Verified
Accenture
Piloting- Professional services
- 814,391 employees
- Copilot rolling out to ~743,000
- ~110,000 AI and data staff
- AI Refinery, built in-house
Accenture deploys AI through platforms it built, AI Refinery on NVIDIA and Reinvention.AI for delivery, and through bought tools: Microsoft 365 Copilot, which CIO Tony Leraris is rolling out to about 743,000 employees, plus ChatGPT Enterprise and Claude Code. Chief AI and Data Officer Lan Guan runs the AI and Data Reinvention Engine. Accenture reported 89% monthly active Copilot use in a tranche of about 200,000 licenses (Apr 2026) but no business return it credits to AI alone.
accenture.comTheir approach to AI
- The platform
- Accenture built AI Refinery on NVIDIA AI Foundry in Jul 2024 to run agents and custom models, and says it has put almost $1 billion over 4 years into Reinvention.AI, its platform for agents and automation in delivery (Oct 2026).
- The models
- Accenture's staff use Microsoft 365 Copilot, which draws on OpenAI and Anthropic models, and Accenture agreed to give them ChatGPT Enterprise, Claude Code and Mistral AI Studio; Accenture builds on open-weight Llama and NVIDIA Nemotron models inside its own platforms (Jul 2024 to Oct 2026).
- The reach
- Accenture and Microsoft said on Apr 27, 2026 that Microsoft 365 Copilot is rolling out to about 743,000 Accenture employees, after monthly active use reached 89% in a tranche of about 200,000 licenses; more than 550,000 Accenture people had completed generative AI fundamentals training by Aug 2025.
- The catch
- Accenture has published no return it credits to AI alone: Copilot users' gains are self-reported speedups with no cost figure, revenue per person rose about 5% in fiscal 2026 partly from delivery efficiencies with AI's share not separated, and Accenture stopped reporting advanced AI revenue in Dec 2025.
Snapshot
- Employees
- ~814,000
- Aug 31, 2026; most in India, the Philippines and the U.S. (FY26 Form 10-K)
- R&D and learning spend
- $2.0B (FY26)
- $1.0B in R&D and $1.0B in learning and development; no internal IT budget published
- AI/ML headcount
- ~110,000
- AI and data professionals, Aug 31, 2026; Faculty added 400+ AI professionals in Mar 2026
- Who owns AI
- Lan Guan, Chief AI and Data Officer
- Leads the AI and Data Reinvention Engine inside Reinvention Services from Mar 31, 2026; CTO Marc Warner shapes technology strategy; CIO Tony Leraris runs enterprise technology
- Platform
- AI Refinery and Reinvention.AI
- Both built in-house; AI Refinery launched Jul 2024 on NVIDIA AI Foundry
- Models
- OpenAI, Anthropic, Mistral, open-weight
- Microsoft 365 Copilot, ChatGPT Enterprise, Claude Code and Mistral AI Studio for staff; Llama and Nemotron in Accenture platforms
- Reach
- ~743,000 Copilot rollout
- Microsoft 365 Copilot, 89% monthly active in a tranche of about 200,000 licenses (Apr 2026); 550,000+ trained in generative AI (Aug 2025)
- AI projects
- 18,000+
- Advanced AI projects for clients since Nov 2022 (Oct 2026); 3,000+ reusable agents (Dec 2025)
Key findings
Accenture had nearly 110,000 AI and data professionals at Aug 31, 2026, out of about 814,000 employees, up from 40,000 when it announced a $3 billion, 3-year AI investment in Jun 2023, and its people completed 46 million training hours in fiscal 2026.
Accenture and Microsoft said on Apr 27, 2026 that Microsoft 365 Copilot is rolling out to about 743,000 Accenture employees, the largest enterprise Copilot deployment to date according to Microsoft, after monthly active use reached 89% in a tranche of about 200,000 licenses; Accenture also agreed in Dec 2025 to give tens of thousands of professionals ChatGPT Enterprise and tens of thousands of developers Claude Code.
CFO Angie Park said on Oct 14, 2026 that AI is making Accenture's delivery more productive and that revenue per person rose about 5% in fiscal 2026, in part from efficiencies in delivery, without separating AI's share; Accenture will pass some AI productivity to clients, reinvest most of what it keeps and expand operating margin 10 to 30 basis points a year.
Accenture reported $4.8 billion of advanced AI revenue and $11.5 billion of bookings across about 11,000 projects from fiscal Q3 2023 through Nov 2025, then stopped publishing the figures in Dec 2025, saying advanced AI was now embedded in nearly all of its work.
Accenture made use of AI tools part of performance evaluations from fiscal 2026 and recorded $615 million in Q4 fiscal 2025 and $308 million in fiscal 2026 for a business optimization program covering severance, including for people whose reskilling was not a viable path, and the divestiture of 2 acquisitions.
Where Accenture sits
- Stage
- Piloting
Accenture has reported no business return it credits to AI alone in the last 18 months, for any internal use case or for the firm: revenue per person rose about 5% in fiscal 2026, which CFO Angie Park credited in part to efficiencies in delivery without separating AI's share (Oct 2026); corporate function costs fell more than 200 basis points of revenue over 4 years, with no AI share given (Oct 2026); agentic recruiting cut time to fill by 28% with no cost figure (Oct 2026); and Microsoft 365 Copilot users' self-reported speed and productivity gains (Apr 2026) carry no cost or headcount equivalent. The $4.8 billion of advanced AI revenue Accenture reported through Nov 2025 came from selling AI to clients, not from its own use. Accenture has reported a company-wide rollout with a user count: Accenture and Microsoft said in a joint statement on Apr 27, 2026 that Microsoft 365 Copilot is rolling out to about 743,000 Accenture employees, after monthly active use reached 89% in a tranche of about 200,000 licenses, a rollout led by CIO Tony Leraris. Accenture is therefore at Piloting; returns it credits to AI in 2 or more functions, or a firm-level figure it credits to AI, would move it to Scaling.
