Cognizant embeds engineers inside client offices for AI at scale

Most enterprise AI projects never leave the pilot stage. Companies spend months building proof-of-concept demos, then discover the gap between a controlled test and a live business process is far wider than expected.

Cognizant is betting on a different model. Instead of sending proposals and managing deployments from a distance, the IT services giant is now placing its own engineers physically inside client offices to run AI systems alongside Google Cloud's teams.

The move comes with a new title. Cognizant earned Google Cloud's Diamond Partner status, the highest tier in Google's newly restructured partner network. It reflects the company's track record of putting AI into actual production, not just drawing up plans for it.

How the forward-deployed model works

The traditional IT services model works like this: a vendor does the work offsite, ships a solution, and hands it over. That approach struggles with AI because context matters enormously. A billing system at a telecoms company has thousands of edge cases that only become visible when you are sitting inside the actual workflow.

Mphasis Bets CAD 30M on Decision AI, Acquires Canadian TAP

Mphasis acquires Continuum AI for up to CAD 30M to close the gap between analytics and business action. Here's what enterprise AI buyers need to know.

Mission MediaMission Media

Cognizant's Gemini Enterprise Practice flips that. Forward-deployed engineers work on-site with clients, alongside Google Cloud's own teams, to build and run AI in the real environment rather than a controlled one. Annadurai Elango, President of Core Technologies and Insights at Cognizant, describes it plainly: "Forward-deployed engineers working directly alongside Google Cloud's teams, solving real problems in production, on site."

The first public proof point: a North American telecoms provider used the practice to deploy AI agents targeting billing discrepancies worth hundreds of millions of dollars. What started as one use case grew to 12 more, spanning revenue recovery, invoice reconciliation, and workflow automation. That organic expansion is the point. On-site engineers see additional problems as they go.

The race to be on-site

Cognizant is not alone in moving this direction. At Google Cloud Next 2026, Google committed US$750 million to support partner-led AI development across its network of more than 2,900 services partners. Accenture simultaneously announced its own version of the forward-deployed engineer model through a Gemini Enterprise Acceleration Program.

The parallel announcements suggest this is becoming the standard approach for enterprise AI, not a niche experiment. Firms that win large AI transformation contracts are increasingly the ones willing to put senior people on-site and take accountability for outcomes.

What this signals for APAC enterprises

For marketing and technology leaders in Asia evaluating AI transformation, the signal is straightforward. The vendors gaining traction are not the ones offering the most sophisticated models. They are the ones with engineers who will sit inside your business and make AI work with your actual data and processes.

Adobe Rebrands Experience Cloud as AI-First CX Enterprise Platform

Adobe replaces Experience Cloud with AI-driven CX Enterprise platform featuring autonomous agents called Coworkers. 1,770+ enterprise customers already have access to production-ready AI agents for marketing automation.

Mission MediaMission Media

Google Cloud has already expanded its Gemini AI infrastructure into Singapore and Southeast Asia, including data residency and a regional AI agent platform. That infrastructure now has a delivery model to match it.

Cognizant's Agentic Employee Platform, built with Google Cloud, gives a concrete example of what on-site implementation looks like. It places AI agents into specific business roles (project management, marketing, learning and development) and automates 60-70% of manual effort in those workflows. Deployment timelines drop from quarters to weeks.

Kevin Ichhpurani, President of Global Partner Ecosystem at Google Cloud, frames the core challenge accurately: "Scaling AI into production requires an enterprise-ready platform combined with partners who can engineer targeted outcomes directly in the field."

The question for enterprise leaders is not whether to invest in AI. It is whether your implementation partner is close enough to the work to actually deliver it.