0
KPMG people globally
A Case Study by Upforge Consulting.
+
Announced: 19 May 2026
Global AI Alliance
Tax & Legal Transformation
Private Equity Channel
Agentic AI
Enterprise Workflow Automation
KPMG–Anthropic Alliance
A platform-embedding alliance pairing KPMG's delivery engine with Anthropic's Claude models across tax, legal, advisory, and private equity workflows.
0
Countries and territories
$0B
KPMG FY25 global revenue
$0B
Tax & Legal FY25 revenue
0%
Tax & Legal growth
Weeks → Minutes
Workflow compression claim
Platform
Claude sits inside Digital Gateway, not beside it as a loose chatbot.
8 / 12
Eight key facts are public; twelve diligence items remain undisclosed.
Human
Tax, legal, and advisory work still needs human review.
Multi-model
Public reporting indicates KPMG remains ecosystem-based.
Summary
Not a chatbot license. A platform-embedding move.
KPMG and Anthropic announced a global alliance on 19 May 2026 to embed Claude into KPMG Digital Gateway, with early emphasis on Tax & Legal and private equity.
The strategic point is where the AI sits: inside a KPMG-controlled client-delivery environment, not as a standalone assistant. The platform layer controls permissions, review, logging, and client experience.
Public materials reference access for 276,000+ KPMG professionals, Claude Cowork, Managed Agents, Claude Code, and KPMG as Anthropic's preferred PE consultant.
Deal Breakdown
What is disclosed, and what it implies strategically.
| Dimension | Publicly Disclosed Detail | Strategic Meaning | Status |
|---|---|---|---|
| Announcement date | 19 May 2026 | Formal launch of the global alliance | Confirmed |
| Launch vehicle | KPMG Digital Gateway Powered by Claude | Productized client-delivery platform | Confirmed |
| Initial focus | Tax clients, legal workflows, private equity | High-value, document-heavy, regulated work | Confirmed |
| Workforce deployment | Claude access for 276,000+ KPMG people | Major internal adoption and training challenge | Confirmed |
| Embedded products | Claude Cowork, Managed Agents, Claude Code | Agentic workflow automation and coding modernization | Partly specified |
| PE channel role | KPMG named preferred consultant for private equity | Gives Anthropic consulting-led market access | Confirmed |
| Commercial terms | Deal value, duration, pricing not publicly disclosed | Important diligence gap | Not publicly disclosed |
| Audit scope | Not directly integrated into audit workflows according to WSJ | Reduces professional-independence risk | Partly specified |
Strategic Rationale
Why both sides wanted this alliance visible.
Why KPMG did this
- Strengthens fast-growing Tax & Legal.
- Moves AI into client workflow, not a side chatbot.
- Differentiates Digital Gateway across tax, legal, advisory, and PE.
- Improves the PE value-creation offer.
- Supports productivity across 276,000+ professionals.
- Keeps KPMG in control of workflow while Claude powers the model layer.
Why Anthropic did this
- Gains a marquee Big Four deployment.
- Reaches regulated enterprise workflows.
- Expands into tax, legal, cyber, and PE.
- Uses KPMG as a global client channel.
- Reinforces its partner-led enterprise strategy.
- Creates proof points against rival model providers.
Data Visuals
The metrics that make the alliance material.
Read the financial scale, service-line growth, adoption data, and impact claims together. Impact claims are company claims or proprietary estimates, not audited outcomes.
KPMG FY25 Revenue Mix
Revenue by function, totaling $39.8B.
$16.4BAdvisory
$14.1BAudit
$9.3BTax & Legal
KPMG Growth by Function
Tax & Legal had the fastest growth among the three major functions.
7.5%Tax & Legal
6.0%Audit
2.9%Advisory
AI Adoption Study
KPMG and UT Austin found a small sophisticated-user cohort in workplace AI usage.
1.4M
Workplace AI interactions analyzed
2,597
Unique users in the study
~5%
Sophisticated AI users
Claimed Operational Impact
Company claims and proprietary estimates only; not audited outcomes.
Tax-regulation agent build time
Weeks → Minutes
KPMG Blaze delivery claim
Half the time
KPMG Blaze cost claim
Half the cost
Cost reduction claim
50%-80%
Speed improvement claim
10x
Label: public company claims / proprietary estimates, not audited outcomes.
Diligence Visibility
Known facts help, but the diligence gap is material.