How the stages are defined - Ownership
- Chief AI and Data Officer Lan Guan, CTO Marc Warner and CIO Tony Leraris, all on the Global Management Committee
- Platform
- Hybrid: AI Refinery and Reinvention.AI built in-house; Copilot and other staff tools bought
- Models
- OpenAI and Anthropic, also through Microsoft 365 Copilot; Mistral; open-weight Llama and Nemotron
- Hosting
- Microsoft 365 Copilot in Microsoft 365 over Accenture's SharePoint and OneDrive data; AI Refinery on NVIDIA AI Foundry, sold as a service on AWS; other internal AI hosting not disclosed
- Reach
- Copilot rolling out to ~743,000; 89% monthly active in a ~200,000-license tranche (Apr 2026)
How they deploy
Run the company on what it sells
Julie Sweet said in Mar 2026 that Accenture deploys on itself all the services it gives clients and is often the experimentation place, and Kate Clifford told investors in Oct 2026 that Accenture is its own best credential. Accenture said in its fiscal 2025 value report that it is reinventing what it sells, how it delivers, how it partners and how it operates with generative, agentic and physical AI.
Built platforms, bought models
Accenture builds its own AI platforms, AI Refinery for agents and custom models and Reinvention.AI for automation in delivery, and buys frontier models and staff tools from Microsoft, OpenAI, Anthropic and Mistral AI; Microsoft 365 Copilot, rolling out to about 743,000 employees (Apr 2026), is the most widely deployed. Julie Sweet said on Oct 1, 2026 that Accenture is the number one partner of the top 3 frontier model companies and now offers full lifecycle services for open-weight models.
The right model for each workflow
Lan Guan said in Oct 2026 that Accenture's own tokenomics research found only 10% to 20% of enterprise workflows justify frontier model capabilities, and that Accenture helps companies pick the right model for each piece of work and run open-weight models in their own secure environments. Julie Sweet said in Jun 2026 that Accenture had built a platform internally to optimize its own token use and track which models suit which problems, and was taking it to clients. Accenture AI Refinery includes a switchboard that selects among models on factors such as cost or accuracy.
Productivity shared 3 ways
Angie Park said in Oct 2026 that Accenture will use AI productivity in 3 ways: passing value to clients, reinvesting the majority of what it retains into growth, and expanding operating margin every year. Julie Sweet said on Oct 1, 2026 that Accenture believes the opportunities from AI are greater than the AI efficiencies it absorbs, and that it gives clients more productivity because of AI while offsetting that with new kinds of work.
Reskill, rehire and buy talent
Accenture spends about $1 billion a year on learning, built a 14-week program that turns software engineers into certified forward deployed engineers, and is re-architecting every job around data and AI, industry, functional and technology skills, Kate Clifford said in Oct 2026. Accenture also buys AI talent: it acquired Faculty, a UK applied AI company with more than 400 professionals, in Mar 2026, and made its CEO Marc Warner Accenture's CTO.
- Deployment model
- Accenture runs AI through Reinvention Services, the unit it created on Sep 1, 2025, the start of fiscal 2026, to bring strategy, consulting, Song, technology and operations together, led by Chief Strategy and Reinvention Services Officer Manish Sharma. Since Mar 31, 2026 Reinvention Services has had 7 client-facing Reinvention Partners and 3 Reinvention Engines, which Accenture describes as where it reinvents itself: AI and Data under Chief AI and Data Officer Lan Guan, Industry and Process under Jason Dess, and Technology under Rajendra Prasad, whom Accenture named chief information and asset engineering officer, responsible for its internal technology, in Jul 2023 and Group Chief Executive, Technology and CTO, still overseeing global IT operations, from Sep 1, 2025. CTO Marc Warner, who joined with Faculty in Mar 2026, shapes Accenture's technology strategy, and Chief Information Officer Tony Leraris, also on the Global Management Committee, is responsible for Accenture's enterprise technology and led the Microsoft 365 Copilot rollout to staff. Corporate functions deploy AI on Accenture itself: Kate Clifford's HR team runs agentic recruiting, Angie Park said Accenture is rebuilding contract to cash, which spans legal, sales and finance, with agents, and COO Kate Hogan is building tech-, data- and AI-enabled corporate growth services, according to her Accenture biography.
- Model providers
- Accenture's most widely deployed staff tool is Microsoft 365 Copilot, rolling out to about 743,000 employees (Apr 2026); CIO Tony Leraris said Accenture chose it in part because it draws on both OpenAI and Anthropic models and sits inside the Microsoft 365 tools employees already use. Accenture agreed in Dec 2025 to give tens of thousands of its professionals in consulting, operations and delivery OpenAI's ChatGPT Enterprise, and to make Anthropic's Claude Code available to tens of thousands of its developers, which Anthropic called its largest ever deployment. Accenture agreed in Feb 2026 to become a Mistral AI customer and equip its professionals with Mistral models and Mistral AI Studio. Accenture AI Refinery was launched in Jul 2024 on NVIDIA AI Foundry with Meta's openly available Llama 3.1 models, Accenture said in Aug 2026 that it had been an early adopter of NVIDIA's Nemotron 3 models across its developer communities, and Accenture AI Factory uses NVIDIA Nemotron open-weight models for agentic operations (Oct 2026). Google DeepMind gives Accenture early access to Gemini models for client work under an Apr 2026 program.
- Data and infrastructure
- Accenture AI Refinery combines model customization on a company's own data, a switchboard that picks models, an enterprise cognitive brain that indexes corporate knowledge, and an agentic architecture, and Accenture said in Oct 2024 that AI Refinery would be available on all public and private cloud platforms; Accenture sells AI Refinery as a service deployed on AWS through AWS Marketplace. Accenture said in Aug 2026 that it was piloting the secure runtime of NVIDIA OpenShell, with access controls and policies, for its own production-ready personal assistant agents. Accenture AI Foundation and AI Factory, expanded for clients in Oct 2026, run agents inside NVIDIA OpenShell sandboxes on Accenture-managed Kubernetes infrastructure with signed agent bundles, deny-by-default sandboxes and central audit, and support fully air-gapped deployments. Accenture's Microsoft 365 Copilot reasons over the 24 petabytes of data Accenture keeps in SharePoint and OneDrive, and Microsoft 365's privacy and security controls let Accenture test new Copilot features with small groups and turn off capabilities that conflict with local regulations, CIO Tony Leraris said in Apr 2026. Accenture has not disclosed which clouds or data centers host its other internal AI workloads or its ChatGPT Enterprise and Claude deployments; its Accenture Dell Business Group, formed in Oct 2026, targets private and on-premises AI for clients.
