AI Threat Assessment · 27 May 2026

CGI Inc.

IT Services
COOKED
7.3/ 10

“CGI spent five decades becoming one of the largest IT consulting firms in the world — 90,000 people across 400 locations, with clients who've trusted them to build the infrastructure that runs governments and Fortune 500s. The problem is that the thing they charge $200 an hour for — translating business requirements into working code — is now something Cursor does in real-time while you type, and GitHub Copilot ships directly to the client's developers without the consultant markup.”

Business Model
8.0
Automation Risk
8.5
Moat Strength
4.5
Adaptability
6.0
Need Survival
7.5
AI Threat Level
BUSINESS MODEL REPLACEABILITY

Their 'end-to-end services' promise to 'design, implement, run and operate' technology solutions, which held up beautifully until AI started doing the designing in Figma, the implementing in Claude, and the operating through autonomous agents that don't bill quarterly retainers.

8.0
WORKFORCE AUTOMATION RISK

Systems integration, application development, and business process consulting — the holy trinity of billable hours — are being compressed into AI workflows that clients can run internally. When a bank can prototype, test, and deploy with Cursor instead of hiring 200 CGI consultants for 18 months, the 'comprehensive suite' becomes comprehensively redundant.

8.5
MOAT STRENGTH

Long-term government contracts and deep enterprise integrations create genuine switching friction — clients don't casually migrate away from systems that run their payroll or power grids. The moat buys them time while clients figure out how to rebuild with AI-native tools, but it's a delay tactic, not a permanent defense.

4.5
AI ADAPTABILITY SIGNALS

They're frantically branding everything as 'AI-enabled' and writing blog posts about 'velocity arbitrage,' but their recent 'insights-driven' repositioning still assumes clients need humans to turn insights into code — exactly the assumption Copilot is busy proving wrong.

6.0
WILL THE NEED SURVIVE AI?

Enterprise digital transformation survives; the headcount delivery model doesn't. AI eliminates the question 'How many consultants do we need?' by changing it to 'How many AIs should our 20 internal engineers manage?' — and the answer is never 90,000.

7.5
Verdict

“The long-term contracts buy CGI a few more years to figure out how to sell AI supervision instead of human hours, but the clock is ticking on a business model that charges enterprise prices for what's becoming consumer software. They built a consulting empire on the assumption that turning strategy into code requires an army — turns out it just requires an API.”

Scores are based on public information and AI analysis. This is an affectionate roast, not a financial assessment. The best companies use this as a mirror, not a verdict.

Roast another →