8Publicly known items
12Not publicly disclosed items
1Key interpretation: multi-model stack
Risk Exposure Profile
Analytical scoring from public disclosures, not company ratings.
5/5Model output risk
4/5Data / residency risk
4/5Independence oversight
Timeline
The alliance follows years of platform, cloud, and AI moves.
KPMG's Digital Gateway/Azure path meets Anthropic's consulting-channel push.
21 Jun 2022
KPMG and Microsoft add new Digital Gateway applications. Source: KPMG
17 Jun 2025
KPMG launches Workbench on Microsoft technology / Azure AI Foundry. Source: KPMG
16 Dec 2025
KPMG reports FY25 global revenue growth. Source: KPMG
3 Feb 2026
KPMG launches Tax AI Accelerator on Digital Gateway and Azure OpenAI. Source: KPMG
19 Mar 2026
KPMG and UT Austin publish workplace AI usage study. Source: KPMG
4 May 2026
Anthropic announces an AI services company with PE backers. Source: Anthropic
14 May 2026
Anthropic expands alliance with PwC. Source: Anthropic
19 May 2026
KPMG and Anthropic announce global alliance. Source: KPMG
26 May 2026
Fortune reports KPMG's wider AI strategy remains non-exclusive. Source: Fortune
28 May 2026
WSJ reports Claude is not directly integrated into audit workflows. Source: WSJ
Technical Architecture
Public sources identify the layers, not the hosting route.
Client Data + KPMG Proprietary Tools
Source systems, documents, tax/legal content, and KPMG assets.
KPMG Digital Gateway on Microsoft Azure
Workflow platform, client access, and orchestration.
KPMG Governance Layer
Trusted AI, cyber, risk, approvals, and audit trails.
Claude Cowork + Managed Agents
Agentic task execution and automation.
Claude Service Endpoint
Model calls, tools, and service routing.
AI Outputs back into Digital Gateway
Summaries, drafts, agents, and reviewable outputs.
Exact model-hosting route is not publicly specified.
KPMG Blaze for PE use cases
PE value creation and IT modernization.
→ Claude Code
Code modernization and delivery acceleration.
→ IT modernization
Treat acceleration as a company claim / proprietary estimate.
Technical Terms Explained
Plain-English glossary for the deal.
Large Language Model
An AI system that understands and generates text, code, and reasoning outputs.
Frontier AI
The newest, most capable AI models for reasoning, coding, document analysis, and workflow automation.
Agentic Workflow
AI performing multiple steps: reading documents, checking rules, drafting outputs, using tools, and returning a deliverable.
API
A software interface that lets one system communicate with another.
Data Sovereignty
Control over where data is stored, processed, and governed.
Zero Data Retention
A setup where prompts and outputs are not stored after processing, except narrow legal or abuse-prevention exceptions.
Model Layer
The AI model used for reasoning and generation.
Platform Layer
The workflow environment controlling data, permissions, UX, approvals, logging, and orchestration.
Human-in-the-loop
A control model where humans review, approve, or override AI outputs before high-stakes use.
Stakeholder Map
Who needs the alliance to work, and why.
Analytical view based on public deal structure and enterprise AI governance needs.
KPMG global leadership
Role Sponsor and allocator.
Incentive AI differentiation.
Concern ROI and risk posture.
Tax & Legal practice
Role Initial workflow owner.
Incentive Faster regulated work.
Concern Accuracy and trust.
Private Equity practice
Role PE go-to-market channel.
Incentive Value creation.
Concern Proof of economics.
Advisory / AI & Data Labs
Role Build use cases.
Incentive Reusable delivery patterns.
Concern Portability and debt.
Cybersecurity and AI assurance teams
Role Governance and assurance.
Incentive Trusted deployment.
Concern Permissions and audit trails.
276,000+ professionals
Role Users and reviewers.
Incentive Productivity.
Concern Training and workflow change.
Anthropic enterprise partnerships
Role Alliance expansion.
Incentive Big Four distribution.
Concern Delivery performance.
Claude product teams
Role Model, agent, and code capabilities.
Incentive Workflow proof points.
Concern Reliability and tooling safety.
Privacy, compliance, and security teams
Role Data controls.
Incentive Regulated adoption.
Concern Cross-border obligations.
Microsoft Azure
Role Platform infrastructure.
Incentive AI workload stickiness.
Concern Routing clarity.