Platforms and internal tools
AI RefineryJul 2024
AI Refinery is Accenture's platform for building custom models and agents, built on NVIDIA AI Foundry, with an agent builder, a model switchboard and an enterprise knowledge index, and supported by AI Refinery engineering hubs in Accenture's Center for Advanced AI. Accenture said in Jul 2024 that it was using AI Refinery to reinvent its own enterprise functions, starting with marketing and communications, and in Oct 2024 that its marketing function was running autonomous agents built on AI Refinery.
- Models
- Meta Llama 3.1 via NVIDIA AI Foundry; other models through the switchboard
Reinvention.AI
Reinvention.AI is Accenture's proprietary platform for agents and automation in managed services, into which Manish Sharma said Accenture has invested almost $1 billion over 4 years (Oct 2026). Julie Sweet said clients turn to Accenture's managed services in part because they do not want to build such an agent platform themselves.
GenWizard
GenWizard is one of Accenture's proprietary delivery platforms, into which Accenture said in its fiscal 2025 value report it had already embedded advanced AI so that it delivers work differently for clients.
SynOps
SynOps is Accenture's AI-powered operations platform combining AI, data and people, developed under Manish Sharma when he led Accenture Operations, and named in Jun 2023 alongside myWizard and MyNav as one of the platforms through which Accenture embeds AI in service delivery.
Accenture TokenomicsOffered to clients Jul 2026
Accenture Tokenomics grew from a platform Accenture built internally to optimize its own token use, which Julie Sweet said in Jun 2026 Accenture was now taking to clients. The client offering, launched on Jul 29, 2026, includes a model engine that routes work across managed and open-weight models and a command center that reports cost per successful business action.
AI Foundation and AI FactoryExpanded Oct 2026
Accenture AI Foundation and AI Factory are Accenture's offerings for running enterprise agents in production, combining NVIDIA AI infrastructure and software with Accenture orchestration, governance and agentic operations and running agents in NVIDIA OpenShell sandboxes on Accenture-managed Kubernetes with central audit and fleet operations.
- Models
- NVIDIA Nemotron open-weight models
Microsoft 365 CopilotPilot Aug 2023
Microsoft 365 Copilot is the bought assistant in Outlook, Teams, Word and other Microsoft 365 apps that Accenture began piloting with a few hundred senior leaders in Aug 2023, scaled to 20,000 users and then to a tranche of about 200,000 licenses, and is rolling out to about 743,000 employees, Accenture and Microsoft said on Apr 27, 2026. CIO Tony Leraris leads the rollout, which Microsoft called the largest enterprise Copilot deployment to date.
- Users
- About 743,000 being rolled out; 89% monthly active in a ~200,000-license tranche (Apr 2026)
- Models
- OpenAI and Anthropic, through Microsoft
ChatGPT EnterpriseDec 2025
OpenAI's ChatGPT Enterprise is a bought tool that Accenture agreed in Dec 2025 to give tens of thousands of its professionals for internal and client work in consulting, operations and delivery, alongside OpenAI certifications; OpenAI said Accenture was putting ChatGPT in their hands.
- Users
- Tens of thousands planned (Dec 2025)
- Models
- OpenAI
Claude CodeDec 2025
Anthropic's Claude Code is the coding agent Accenture said in Dec 2025 it would make available to tens of thousands of its developers, when about 30,000 Accenture professionals were to be trained on Claude.
- Users
- Tens of thousands of developers planned (Dec 2025)
- Models
- Anthropic Claude
Vendors and partners
Microsoft
Microsoft 365 Copilot for Accenture's workforce; Avanade joint venture · Accenture began piloting Microsoft 365 Copilot in Aug 2023, and Accenture and Microsoft said in a joint statement on Apr 27, 2026 that Copilot is rolling out to about 743,000 Accenture employees.
OpenAIDec 2025
ChatGPT Enterprise for Accenture staff; one of Accenture's primary AI partners for client services · OpenAI and Accenture launched a flagship AI client program in Dec 2025, and Accenture said it would have the largest number of professionals upskilled through OpenAI certifications.
AnthropicDec 2025
Claude Code for Accenture developers; Accenture Anthropic Business Group; AI safety evaluators · Accenture and Anthropic agreed in Sep 2026 to build a team of embedded evaluators to red-team Anthropic models, and each expects to invest at least $1 billion in AI safety over 5 years.
NVIDIAJul 2024
Infrastructure and software under AI Refinery, AI Foundation and AI Factory; OpenShell runtime piloted for Accenture's own assistant agents · Accenture formed an NVIDIA Business Group in Oct 2024 with more than 30,000 professionals to be trained, and expanded AI Foundation and AI Factory on NVIDIA technology in Oct 2026.
Mistral AIFeb 2026
Model provider for Accenture staff and operations; co-developed client solutions · Accenture became a Mistral AI customer in Feb 2026, equipping professionals with Mistral models and Mistral AI Studio.
Google Cloud and Google DeepMindApr 2026
Gemini Enterprise business group; early access to Gemini models · Accenture formed the Accenture Gemini Enterprise Business Group in Sep 2026, with nearly 50,000 Google Cloud-skilled professionals and a planned 1,000-person forward deployed engineering workforce.
Workday
HR, finance and IT platform whose agentic capabilities Accenture deploys on itself · Accenture and Workday launched a joint forward deployed engineering program in Oct 2026 that draws on Accenture's own enterprise-scale deployment of Workday's agentic capabilities.
MetaJul 2024
Llama open models behind AI Refinery's custom model service
Dell TechnologiesOct 2026
Private and on-premises AI infrastructure through the Accenture Dell Business Group · Accenture plans to train more than 3,000 practitioners on the Dell portfolio.
Stanford Institute for Human-Centered AI
Co-created Accenture's agentic AI fundamentals course · 192,000 Accenture people had completed the course by Mar 2026.
