Share with your CHRO
Mphasis is restructuring how it deploys technical talent as AI-led work takes over its pipeline, shifting employees from offshore delivery centers to client sites during the early stages of engagements. The company reported $466 million in new contract wins in Q2 FY27, with 66% of that AI-led, and an overall pipeline up 23% year-on-year to a record high, 71% of which is now AI-driven. The onsite deployment surge is already compressing margins, with operating profit landing at 14.8%, down 50 basis points year-on-year.
What this means for your business
The workforce planning assumption embedded in decades of IT services is that AI would eventually let vendors do more offshore, more cheaply. Mphasis is reporting the opposite, at least in the early innings of AI projects. AI engagements front-load the relationship-intensive work: understanding a client’s data environment, mapping business processes, and validating that a solution actually fits the problem. That work can’t be done asynchronously from another timezone, and CHROs at enterprises buying these services should expect their IT service partners to push hard for onsite access and embedded team arrangements.
The pattern here is consistent with what happens when any genuinely complex technology enters production at scale. Requirements discovery for AI systems is harder than for conventional software because the failure modes are less predictable and the tolerance for misaligned outputs is low in regulated industries. Mphasis is essentially admitting that its “forward-deployed engineering” model, teams physically embedded with clients to co-develop solutions, is not a premium upsell but a structural necessity. The margin hit makes that candid: this isn’t a choice they’d make if they could avoid it.
The budget consequence for enterprise buyers is that AI services will carry a higher onsite labor component than traditional IT outsourcing, and that cost won’t disappear once the engagement matures because the certification of client-facing leaders on “value-stream playbooks” suggests Mphasis is building the embedded model into how it measures outcomes. CHROs who have workforce strategies anchored to the assumption that AI reduces vendor headcount on-premises should revisit that assumption now, before the next renewal cycle, because the contracts coming through today’s pipeline are being priced on the opposite logic.
Concept deep-dive: Forward-deployed engineering
Forward-deployed engineering places a vendor’s technical staff inside the client’s own environment rather than behind a service desk or in a delivery center. The model borrows from how defense contractors embed engineers at military sites. The business case is that complex systems, especially AI ones where requirements shift as the model’s outputs surface new questions, need engineers who can observe operations directly rather than work from documented specs. The tradeoff is higher cost and lower scalability compared to offshore delivery.
Based on reporting from Mphasis Increases Onsite Talent Deployment as AI Projects Surge, ETHRWorld, originally published 2026-11-09 07:56:00.