Clients
Role Tax, legal, PE, and regulated users.
Incentive Faster KPMG expertise.
Concern Data, review, accountability.
Regulators / risk functions
Role Oversight bodies and committees.
Incentive Transparent controls.
Concern Misuse and independence.
3
Primary workflow markets: tax, legal, and private equity value creation.
4
Control domains to watch: privacy, security, independence, and model quality.
1
Core strategic question: who owns the client workflow layer?
Pros and Cons by Party
Upside is real, but governance has to work.
A concise view of public disclosures and standard enterprise AI diligence.
KPMG
Pros
- Stronger Digital Gateway differentiation.
- Better Tax & Legal automation.
- Access to Claude across workforce.
- Stronger private equity value proposition.
- Keeps client workflow inside KPMG's platform.
- Sharper Big Four AI positioning.
Cons / Risks
- Vendor dependence on Anthropic.
- Model-risk exposure.
- Data privacy and residency ambiguity.
- Change burden across 276,000+ people.
- Professional-independence constraints.
- Contract terms not publicly disclosed.
Anthropic
Pros
- Big Four enterprise distribution.
- Regulated workflow proof point.
- Access to KPMG's global client base.
- Expansion into tax, legal, cyber, and PE.
- Stronger position against OpenAI.
Cons / Risks
- Pressure in high-stakes workflows.
- Multi-jurisdiction compliance burden.
- Reputational risk if AI outputs fail.
- Possible channel conflict.
- Enterprise privacy and security expectations.
Clients
Pros
- Faster workflows.
- KPMG expertise and AI in one platform.
- Reduced context switching.
- Potential cost and cycle-time gains.
Cons / Risks
- Need clarity on retention and hosting.
- Need human review for high-stakes advice.
- Risk of overreliance on AI outputs.
- Need clear accountability.
Risk Heatmap
The real risk is uncontrolled adoption.
Analytical judgments from public disclosures; not KPMG or Anthropic risk ratings.
| Risk | Likelihood | Impact | Mitigation |
|---|---|---|---|
| Model hallucination / incorrect outputs | High | High | Human review, QA sampling, audit trails |
| Data privacy / residency uncertainty | Medium | High | Explicit contractual clauses |
| Vendor lock-in | Medium | Medium-High | Multi-model architecture |
| Professional-independence concerns | Medium | High | Ring-fence audit and restricted-client work |
| Adoption failure | High | Medium | Training, champions, workflow redesign |
| IP ownership ambiguity | Medium | Medium | Contractual IP provisions |
| Security exposure from agentic tools | Medium | High | Sandboxing, permission controls, logging |
| Reputational risk | Medium | High | Governance, disclosure, controlled rollout |
Risk ratings here are analytical judgments from public disclosures, not disclosed KPMG or Anthropic ratings.
Model hallucination
Tax and legal work needs QA, human signoff, and traceable source grounding.
Data privacy
KPMG-specific hosting, residency, and zero-data-retention terms are not public.
Independence
WSJ says Claude is not directly integrated into audit workflows, reducing but not removing oversight concerns.
Agentic tools
Agents need scoped permissions, logs, sandboxing, and approval workflows.
Competitive Landscape
Claude and OpenAI are using consulting channels to reach enterprise workflows.