Leadership
AI, data and platform leaders
Executives
AI org map
AI and Data Reinvention Engine inside Reinvention Services, a CTO from Faculty, and agents deployed by corporate functions
Accenture combined strategy, consulting, Song, technology and operations into Reinvention Services on Sep 1, 2025, which Manish Sharma leads, and from Mar 31, 2026 split it into 7 client-facing Reinvention Partners and 3 Reinvention Engines, with Lan Guan running the AI and Data engine as Chief AI and Data Officer. Accenture added a technology chief from outside when it bought Faculty in Mar 2026 and made Faculty CEO Marc Warner its CTO, while CIO Tony Leraris runs Accenture's enterprise technology and its Microsoft 365 Copilot rollout. Corporate functions under COO Kate Hogan, CFO Angie Park and HR chief Kate Clifford deploy AI on Accenture's own operations, and the board's Audit Committee oversees AI risk.
Accenture had nearly 110,000 AI and data professionals at Aug 31, 2026, up from about 40,000 in Jun 2023, 57,000 AI practitioners in Oct 2024, about 77,000 at the end of fiscal 2025 and more than 85,000 in Mar 2026. Accenture reached the total through hiring, acquisitions such as Faculty, which added more than 400 AI professionals in Mar 2026, and training, including partner-specific programs for about 30,000 professionals each on Anthropic's Claude and on NVIDIA technology. Lan Guan's Accenture biography credits her with a practice of nearly 80,000 AI and data experts, and in Oct 2026 Accenture formed the Advanced Intelligence Group, more than 2,000 specialists such as forward deployed engineers inside that practice; Accenture does not publish where the rest of the AI and data workforce sits.
Reporting lines
Chair and CEO
Reinvention Services
Created at the start of fiscal 2026; 7 Reinvention Partners and 3 Reinvention Engines from Mar 31, 2026
AI and Data Reinvention Engine
Lan Guan, Chief AI and Data Officer
Builds AI and data capabilities and training across Accenture with the Center for Advanced AI; nearly 80,000 AI and data experts; Advanced Intelligence Group of 2,000+ specialists (Oct 2026)
Technology Reinvention Engine
Prasad was named chief information and asset engineering officer, responsible for internal technology, in Jul 2023, and kept oversight of global IT operations as Group Chief Executive, Technology and CTO from Sep 2025
Industry and Process Reinvention Engine
Develops AI-enabled methods, including AI-enabled ERP methods
Office of the CTO and Faculty
Reporting line not published; Warner sits on the Global Management Committee and Faculty is one of Accenture's network businesses
Corporate operations
Kate Hogan, COO
Leadership and Human Resources
Kate Clifford
Agentic recruiting and AI skilling
Enterprise technology
Tony Leraris, CIO
Reporting line not published; Leraris sits on the Global Management Committee and leads the Microsoft 365 Copilot rollout
Board Audit Committee
Oversees AI risk exposures and receives reports on the Responsible AI compliance program at least annually
Accenture does not publish an organization chart. The lines to the Chair and CEO are inferred from the executive officer list in the fiscal 2026 Form 10-K, the Reinvention Engines' place under Manish Sharma comes from Accenture's Mar 13, 2026 announcement, and Accenture has not said to whom CTO Marc Warner or CIO Tony Leraris reports. Tony Leraris's undated Accenture biography makes him responsible for the enterprise technology that Rajendra Prasad oversaw as chief information and asset engineering officer from Jul 2023; Accenture has not said when Leraris took the role.
Where the AI workforce sits
AI and data practice under Lan Guan
Estimate: Lan Guan's undated Accenture biography cites nearly 80,000 AI and data experts, against nearly 110,000 AI and data professionals firmwide at Aug 31, 2026.
~70%Faculty
More than 400 applied AI professionals joined in Mar 2026.
<1%AI and data professionals elsewhere in Accenture
Estimate: the remainder, across industry and functional teams, Song, operations and network businesses; Accenture does not publish the split.
~25-30%
Named people
Use cases
Accenture describes AI inside the company in 3 places: service delivery, where its Reinvention.AI and GenWizard platforms and a library of more than 3,000 reusable agents support client work; corporate functions such as recruiting, Workday agents, contract to cash, marketing and workforce planning; and staff tools such as Microsoft 365 Copilot, which is rolling out to about 743,000 employees (Apr 2026), ChatGPT Enterprise, Claude Code and personal assistant agents in pilot. Accenture has not published a business return credited to AI alone for any of these use cases; the Copilot gains it reported are users' own estimates of speed and productivity. At the firm level, revenue per person rose about 5% in fiscal 2026, which Accenture credits only in part to efficiencies in delivery, and Accenture stopped reporting advanced AI revenue from clients after $4.8 billion through Nov 2025.
| Use case | Area | Status | ROI |
|---|---|---|---|
| AI in delivery: Reinvention.AI, GenWizard and reusable agents Accenture embeds advanced AI and agents in the platforms its teams use to deliver consulting and managed services, including Reinvention.AI and GenWizard, and had built a library of more than 3,000 reusable agents used in client environments by Dec 2025. Angie Park said in Oct 2026 that AI is making delivery more productive and that revenue per person rose about 5% in fiscal 2026, in part from efficiencies in how Accenture delivers, without separating AI's share. | Service delivery | In production | |
| Agentic recruiting Agents schedule all interviews, screen resumes and run skills assessments in Accenture's recruiting, while a hiring manager always makes the hiring decision; Kate Clifford said in Oct 2026 that time to fill fell 28% and candidate quality rose 30%, and that entry-level hires now become productive in 6 weeks instead of 6 months. | Human resources | In production | |
| Workday agents Accenture deploys and extends Workday's agentic capabilities on itself at enterprise scale, experience it said in Oct 2026 would feed a joint forward deployed engineering program with Workday that puts Workday agents to work in clients' HR and finance. | Corporate functions | In production | |
| Agentic contract to cash Accenture is rebuilding contract to cash with agents from contract review to bookings, billing and collections, a process spanning legal, sales and finance that touches 9,000 clients, 60,000 contracts and 900,000 invoices. | Finance | Announced | |
| Campaign agents on AI Refinery Accenture's marketing function integrated AI Refinery with autonomous agents to create and run campaigns; Accenture said in Oct 2024 that this would cut manual steps 25-35% and costs 6%, and has not published a measured result since. | Marketing | In production | |
