← VuduVationsThe Intelligence Bureau
VuduIntelConsulting-AI Intelligence
← Firm Dossiers
Firm DossierBig Four + Accenture● Live data
Deloitte
Largest enterprise Claude deployment; 1,000+ Google Cloud agents.
Revenue (FY25)
~$67B global
Headcount
~460,000
AI / GenAI book
Not disclosed
Market's Verdict
Private
Owns the loop?
Rents
The Read

Deloitte is making the most optimistic bet in the Big Four: that AI will create more consulting work than it destroys. As a private partnership, it answers to no stock market pricing in job losses, so it is free to tell clients the wave of AI adoption will grow the business faster than automation shrinks it, and it is putting real product behind that story.

That product is Zora AI: a set of ready-to-use AI “agents” for finance, HR, supply chain and customer service, sold on subscription rather than by the hour. Deloitte runs it on itself first, using its own finance team as the test case and chasing a quarter less cost with forty percent more output, and it is reshaping its workforce around senior “architects” who design solutions rather than armies of juniors who build them. At roughly $67 billion and 460,000 people, the largest of the Big Four, it has the scale and the partnerships to push this harder than anyone: NVIDIA, Oracle and SAP on the technology side, and the single largest rollout of Anthropic’s Claude anywhere, to all 470,000 of its people.

The optimism doubles as a hedge. “AI is a demand driver, not a disruption,” as CEO Joe Ucuzoglu puts it, is exactly the message that protects a business still built on billable hours, and Zora itself runs on NVIDIA’s technology, not Deloitte’s own. In plain terms: Deloitte rents the AI engine it sells, and frames renting as winning.

The net of it: no firm is better placed to cash in on the implementation boom. But the same story that sells the boom is how Deloitte sidesteps the harder question its publicly traded rivals are already being punished for: what happens to a people-heavy business when the AI does the work. The market is starting to ask. Deloitte’s answer, for now, is to sell more help getting there.

Where They Stand, 10 axes
Replace ↔ Substrate42
Deflation ↔ Expansion74
Tech ↔ Work redesign68
Accountability38
Pricing model22
RAI88
CEO ↔ CAIO78
Agents ready20
Growth ↔ Cost22
Multi-model45
Recent Signals
The technology is unbelievably powerful, but this is hard work, to go into entrenched, large institutions and redesign workflows and process.
Joe Ucuzoglu, Global CEOColumbia Business · Jun 16
The value is the ability to think cross-functionally, to bring together a multidisciplinary set of capabilities and architect solutions.
The gap between what is possible in theory and what works in practice is where much of today’s opportunity resides.
Alliances
Anthropic
Confirmed Jul 2024; expanded Oct 2025 to all 470,000 professionals; $2B IndustryAdvantage.
Google Cloud
Dedicated Agentic Transformation Practice; 1,000+ agents; Gemini Enterprise (Apr 2026).
Deloitte press room
NVIDIA
Zora AI agentic platform built on the NVIDIA stack.
Wiz · Oracle · SAP
Deloitte-Google-Wiz AI cyber-defense collaboration (Jun 2026); ERP-embedded agents.
Deloitte press room
The People
Joe Ucuzoglu
Global CEO
Nitin Mittal
Global AI Leader; Principal
Costi Perricos
Global GenAI Business Leader; founder, Deloitte AI Institute
Dounia Senawi
Principal; US Chief Commercial Officer
Jim Rowan
US Head of AI; Principal
Beena Ammanath
Exec Director, Global Deloitte AI Institute
The VuduVations Read
Deloitte's seven dimensions of Trustworthy AI are the right specification, delivered as advisory. MCOS implements four of them as deployed code before the LLM ever runs. The framework is theirs; the infrastructure is ours.
Stack & Stance, Deloitte vs Consulting-as-Code
Their stackZora AI + Agentic Network (NVIDIA-built)
Capability, output-level
Deloitte
Consulting-as-Code™
via MCOS
Scale & delivery muscle
Client relationships / domain depth
Frontier-model access
Owns the loop vs rents the model
rents
Auditable, source-cited outputs
Client-owned procedure (sovereignty)
locked
Outcome-underwritten vs effort-billed
hours
VerdictDeloitte wins the top of the table, muscle, access, relationships. The Consulting-as-Code model, delivered by MCOS, wins the bottom: ownership, auditability, sovereignty, outcome. The tape is repricing the top while the bottom holds.
The Single Diagnostic
Does Deloitte own the loop, or rent it?, the firm-specific ownership test.
In the Archive, 8 dispatches
No. 18 · Pricing
The billable hour is dying, and only a product survives it.
No. 18 · Governance
A competitor named the governance crisis and sold you the fix in the same breath.
No. 18 · Special Report
The Standards War: how the US and China are building two AI operating systems
No. 6 · Consulting Layer
The firms with the most to lose from OpenAI's move said nothing the week it landed.
No. 3 · Architecture
The Big Five want weeks of data plumbing before their AI runs, and the market is quietly walking away from it.
No. 3 · Consulting Layer
The consultants are printing precise AI return numbers, but not the path from a raw document to the number.
No. 2 · Consulting Layer
The firms that bet their AI practices on Microsoft are now defending a moat that no longer exists.
No. 1 · Consulting Layer
The big firms are naming the compute crunch so they can sell you the cure forever.
← All Firm DossiersThis Week →