KPMGAnthropic
Claude'%20gradientUnits='userSpaceOnUse'%3e%3cstop%20offset='0'%20stop-color='%23114a8b'/%3e%3cstop%20offset='1'%20stop-color='%230669bc'/%3e%3c/linearGradient%3e%3clinearGradient%20id='ac2a6fc2-ca48-4327-9a3c-d4dcc3256e15'%20x1='-1023.725'%20x2='-1029.98'%20y1='108.083'%20y2='105.968'%20gradientTransform='matrix(1%200%200%20-1%201075%20158)'%20gradientUnits='userSpaceOnUse'%3e%3cstop%20offset='0'%20stop-opacity='.3'/%3e%3cstop%20offset='.071'%20stop-opacity='.2'/%3e%3cstop%20offset='.321'%20stop-opacity='.1'/%3e%3cstop%20offset='.623'%20stop-opacity='.05'/%3e%3cstop%20offset='1'%20stop-opacity='0'/%3e%3c/linearGradient%3e%3clinearGradient%20id='a7fee970-a784-4bb1-af8d-63d18e5f7db9'%20x1='-1027.165'%20x2='-997.482'%20y1='147.642'%20y2='68.561'%20gradientTransform='matrix(1%200%200%20-1%201075%20158)'%20gradientUnits='userSpaceOnUse'%3e%3cstop%20offset='0'%20stop-color='%233ccbf4'/%3e%3cstop%20offset='1'%20stop-color='%232892df'/%3e%3c/linearGradient%3e%3c/defs%3e%3cpath%20fill='url(%23e399c19f-b68f-429d-b176-18c2117ff73c)'%20d='M33.338%206.544h26.038l-27.03%2080.087a4.152%204.152%200%200%201-3.933%202.824H8.149a4.145%204.145%200%200%201-3.928-5.47L29.404%209.368a4.152%204.152%200%200%201%203.934-2.825z'/%3e%3cpath%20fill='%230078d4'%20d='M71.175%2060.261h-41.29a1.911%201.911%200%200%200-1.305%203.309l26.532%2024.764a4.171%204.171%200%200%200%202.846%201.121h23.38z'/%3e%3cpath%20fill='url(%23ac2a6fc2-ca48-4327-9a3c-d4dcc3256e15)'%20d='M33.338%206.544a4.118%204.118%200%200%200-3.943%202.879L4.252%2083.917a4.14%204.14%200%200%200%203.908%205.538h20.787a4.443%204.443%200%200%200%203.41-2.9l5.014-14.777%2017.91%2016.705a4.237%204.237%200%200%200%202.666.972H81.24L71.024%2060.261l-29.781.007L59.47%206.544z'/%3e%3cpath%20fill='url(%23a7fee970-a784-4bb1-af8d-63d18e5f7db9)'%20d='M66.595%209.364a4.145%204.145%200%200%200-3.928-2.82H33.648a4.146%204.146%200%200%201%203.928%202.82l25.184%2074.62a4.146%204.146%200%200%201-3.928%205.472h29.02a4.146%204.146%200%200%200%203.927-5.472z'/%3e%3c/svg%3e)
Microsoft Azure%20--%3e%3csvg%20version='1.1'%20id='Layer_1'%20xmlns='http://www.w3.org/2000/svg'%20xmlns:xlink='http://www.w3.org/1999/xlink'%20x='0px'%20y='0px'%20viewBox='0%200%201180%20320'%20style='enable-background:new%200%200%201180%20320;'%20xml:space='preserve'%3e%3cpath%20d='M444.7,68.2c-47.3,0-86.1,38.7-86.1,86.1s38.7,86.1,86.1,86.1s86.1-38.5,86.1-86.1S492.2,68.2,444.7,68.2z%20M444.7,209.5%20c-29.4,0-53.1-24.1-53.1-55.2S415.3,99,444.7,99c29.4,0,53.1,24.1,53.1,55.2S474.1,209.5,444.7,209.5z'/%3e%3cpath%20d='M617.7,116c-15.5,0-30.6,6.2-38.5,16.7v-14.3h-31.1v167.3h31.1v-60.5c7.9,9.8,22.5,15.1,38.5,15.1%20c33.5,0,59.8-26.3,59.8-62.2S651.2,116,617.7,116z%20M612.5,213.3c-17.7,0-