| Microsoft 365 Copilot for the workforce Accenture people use Microsoft 365 Copilot to draft, research and analyze in Outlook, Teams and Word, and marketing writers use it to check new content against past material; Accenture and Microsoft said on Apr 27, 2026 that Copilot is rolling out to about 743,000 employees. In 2025 company data on about 200,000 users, 97% reported finishing routine tasks up to 15 times faster and 53% significant productivity gains; Accenture has given no cost or capacity figure. | Firmwide | In production | |
| ChatGPT Enterprise for consulting, operations and delivery Accenture agreed in Dec 2025 to give tens of thousands of its professionals ChatGPT Enterprise for internal and client work, with OpenAI certifications to build skills; OpenAI said Accenture was putting ChatGPT in their hands. | Firmwide | In production | |
| Claude Code for developers Accenture said in Dec 2025 that it would make Anthropic's Claude Code available to tens of thousands of its developers; Accenture has not reported how many use it. | Technology | Announced | |
| Personal assistant agents on NVIDIA OpenShell Accenture said in Aug 2026 that it was piloting the secure runtime of NVIDIA OpenShell for its own production-ready personal assistant agents, setting access controls and policies before widening their use. | Firmwide | Pilot | |
| Token and model routing Accenture built a platform internally to optimize its token use and choose which models to use for which problems, which Julie Sweet said in Jun 2026 Accenture was taking to clients as Accenture Tokenomics. | Technology | In production | |
| Skills matching and workforce planning An AI-powered skills engine tracks the skills of Accenture's people, and a Talent Orchestration dashboard supports workforce planning across demand, skills, revenue and payroll; Angie Park said in Oct 2026 that Accenture is deploying advanced AI to match skills and capacity with demand and to shorten recruiting times. | Talent and staffing | In production | |
| Coaching conversation simulator Accenture leaders practice coaching conversations with their teams on a generative AI conversation simulator, according to Accenture's fiscal 2025 value report. | Leadership development | In production |
Adoption
Accenture spread AI first through training: more than 550,000 people had completed generative AI fundamentals by Aug 31, 2025, 192,000 had completed an agentic AI fundamentals course co-created with Stanford by Mar 2026, and Accenture people completed 46 million training hours in fiscal 2026. Accenture's broadest AI tool is Microsoft 365 Copilot: a pilot with a few hundred senior leaders in Aug 2023 grew to 20,000 users and then to a tranche of about 200,000 licenses, where monthly active use reached 89%, and Accenture and Microsoft said on Apr 27, 2026 that Copilot is rolling out to about 743,000 employees under CIO Tony Leraris. Accenture also agreed in Dec 2025 to give ChatGPT Enterprise to tens of thousands of professionals and Claude Code to tens of thousands of developers, and from fiscal 2026 made use of AI tools part of performance evaluations; the Financial Times reported in Feb 2026 that Accenture was collecting weekly logins to its AI tools, including AI Refinery, for some senior staff as an input to promotions. Accenture has not published how many employees use Copilot each week since the wider rollout, or how many use ChatGPT Enterprise and Claude Code, and its people held more than 134,000 unique certifications across AI and data ecosystem partners in Oct 2026.
By the numbers
Rollout, first pilot to today
2022
Accenture implemented an enterprise-wide responsible AI program that tracks where AI is used inside the company, assesses each system's risk and monitors systems after deployment.
Jun 2023
Accenture announced a $3 billion, 3-year investment in its Data & AI practice, a goal to double AI talent to 80,000, an AI Navigator for Enterprise platform and a Center for Advanced AI to reimagine service delivery.
Aug 2023
Accenture began piloting Microsoft 365 Copilot with a few hundred senior leaders and select employees, then scaled it to 20,000 users while it worked on data governance and access controls.
Jul 2024
Accenture launched AI Refinery on NVIDIA AI Foundry with Llama 3.1 models and began using it to reinvent its own enterprise functions, starting with marketing and communications.
Oct 2024
Accenture formed an NVIDIA Business Group, opened AI Refinery engineering hubs serving 57,000 AI practitioners, and said its marketing function was running autonomous agents built on AI Refinery.
Aug 2025
Accenture ended fiscal 2025 with about 77,000 AI and data professionals and more than 550,000 people who had completed training in generative AI fundamentals.
Sep 2025
Accenture reported $5.9 billion of generative AI bookings for fiscal 2025 and began a 6-month business optimization program that included exiting people on a compressed timeline where reskilling was not viable.
Dec 1, 2025
Accenture agreed to give tens of thousands of its professionals ChatGPT Enterprise for consulting, operations and delivery work.
Dec 9, 2025
Accenture said it would make Claude Code available to tens of thousands of its developers and train about 30,000 professionals on Claude.
Dec 18, 2025
Accenture reported advanced AI metrics for the last time: $4.8 billion of revenue and $11.5 billion of bookings across about 11,000 projects since fiscal Q3 2023, and a library of more than 3,000 reusable agents.
Feb 2026
Accenture began collecting weekly logins to its AI tools, including AI Refinery, for some senior staff ahead of its summer promotion round, exempting staff in 12 European countries and on U.S. federal contracts, the Financial Times reported.
Mar 16, 2026
Accenture completed its acquisition of Faculty, adding more than 400 AI professionals, and named Faculty CEO Marc Warner its CTO.
Mar 19, 2026
Accenture reported more than 85,000 AI and data professionals and 192,000 people through agentic AI fundamentals, and made use of AI tools a formal part of performance evaluations.
Mar 31, 2026
Accenture's new Reinvention Services design took effect, with an AI and Data Reinvention Engine under Lan Guan charged with AI training across Accenture and embedding AI across the enterprise.
Apr 27, 2026
Accenture and Microsoft said Microsoft 365 Copilot is rolling out to about 743,000 Accenture employees, after monthly active use reached 89% in a tranche of about 200,000 licenses.
Aug 2026
Accenture said it was piloting the secure runtime of NVIDIA OpenShell for its own production-ready personal assistant agents.
Oct 13, 2026
Accenture said it deploys and extends Workday's agentic capabilities on itself at enterprise scale, as it launched a joint forward deployed engineering program with Workday.
Oct 14, 2026
Accenture reported nearly 110,000 AI and data professionals and described agentic recruiting and contract-to-cash agents at its investor day.
