Cognizant to Certify 10,000 Engineers on Google's Gemini Enterprise

Cognizant said on Tuesday it will certify at least 10,000 of its professionals on Google Cloud's Gemini Enterprise platform and roll the tool out to 100,000 of its own employees this year, with plans to reach 200,000. The expanded partnership shows how the companies that build enterprise AI systems for clients are retraining their own workforces first.

The announcement, distributed through PR Newswire and published by MarTech Series, builds on a dedicated Gemini Enterprise practice Cognizant set up in April. Cognizant is also deploying Google Workspace across its global organization, though the companies did not give a seat count or timeline for that part of the deal.

What is Cognizant deploying with Google Cloud?

The deal has two sides: what Cognizant uses internally and what it sells to clients.

Internally, the target is 100,000 associates on Gemini Enterprise in 2026, scaling to 200,000. For clients, Cognizant is packaging four offerings:

• A "Frontier Certified Engineer" delivery model that places senior engineers alongside Google Cloud teams.
• An agentic employee platform that deploys AI agents into specific enterprise roles.
• Contact center services built on Gemini Enterprise for Customer Experience.
• Agent Foundry, a rapid agent development program that the companies say has produced more than 2,000 agents so far.

Cognizant is targeting retail and consumer goods, healthcare and life sciences, communications, media and technology, and financial services.

"This partnership reflects exactly what our AI Builder strategy is built to do," Cognizant chief executive Ravi Kumar S said in the announcement. Karthik Narain, chief product and business officer at Google Cloud, described Cognizant as leading by example by using the tools across its own workforce.

What results does Cognizant claim?

The companies cited several early metrics, all self-reported:

• Up to 30% faster software development, based on internal Cognizant benchmarks.
• Role-based agents that can automate 60% to 70% of manual effort in targeted workflows.
• For an unnamed US communications and entertainment client, a 17% rise in first-contact resolution in its contact center against pre-deployment baselines, with nearly one-third of appointment requests resolved through automation. Agents went live in three months, followed by more than 500 model optimizations in the first year.
• Agent Foundry, which the companies say can shrink multi-month development cycles into two-week sprints.

Enterprise buyers should read these as vendor claims rather than audited results. The qualifier about targeted workflows in particular signals that the 60% to 70% figure applies to selected processes, not to whole job roles.

Why are IT services firms retraining their own staff on AI?

Services firms sell expertise, and clients increasingly want partners who have already run AI at scale inside their own operations. Certifying engineers on a specific platform also deepens the alliance with that cloud provider, which matters as hyperscalers compete for the integrators that steer enterprise platform decisions.

There is a workforce question underneath. Cognizant's plan to reach 200,000 users gives a sense of the scale. Much of the work done at large services firms, from code maintenance to help desks, is exactly what agents are being built to automate. Retraining at this scale is partly a bet that those people can move up to designing, supervising and tuning agents rather than doing the tasks themselves.

Practitioners at AI Connect events have argued that this kind of training only works when it is tied to real work. "Meaningful upskilling really comes from when it's contextual and not generic," Arpit Gularti, associate director of analytics at The Clorox Company, said at COLLIDE 2025, in a panel on AI change management and workforce trust. At the same session, Katie McDonald of UnitedHealth Group's chief data office said her organization measures AI differently from headcount: "It's not headcounts, it's business value."

What should enterprise buyers ask their AI services partners?

For CIOs who buy from Cognizant or its rivals, the announcement raises practical questions:

• Certification versus delivery. How many engineers on a given account have shipped production agents, not just passed a platform exam?
• Platform neutrality. Will a partner tied closely to one cloud recommend another when it fits the problem better?
• Measurement. Will productivity claims be measured against the client's own baselines, with access to the data?
• Ownership. Who owns the agents, prompts and evaluation sets built during an engagement?

These questions echo lessons from AI Connect panels on moving AI pilots to measurable ROI, where speakers stressed that stalled projects usually fail for organizational reasons, not technical ones.

What happens next?

The test is whether Cognizant hits its 100,000-user target this year and whether client results match the internal benchmarks. Other large services firms are likely to face client questions about their own AI deployments, and the answers will show which partners are using the tools they sell.

Frequently asked questions

What did Cognizant announce with Google Cloud?

Plans to certify at least 10,000 professionals on Gemini Enterprise, deploy it to 100,000 associates this year with a path to 200,000, and roll out Google Workspace globally.

What results has Cognizant reported?

Self-reported figures include up to 30% faster software development and, for one US client, a 17% rise in first-contact resolution in its contact center.

What should enterprises ask AI services partners?

How many engineers have shipped production agents, whether the partner is platform-neutral, how results are measured, and who owns the agents and evaluation sets built.

Filed under: People, Culture & Leadership, Agentic AI