33.5-13.9-33.5-35.1s15.8-35.1,33.5-35.1%20c17.7,0,33.5,13.9,33.5,35.1S630.2,213.3,612.5,213.3z'/%3e%3cpath%20d='M750.6,116c-33.9,0-60.7,26.5-60.7,62.2s23.4,62.2,61.7,62.2c31.3,0,51.4-18.9,57.6-40.2h-30.4%20c-3.8,8.8-14.6,15.1-27.5,15.1c-16,0-28.2-11.2-31.1-27.3h90.4v-12.2C810.6,143.2,787.9,116,750.6,116z%20M720.5,165.9%20c3.3-15.1,15.8-24.9,30.8-24.9c16,0,28.2,10.5,29.6,24.9H720.5z'/%3e%3cpath%20d='M893.3,116c-13.9,0-28.4,6.2-35.1,16.5v-14.1h-31.1v119.5h31.1v-64.3c0-18.6,10-30.8,26.3-30.8c15.1,0,23.2,11.5,23.2,27.5%20v67.6h31.1v-72.7C938.8,135.6,920.6,116,893.3,116z'/%3e%3cpath%20d='M1018.3,70.6l-67.6,167.3h33.2l14.3-36.6h77l14.3,36.6h33.7l-67.2-167.3H1018.3z%20M1009,173.6l27.7-70l27.5,70H1009z'/%3e%3cpath%20d='M1167.5,70.6H1136v167.3h31.6V70.6z'/%3e%3cpath%20id='mark'%20d='M123.2,118.3V85c0-2.2,0.6-3.8,2.9-5.1L187.9,44c8.3-4.8,18.9-7,29.2-7c39.1,0,63.8,30.1,63.8,62.5%20c0,2.6,0,6.1-0.6,9l-64.7-37.8c-3.2-1.9-6.7-2.2-10.6,0L123.2,118.3z%20M266.1,236.6v-74c0-4.2-1.6-7-5.4-9.3l-82-47.7l28.8-16.7%20c1.6-1,4.2-1,5.8,0l62.2,35.9c17.6,10.3,29.8,32.7,29.8,54.1C305.2,204.2,289.8,227.6,266.1,236.6z%20M106.2,172.8l-28.5-17%20c-2.2-1.3-2.9-2.9-2.9-5.1V79.3c0-34.9,26.6-61.2,62.8-61.2c14.1,0,27.6,4.8,38.4,13.5L111.7,69c-3.8,2.2-5.4,5.1-5.4,9.3V172.8z%20M162,204.9l-38.8-21.8v-46.1l38.8-21.8l38.4,21.8v46.1L162,204.9z%20M186,301.9c-14.1,0-27.6-4.8-38.4-13.5L212,251%20c3.8-2.2,5.4-5.1,5.4-9.3v-94.5l28.8,17c2.2,1.3,2.9,2.9,2.9,5.1v71.5C249.1,275.7,222.2,301.9,186,301.9z%20M110.4,231.1l-62.2-35.9%20c-17.6-10.3-29.8-32.7-29.8-54.1c0-25.6,15.7-48.7,39.4-57.7v74.3c0,4.2,1.6,7,5.4,9.3l81.7,47.4l-28.8,16.7%20C114.6,232.1,112,232.1,110.4,231.1z%20M106.5,283c-36.8,0-63.8-27.6-63.8-61.8c0-3.2,0.3-6.4,0.6-9.3l64.4,37.2c3.8,2.2,7,2.2,10.9,0%20l81.7-47.4V235c0,2.2-0.6,3.8-2.9,5.1L135.7,276C127.4,280.8,116.8,283,106.5,283z%20M186,319.2c38.4,0,70.5-27.6,77.5-64.1%20c35.9-9,59-42.3,59-76.3c0-22.4-9.6-43.9-27.2-59.6c1.6-6.7,2.9-13.8,2.9-20.5c0-45.2-36.8-79.1-79.1-79.1c-8.7,0-17.3,1.6-25.6,4.5%20C179,9.7,159.4,0.8,137.6,0.8c-38.4,0-70.5,27.6-77.5,64.1c-35.9,9-59,42.3-59,76.3c0,22.4,9.6,43.9,27.2,59.6%20c-1.6,6.7-2.9,13.8-2.9,20.5c0,45.2,36.8,79.1,79.1,79.1c8.7,0,17.3-1.6,25.6-4.5C144.7,310.3,164.2,319.2,186,319.2z'/%3e%3c/svg%3e)
OpenAI
Deloitte'%3e%3cpath%20d='m179.01-17.156-12.299%2022.126%2078.703%200.0082282%2013.131-22.143v-3e-6z'%20fill='%23fd5108'%20stroke-width='2.1088'/%3e%3cpath%20d='m271.66-39.284-12.299%2022.126%2078.703%200.0082%2013.131-22.143v-2e-6z'%20fill='%23fd5108'%20stroke-width='2.1088'/%3e%3cg%20transform='matrix(.39636%200%200%20.39636%20-59.129%20-37.953)'%3e%3cpath%20d='m805.26%20177.46c-31.69%205.6338-48.592%2029.577-48.592%2071.127%200%2042.254%2022.535%2070.423%2055.634%2070.423%2016.