Economics & controls
Spend and ROI
AI investment
Accenture announced a $3 billion, 3-year AI investment in Jun 2023, and Manish Sharma said in Oct 2026 that Accenture has invested almost $1 billion over 4 years in its Reinvention.AI platform. Accenture spent $1.0 billion on R&D, $1.0 billion on learning and development and $4.9 billion on 17 acquisitions, including Faculty, in fiscal 2026, and its key operating investments totaled $13.0 billion from fiscal 2023 to 2026. Accenture said in Sep 2026 that it expects to invest at least $1 billion over 5 years in AI safety, and Julie Sweet told investors in Oct 2026 that Accenture had created an AI safety business using Faculty. Accenture has not disclosed what its own AI tools, tokens or model licenses cost.
What AI earns Accenture
Accenture tripled its revenue from generative and agentic AI to $2.7 billion in fiscal 2025 and nearly doubled generative AI bookings to $5.9 billion, figures that exclude data, classical AI and AI used in delivering its services. Accenture reported $2.2 billion of advanced AI bookings and about $1.1 billion of revenue in the quarter to Nov 30, 2025, $4.8 billion of revenue and $11.5 billion of bookings from fiscal Q3 2023 to that date, and then stopped publishing the figures. In fiscal 2026, bookings with Accenture's 8 emerging AI and data partners more than tripled and revenue more than doubled.
Reported ROI
Accenture has not reported a business return that it credits to AI alone, for the firm or for any internal use case. Angie Park said on Oct 14, 2026 that AI is making Accenture's delivery more productive and that revenue per person rose about 5% in fiscal 2026, in part because of efficiencies in how Accenture delivers services; Julie Sweet said on Oct 1, 2026 that the rise was in part due to AI, and Manish Sharma said part of it came from higher-value work. Park called revenue per person one of the clearest indicators of how AI is changing the business and set a goal for it to rise again in fiscal 2027, although in Dec 2025 she said a 7% rise in Q1 fiscal 2026 was driven primarily by talent rotation. Accenture also reduced the cost of its corporate functions by more than 200 basis points of revenue over 4 years and reinvested the savings, but has not said how much of that came from AI. Accenture and Microsoft reported in Apr 2026 that, in 2025 company data on about 200,000 Microsoft 365 Copilot users, 97% reported finishing routine tasks up to 15 times faster and 53% significant productivity gains; the figures are self-reported and carry no cost or headcount equivalent.
Headcount
Accenture had 814,391 employees at Aug 31, 2026, up 5% in a year, after headcount fell from 790,692 to 779,273 in the quarter to Aug 31, 2025. Accenture recorded $615 million in Q4 fiscal 2025 and $308 million in fiscal 2026 for a business optimization program that covered severance, including for people exited on a compressed timeline where reskilling was not viable, and the divestiture of 2 acquisitions. Julie Sweet said on Oct 1, 2026 that Accenture will hire in fiscal 2027, including more entry-level people, but at a lower rate in part because of AI.
Pricing and AI efficiencies
Julie Sweet said on Oct 1, 2026 that Accenture is giving clients more productivity because of AI, that the impact has been steady and offset by new work, and that pricing was lower in many areas of the business in Q4 fiscal 2026. Accenture's 3-year plan, presented on Oct 14, 2026, targets at least mid-single-digit revenue growth in fiscal 2027 and 2028 while absorbing AI efficiencies, and Accenture's fiscal 2026 Form 10-K warns that AI-enabled automation has replaced and will continue to replace some services and tasks performed by its people or its clients' people.
Governance and risk
Board oversight
Accenture's board Audit Committee oversees AI risk exposures and receives reports on AI matters, including the Responsible AI compliance program, at least annually, and the full board reviews AI, including generative, agentic and physical AI, in its annual review of enterprise risk, according to Accenture's Dec 2025 proxy statement.
Responsible AI compliance program
Accenture's Responsible AI compliance program, which has CEO sponsorship, applies the same principles to Accenture's internal AI systems and to client work: human by design; fairness; transparency, explainability and accuracy; safety; accountability; compliance, data privacy and cybersecurity; and sustainability. Accenture said in May 2024 that the enterprise-wide program, in place since 2022, tracks where AI is used and for what, assesses each system's risk, applies mitigations and monitors systems after deployment, and Accenture named Arnab Chakraborty its first chief responsible AI officer that month. Accenture includes responsible AI in the ethics and compliance training required of all employees and gives deeper training to people who build and deploy AI.
Humans decide, agents are sandboxed
Kate Clifford said in Oct 2026 that in Accenture's agentic recruiting a hiring manager always makes the hiring decision. Accenture said in Aug 2026 that it was piloting NVIDIA OpenShell's secure runtime, with access controls and policies, for its own personal assistant agents. Accenture's agent platform for clients, AI Foundation, expanded in Oct 2026, runs agents in deny-by-default NVIDIA OpenShell sandboxes with IT-signed agent bundles, governs which tools agents may call and where inference is routed, and keeps a central audit trail.
AI use in performance reviews
Julie Sweet said on Mar 19, 2026 that after significant investment and training, Accenture had made the use of AI tools, and contributions to making Accenture the most AI-enabled company, a formal part of its performance evaluations from fiscal 2026. The Financial Times reported in Feb 2026 that Accenture had begun collecting weekly logins to its AI tools for some senior staff, that an internal email called use of key tools a visible input to talent discussions, and that staff in 12 European countries and on U.S. federal contracts were exempt.
AI safety
Accenture bought Faculty in part for its AI safety experience, and in Sep 2026 agreed with Anthropic to build a team of embedded evaluators to red-team models, conduct alignment assessments and test safeguards, with each company expecting to invest at least $1 billion in AI safety over 5 years. Julie Sweet said the arrangement is non-exclusive and that Accenture expects to work with several AI labs.
Disclosed risks
Accenture's fiscal 2026 Form 10-K says that using autonomous or semi-autonomous agents in delivery and internal functions may not deliver the expected productivity, may take longer or may need more oversight and human intervention than planned, that many of its AI services rely on third-party models it has limited control over, and that its AI safety services could draw regulatory scrutiny and litigation.
Recent signals
Oct 21, 2026
Accenture expanded AI Foundation and AI Factory with NVIDIA, adding secure agent execution in NVIDIA OpenShell, agentic operations on Nemotron open-weight models, and autonomous agents on Accenture-managed Kubernetes.
