorial standards
Suggested citation: Okhrem, P. (2026). Companies using AI in 2026: examples by industry. paul-okhrem.com/companies-using-ai/. Free to cite under CC BY 4.0.
Use the cases on this page to examine the process and ownership changes that a deployment would need. For that work, see AI transformation consulting.
Change log
- : Reviewed service scope, decision guidance and supporting resources for the October release.
- : Reviewed service scope, source evidence, links and buyer answers for the October release.
- : Clarified dense passages while preserving source limits and commercial terms.
Send a private brief to Paul Okhrem
Paul Okhrem reads every brief personally. Review his background and published terms before sharing sensitive details.