197%200%2030.282-5.6338%2059.859-20.423v34.507c-35.915%2016.901-57.042%2021.127-85.211%2021.127-31.69%200-54.225-7.7465-71.831-26.761-18.31-18.31-28.169-42.254-28.169-69.718%200-59.859%2045.07-101.41%20110.56-101.41%2042.958%200%2073.239%2020.423%2073.239%2049.296%200%2019.014-14.085%2032.394-34.507%2032.394-10.563%200-19.014-3.5211-30.986-9.8592'/%3e%3cpath%20d='m646.81%20257.74c28.873-35.915%2039.437-50.704%2039.437-68.31%200-17.606-14.085-33.099-33.803-33.099-11.268%200-21.831%206.338-26.761%2011.972v73.944c-47.183%2063.38-47.183%2063.38-47.183%2063.38v-143.66h-45.07l-74.648%20123.24v-123.24h-25.352l-68.31%2015.493v17.606l37.324%203.5211v152.82h47.183l71.831-118.31v118.31h52.817'/%3e%3cpath%20d='m222.16%20327.46h10.563c38.028%200%2059.155-23.239%2059.155-69.718%200-38.732-16.901-59.859-48.592-59.859-4.2254%200-10.563%200.70422-21.127%202.8169zm0%2083.099c30.986%207.0423%2030.986%207.0423%2030.986%207.0423v17.606h-121.83v-17.606c28.169-7.0423%2028.169-7.0423%2028.169-7.0423v-212.68h-30.986v-17.606l73.944-22.535h19.718v25.352c34.507-21.127%2046.479-24.648%2064.789-24.648%2040.845%200%2071.831%2038.028%2071.831%2090.845%200%2060.563-41.549%20102.11-104.93%20102.11-7.0422%200-19.014-0.70422-31.69-1.4084'/%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
PwC
McKinsey
BCG%20--%3e%3csvg%20version='1.1'%20id='Layer_1'%20x='0px'%20y='0px'%20viewBox='0%200%20163.4%2043'%20xml:space='preserve'%20width='163.39999'%20height='43'%20xmlns='http://www.w3.org/2000/svg'%20xmlns:svg='http://www.w3.org/2000/svg'%3e%3cdefs%20id='defs1'%20/%3e%3cstyle%20type='text/css'%20id='style1'%3e%20.st0{fill:%23A100FF;}%20%3c/style%3e%3cg%20id='g1'%3e%3cpolygon%20class='st0'%20points='95.1,4.9%2095.1,0%20111.2,6.5%20111.2,10.5%2095.1,17%2095.1,12%20104.5,8.5%20'%20id='polygon1'%20/%3e%3cpath%20d='M%206.2,43%20C%202.8,43%200,41.3%200,37.5%20v%20-0.2%20c%200,-4.6%204,-6.2%208.9,-6.2%20h%202.3%20v%20-0.9%20c%200,-1.9%20-0.8,-3%20-2.8,-3%20-1.8,0%20-2.7,1%20-2.8,2.4%20h%20-5%20c%200.4,-4.2%203.7,-6.2%208.1,-6.2%204.5,0%207.8,1.9%207.8,6.6%20V%2042.6%20H%2011.4%20V%2040.4%20C%2010.4,41.8%208.7,43%206.2,43%20Z%20m%205,-6.6%20V%2034.6%20H%209.1%20c%20-2.6,0%20-3.9,0.7%20-3.9,2.4%20v%200.2%20c%200,1.3%200.8,2.2%202.6,2.2%201.8,-0.1%203.4,-1.1%203.4,-3%20z%20M%2028.4,43%20c%20-5.2,0%20-9,-3.2%20-9,-9.6%20v%20-0.3%20c%200,-6.4%204,-9.8%209,-9.8%204.3,0%207.8,2.2%208.2,7.1%20h%20-5%20c%20-0.3,-1.8%20-1.3,-3%20-3.1,-3%20-2.2,0%20-3.8,1.8%20-3.8,5.5%20v%200.6%20c%200,3.8%201.4,5.5%203.8,5.5%201.8,0%203.1,-1.3%203.4,-3.4%20h%204.8%20C%2036.4,40%2033.5,43%2028.4,43%20Z%20M%2048,43%20c%20-5.2,0%20-9,-3.2%20-9,-9.6%20v%20-0.3%20c%200,-6.4%204,-9.8%209,-9.8%204.3,0%207.8,2.2%208.2,7.1%20h%20-5%20c%20-0.3,-1.8%20-1.3,-