Oct 14, 2026
At its investor day Accenture said revenue per person rose about 5% in fiscal 2026, in part from efficiencies in delivery, agentic recruiting cut time to fill by 28%, it has invested almost $1 billion in Reinvention.AI over 4 years, and it will invest in a Faculty Lab.
Oct 13, 2026
Accenture and Workday launched a forward deployed engineering program for clients' HR and finance that draws on Accenture's own enterprise-scale deployment of Workday's agentic capabilities.
Oct 5, 2026
Accenture formed the Advanced Intelligence Group, more than 2,000 specialists such as forward deployed engineers and AI solutions engineers inside its AI and Data practice, in a post quoting Chief AI and Data Officer Lan Guan.
Oct 1, 2026
Accenture reported fiscal 2026 revenue of $74.2 billion, nearly 110,000 AI and data professionals and more than 400 clients starting advanced AI work in the year, and said it will hire at a lower rate in fiscal 2027 partly because of AI.
Sep 18, 2026
Accenture and Anthropic agreed to build a team of embedded evaluators to red-team Anthropic models, each expecting to invest at least $1 billion in AI safety over 5 years.
Aug 26, 2026
Accenture said it was piloting NVIDIA OpenShell's secure runtime for its own production-ready personal assistant agents and launched full lifecycle services for open-weight models.
Jun 18, 2026
Julie Sweet said Accenture had built a platform internally to optimize its token use and was taking it to clients; Accenture launched it as Accenture Tokenomics on Jul 29, 2026.
Apr 27, 2026
Accenture and Microsoft said in a joint statement that Microsoft 365 Copilot is rolling out to about 743,000 Accenture employees, and Julie Sweet said Accenture teams are already doing higher-value work because of it.
Mar 19, 2026
Julie Sweet said Accenture had more than 85,000 AI and data professionals, 192,000 people had completed agentic AI fundamentals, and use of AI tools is now a formal part of performance evaluations.
Mar 16, 2026
Accenture completed its acquisition of Faculty, adding more than 400 AI professionals, and Faculty CEO Marc Warner became Accenture's CTO.
Mar 13, 2026
Accenture named the leaders of its redesigned Reinvention Services, with Lan Guan heading an AI and Data Reinvention Engine charged with AI training across Accenture, effective Mar 31, 2026.
Feb 26, 2026
Accenture and Mistral AI began a multi-year collaboration under which Accenture became a Mistral customer, equipping its professionals with Mistral models and Mistral AI Studio.
Feb 2026
The Financial Times reported that Accenture had begun collecting weekly logins to its AI tools, including AI Refinery, for some senior staff, and that regular AI use would count in promotions to leadership roles.
Dec 18, 2025
Accenture reported $2.2 billion of advanced AI bookings and about $1.1 billion of revenue for the quarter, said it had built more than 3,000 reusable agents, and stopped publishing advanced AI metrics.
Quick answers
- How does Accenture deploy AI?
- Accenture deploys AI through platforms it built, AI Refinery on NVIDIA and Reinvention.AI for delivery, and through bought tools: Microsoft 365 Copilot, which CIO Tony Leraris is rolling out to about 743,000 employees, plus ChatGPT Enterprise and Claude Code. Chief AI and Data Officer Lan Guan runs the AI and Data Reinvention Engine. Accenture reported 89% monthly active Copilot use in a tranche of about 200,000 licenses (Apr 2026) but no business return it credits to AI alone.
- What does Accenture use AI for?
- Accenture describes AI inside the company in 3 places: service delivery, where its Reinvention.AI and GenWizard platforms and a library of more than 3,000 reusable agents support client work; corporate functions such as recruiting, Workday agents, contract to cash, marketing and workforce planning; and staff tools such as Microsoft 365 Copilot, which is rolling out to about 743,000 employees (Apr 2026), ChatGPT Enterprise, Claude Code and personal assistant agents in pilot. Accenture has not published a business return credited to AI alone for any of these use cases; the Copilot gains it reported are users' own estimates of speed and productivity. At the firm level, revenue per person rose about 5% in fiscal 2026, which Accenture credits only in part to efficiencies in delivery, and Accenture stopped reporting advanced AI revenue from clients after $4.8 billion through Nov 2025.
- Who leads AI at Accenture?
- AI at Accenture is led by Lan Guan (Chief AI and Data Officer; AI and Data Reinvention Engine Lead); Marc Warner (Chief Technology Officer; CEO, Faculty); Rajendra Prasad (Technology Reinvention Engine Lead). Ownership of AI on the org chart: Chief AI and Data Officer Lan Guan, CTO Marc Warner and CIO Tony Leraris, all on the Global Management Committee.
- What AI platform does Accenture use?
- Accenture built AI Refinery on NVIDIA AI Foundry in Jul 2024 to run agents and custom models, and says it has put almost $1 billion over 4 years into Reinvention.AI, its platform for agents and automation in delivery (Oct 2026).
- Which AI models does Accenture use?
- Accenture's staff use Microsoft 365 Copilot, which draws on OpenAI and Anthropic models, and Accenture agreed to give them ChatGPT Enterprise, Claude Code and Mistral AI Studio; Accenture builds on open-weight Llama and NVIDIA Nemotron models inside its own platforms (Jul 2024 to Oct 2026).
- How many Accenture employees use AI?
- Accenture and Microsoft said on Apr 27, 2026 that Microsoft 365 Copilot is rolling out to about 743,000 Accenture employees, after monthly active use reached 89% in a tranche of about 200,000 licenses; more than 550,000 Accenture people had completed generative AI fundamentals training by Aug 2025.
- What return has Accenture reported from AI?
- Accenture has reported no business return it credits to AI alone in the last 18 months, for any internal use case or for the firm: revenue per person rose about 5% in fiscal 2026, which CFO Angie Park credited in part to efficiencies in delivery without separating AI's share (Oct 2026); corporate function costs fell more than 200 basis points of revenue over 4 years, with no AI share given (Oct 2026); agentic recruiting cut time to fill by 28% with no cost figure (Oct 2026); and Microsoft 365 Copilot users' self-reported speed and productivity gains (Apr 2026) carry no cost or headcount equivalent. The $4.8 billion of advanced AI revenue Accenture reported through Nov 2025 came from selling AI to clients, not from its own use. Accenture has reported a company-wide rollout with a user count: Accenture and Microsoft said in a joint statement on Apr 27, 2026 that Microsoft 365 Copilot is rolling out to about 743,000 Accenture employees, after monthly active use reached 89% in a tranche of about 200,000 licenses, a rollout led by CIO Tony Leraris. Accenture is therefore at Piloting; returns it credits to AI in 2 or more functions, or a firm-level figure it credits to AI, would move it to Scaling.
