As artificial intelligence moves from pilot projects to the core of enterprise technology stacks, it is not just client workflows that are changing. Tech Mahindra is reshaping its operating model, shifting towards outcome-linked delivery structures and deeper consulting-led engagements as AI becomes central to customer demand.
At the centre of this shift is the company’s use of “service tokens” — an internal unit that represents the scope of services delivered across the full AI stack rather than just traditional billing based on man-hours. “The tokens function as a form of service currency, mapping the effort required across design, automation, model work, infrastructure, and application development within a project,” Nikhil Malhotra, chief innovation officer and head of AI and emerging technologies at Tech Mahindra, said on the sidelines of the India AI Impact Summit.
Tokenisation of IT
Instead of approaching engagements horizontally — such as handling only infrastructure or application layers — the company now structures projects vertically around a client’s business problem. The required services across the stack are broken into measurable components, and tokens are assigned based on how much design work, manual effort, automation, or model integration is needed. This framework allows the company to define scope, estimate effort and structure billing in terms of delivered services rather than time spent.
Across industries, the company sees AI implementations typically built on a common core architecture with significant customisation for each client’s processes, data environment and compliance requirements. As enterprises push towards personalised, domain-specific AI systems, this combination of vertical delivery, consulting-led engagement and token-based service measurement is increasingly shaping how Tech Mahindra structures its next phase of growth.
AI adoption is also strengthening Tech Mahindra’s consulting business. As enterprises explore agentic AI and automation initiatives, many approach the company with broadly defined use cases.
“This has increased demand for advisory work covering ROI evaluation, process mapping, data readiness and implementation strategy. Consulting teams now often engage before any deployment begins, helping customers determine whether AI is necessary, what outcomes are achievable, and how investments should be structured,” Malhotra said.
The company follows a multi-stage workflow for AI projects. The first phase focuses on understanding the client’s problem statement and defining expected outcomes, including productivity targets or functional goals such as accuracy or response speed. This is followed by ROI assessment and feasibility analysis, after which the company provides a structured implementation plan outlining investment needs and expected returns.
Tech Mahindra is also deploying AI internally before scaling solutions externally, treating itself as the first customer for several initiatives.
Proving Agentic AI in HR and Finance
“For Orion, the agentic enterprise, the customer zero is TechM. And we are doing it (building use cases) in two areas – HR and finance as well (where TechM will be client zero),” Malhotra said.