3%20-3.1,-3%20-2.2,0%20-3.8,1.8%20-3.8,5.5%20v%200.6%20c%200,3.8%201.4,5.5%203.8,5.5%201.8,0%203.1,-1.3%203.4,-3.4%20h%204.8%20C%2056,40%2053.1,43%2048,43%20Z%20m%2019.7,0%20c%20-5.4,0%20-9.1,-3.2%20-9.1,-9.5%20v%20-0.4%20c%200,-6.3%203.9,-9.8%209,-9.8%204.7,0%208.6,2.6%208.6,8.9%20v%202.3%20H%2063.9%20c%200.2,3.4%201.7,4.7%203.9,4.7%202,0%203.1,-1.1%203.5,-2.4%20h%204.9%20C%2075.6,40.3%2072.6,43%2067.7,43%20Z%20M%2064,31%20h%207%20c%20-0.1,-2.8%20-1.4,-4%20-3.5,-4%20-1.6,0.1%20-3.1,1%20-3.5,4%20z%20m%2015.4,-7.2%20h%205.3%20v%202.8%20c%200.9,-1.8%202.8,-3.2%205.7,-3.2%203.4,0%205.7,2.1%205.7,6.6%20V%2042.6%20H%2090.8%20V%2030.8%20c%200,-2.2%20-0.9,-3.2%20-2.8,-3.2%20-1.8,0%20-3.3,1.1%20-3.3,3.5%20v%2011.5%20h%20-5.3%20z%20m%2026.4,-5.7%20v%205.7%20h%203.6%20v%203.9%20h%20-3.6%20v%208.9%20c%200,1.4%200.6,2.1%201.9,2.1%200.8,0%201.3,-0.1%201.8,-0.3%20v%204.1%20c%20-0.6,0.2%20-1.7,0.4%20-3,0.4%20-4.1,0%20-6,-1.9%20-6,-5.7%20v%20-9.5%20h%20-2.2%20v%20-3.9%20h%202.2%20v%20-3.5%20z%20m%2023.4,24.5%20H%20124%20v%20-2.8%20c%20-0.9,1.8%20-2.7,3.2%20-5.5,3.2%20-3.4,0%20-5.9,-2.1%20-5.9,-6.5%20V%2023.8%20h%205.3%20v%2012%20c%200,2.2%200.9,3.2%202.7,3.2%201.8,0%203.3,-1.2%203.3,-3.5%20V%2023.8%20h%205.3%20z%20m%203.9,-18.8%20h%205.3%20v%203.5%20c%201.1,-2.5%202.9,-3.7%205.7,-3.7%20v%205.2%20c%20-3.6,0%20-5.7,1.1%20-5.7,4.2%20v%209.7%20h%20-5.3%20z%20M%20154.8,43%20c%20-5.4,0%20-9.1,-3.2%20-9.1,-9.5%20v%20-0.4%20c%200,-6.3%203.9,-9.8%209,-9.8%204.7,0%208.6,2.6%208.6,8.9%20v%202.3%20h%20-12.2%20c%200.2,3.4%201.7,4.7%203.9,4.7%202,0%203.1,-1.1%203.5,-2.4%20h%204.9%20c%20-0.8,3.5%20-3.7,6.2%20-8.6,6.2%20z%20M%20151,31%20h%207.1%20c%20-0.1,-2.8%20-1.4,-4%20-3.5,-4%20-1.6,0.1%20-3.1,1%20-3.6,4%20z'%20id='path1'%20/%3e%3c/g%3e%3c/svg%3e)
Accenture
Capgemini| Alliance / Firm | Model Provider | Nature of Relationship | Strategic Meaning |
|---|---|---|---|
| KPMG + Anthropic | Claude | Digital Gateway platform embedding | Workflow-level enterprise AI |
| Deloitte + Anthropic | Claude | Workforce and industry solutions | Anthropic Big Four expansion |
| PwC + Anthropic | Claude | Expanded strategic alliance | Consulting-led Claude deployment |
| McKinsey + OpenAI | OpenAI | Frontier Alliance | Enterprise AI transformation |
| BCG + OpenAI | OpenAI | Frontier Alliance | Strategic AI consulting deployment |
| Accenture + OpenAI | OpenAI | Enterprise partnership | AI transformation at scale |
| Capgemini + OpenAI | OpenAI | Frontier Alliance | Systems integration channel |
Publicly Known vs. Not Disclosed
The diligence view: known facts versus open questions.
No assumptions are made on deal value, duration, pricing, revenue share, or minimum spend.
Publicly Known
- Alliance announced on 19 May 2026.