Open questions
- How many Accenture employees use Microsoft 365 Copilot each week now that it is rolling out to about 743,000, and how many use ChatGPT Enterprise and Claude Code?
- What cost or capacity return has Microsoft 365 Copilot delivered at Accenture, beyond users' self-reported speed and productivity gains?
- Which clouds and data centers host Accenture's internal AI workloads, including AI Refinery, and its ChatGPT Enterprise, Claude and Mistral deployments?
- To whom do CIO Tony Leraris and CTO Marc Warner report, and how does Leraris's enterprise technology role divide with Rajendra Prasad's Technology Reinvention Engine?
- How much of the roughly 5% rise in revenue per person in fiscal 2026 came from AI, as opposed to talent rotation, business mix and higher-value work?
- What did AI Refinery agents deliver in Accenture's marketing function against the 6% cost savings Accenture expected in Oct 2024?
- What cost or capacity return do agentic recruiting, the Workday agents and the contract-to-cash agents deliver, beyond time to fill and candidate quality?
- What does Accenture spend on AI tokens, licenses and infrastructure for its own use, and what has its internal token platform saved?
- Who leads Accenture's Responsible AI compliance program in late 2026, and does Arnab Chakraborty still hold the chief responsible AI officer role?
- How many people left Accenture in the 2025 business optimization program because reskilling was not viable?
- How will Accenture show AI's financial effect now that it no longer publishes advanced AI bookings and revenue?
- How much weight does AI tool use carry in Accenture's performance evaluations, and how is it measured?
Sources
Accenture
- Leadership
- Julie Sweet, Chair and CEO, biography
- Lan Guan, Chief AI and Data Officer, biography
- Marc Warner, CTO and CEO of Faculty, biography
- Kate Hogan, COO, biography
- Manish Sharma, Chief Strategy and Reinvention Services Officer, biography
- Kate Clifford, Chief Leadership and Human Resources Officer, biography
- Angie Park, CFO, biography
- Rajendra Prasad biography (still shows his Sep 2025 CTO title)
- Tony Leraris, Chief Information Officer, biography
- AI Refinery
- Accenture to invest $3 billion in AI (Jun 13, 2023)
- Accenture appoints Rajendra Prasad chief information and asset engineering officer (Jun 29, 2023)
- Accenture appoints Arnab Chakraborty chief responsible AI officer (May 14, 2024)
- Accenture launches AI Refinery with NVIDIA AI Foundry and Llama (Jul 23, 2024)
- Accenture and NVIDIA form a business group; marketing agents on AI Refinery (Oct 2, 2024)
- Accenture changes its growth model; Rajendra Prasad to become CTO (Jun 20, 2025)
- Accenture to acquire Faculty (Jan 6, 2026)
- Accenture announces Reinvention Services leadership (Mar 13, 2026)
- Accenture completes acquisition of Faculty; Marc Warner becomes CTO (Mar 16, 2026)
- Accenture Tokenomics launched (Jul 29, 2026)
- Accenture on open-weight models and its OpenShell pilot (Aug 26, 2026)
- Accenture appoints Emma Chalwin chief marketing officer (Sep 8, 2026)
- Accenture launches the Advanced Intelligence Group (Oct 5, 2026)
- Accenture expands AI Foundation and AI Factory (Oct 21, 2026)
Filings, letters and investor materials
- Accenture fiscal 2026 Form 10-K (Oct 2026)
- Q4 and full year fiscal 2026 earnings release (Oct 1, 2026)
- Q4 fiscal 2026 earnings call transcript (Oct 1, 2026)
- Q4 fiscal 2026 earnings presentation (Oct 1, 2026)
- Investor Day FY27 page (Oct 14, 2026)
- Investor Day FY27 transcript (Oct 14, 2026)
- Investor Day FY27 financial presentation (Oct 14, 2026)
- Q3 fiscal 2026 earnings call transcript (Jun 18, 2026)
- Q2 fiscal 2026 earnings call transcript (Mar 19, 2026)
- Letter to shareholders, fiscal 2026 annual report
- 2025 proxy statement (Dec 2025)
- Q4 and full year fiscal 2025 earnings release (Sep 25, 2025)
- 360° Value Report 2025
Partners
- Microsoft Source: Accenture is rolling out Copilot to about 743,000 people (Apr 27, 2026)
- OpenAI and Accenture: ChatGPT Enterprise for Accenture professionals (Dec 1, 2025)
- Accenture and Anthropic: business group and Claude Code (Dec 9, 2025)
- Accenture and Mistral AI collaboration (Feb 26, 2026)
- Accenture and Google Cloud: Gemini Enterprise Acceleration Program (Apr 22, 2026)
- Accenture Gemini Enterprise Business Group (Sep 8, 2026)
- Accenture and Anthropic: embedded AI safety evaluators (Sep 18, 2026)
- Accenture Dell Business Group (Oct 8, 2026)
- Accenture and Workday forward deployed engineering program (Oct 13, 2026)
- AWS Marketplace: AI Refinery, sold by Accenture and deployed on AWS
News
- The Motley Fool: Accenture Q4 fiscal 2025 earnings call transcript (Sep 25, 2025)
- The Motley Fool: Accenture Q1 fiscal 2026 earnings call transcript (Dec 18, 2025)
- TechRadar, reporting the Financial Times: Accenture ties promotions to regular AI use (Feb 19, 2026)
- Reuters, via Yahoo Finance: Accenture to roll out Copilot to 743,000 employees (Apr 27, 2026)
- Fortune: Ryoji Sekido on why Asia's AI moment depends on leaders (Oct 13, 2026)
Data confidence
- Confirmed
- Stated in a primary or reputable public source.
- Estimated
- Derived or approximated from public reporting.
- Unconfirmed
- Reported, but not yet corroborated.
Accenture profile · Enterprise AI diffusion report by Imagine · facts checked against their sources on Nov 1, 2026