- Claude embedded into KPMG Digital Gateway.
- Initial focus includes Tax & Legal and private equity.
- Claude access planned for 276,000+ KPMG people.
- KPMG named preferred consultant for private equity.
- Digital Gateway is built on Microsoft Azure.
- KPMG Workbench is built on Microsoft technology / Azure AI Foundry.
- Claude not directly integrated into audit workflows according to WSJ.
Not Publicly Disclosed
- Deal value.
- Contract duration.
- Pricing model.
- Minimum spend.
- Revenue-sharing arrangement.
- Exact exclusivity boundaries.
- Exact Claude model versions.
- Exact model-hosting route.
- Zero-data-retention status for KPMG.
- Specific data-residency arrangement.
- IP ownership for co-developed PE products.
- SLAs / indemnities / termination rights.
Consulting Takeaways
The lesson is not "pick Claude." It is: control workflow, keep models flexible, and build governance early.
Own the platform layer, not just the AI license.
Start with high-value, document-heavy, judgment-intensive workflows.
Use AI to compress cycle time, not remove professional accountability.
Negotiate data, retention, residency, and IP rights upfront.
Preserve multi-model portability.
Build human-in-the-loop controls.
Measure economics, not just adoption.
Treat training and change management as core value drivers.
Source Library
Primary sources, reporting, and technical docs.
Use the filter to find supporting source cards.
PrimaryKPMG press releasehttps://kpmg.com/xx/en/media/press-releases/2026/05/kpmg-and-anthropic-sign-global-alliance-and-launch-digital-gateway-powered-by-claude.html
PrimaryAnthropic announcementhttps://www.anthropic.com/news/anthropic-kpmg
ReportingWSJ coveragehttps://www.wsj.com/cfo-journal/kpmg-taps-anthropic-to-revamp-global-tax-advisory-platforms-853093bb
ReportingFortune interviewhttps://fortune.com/2026/05/26/kpmg-anthropic-claude-partnership-big-four-ai/
FinancialKPMG FY25 revenue releasehttps://kpmg.com/xx/en/media/press-releases/2025/12/kpmg-delivers-rise-in-global-revenue.html
FinancialKPMG performance summaryhttps://kpmg.com/pk/en/about/performance.html
PlatformKPMG Digital Gatewayhttps://kpmg.com/xx/en/our-insights/ai-and-technology/kpmg-digital-gateway.html
PlatformKPMG Workbenchhttps://kpmg.com/us/en/capabilities-services/ai/kpmg-workbench.html
PlatformKPMG Workbench launchhttps://kpmg.com/xx/en/media/press-releases/2025/06/kpmg-launches-a-multi-agent-ai-platform-transforming-client-delivery-and-ways-of-working-across-the-global-organization.html
PlatformKPMG Tax AI Acceleratorhttps://kpmg.com/us/en/media/news/tax-ai-accelerator-program.html
StudyKPMG + UT Austin AI studyhttps://kpmg.com/us/en/media/news/utaustin-kpmg-study.html
ClaimsKPMG Blaze PDFhttps://kpmg.com/kpmg-us/content/dam/kpmg/pdf/2026/kpmg-blaze.pdf
GovernanceKPMG Trusted AIhttps://kpmg.com/xx/en/what-we-do/services/ai/trusted-ai-framework.html
PartnerAnthropic Claude Partner Networkhttps://www.anthropic.com/news/claude-partner-network
DocsAnthropic Managed Agentshttps://platform.claude.com/docs/en/managed-agents/overview
ProductClaude Codehttps://claude.com/product/claude-code
DocsClaude documentation introhttps://platform.claude.com/docs/en/intro
DocsClaude API and data retentionhttps://platform.claude.com/docs/en/manage-claude/api-and-data-retention
DocsClaude in Microsoft Foundryhttps://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry
CompetitiveAnthropic Deloitte partnershiphttps://www.anthropic.com/news/deloitte-anthropic-partnership
CompetitiveAnthropic PwC partnershiphttps://www.anthropic.com/news/pwc-expanded-partnership
CompetitiveAnthropic enterprise AI services companyhttps://www.anthropic.com/news/enterprise-ai-services-company
CompetitiveOpenAI Frontier Allianceshttps://openai.com/index/frontier-alliance-partners/
CompetitiveOpenAI Accenture partnershiphttps://openai.com/index/accenture-partnership/