By Sumed Marwaha, Managing Director, AHEAD-India
As enterprises accelerate their AI ambitions, a structural gap is becoming increasingly visible across organisations: the gap between AI strategy and execution. Despite significant investment in AI, many organisations have not yet adjusted how AI initiatives are run relative to where delivery actually happens.
In my experience, a familiar pattern plays out. AI and automation programmes are conceived and piloted in the US or Europe, often in innovation teams or corporate functions, while the India-based delivery centres are left out of the early design. Strategy conversations happen between headquarters stakeholders, models are scoped and tools are selected, and only then are the GCCs informed about what will change in their environment.
One recent example from a large global services organisation is instructive. A US-based team was exploring an autonomous service desk and broader ServiceNow automation with stakeholders in the US. The largest delivery footprint for that work, however, sat in India, and the India delivery leaders were not initially at the table. Once the organisation brought those leaders into the conversation, mapped how the service desk actually worked and identified concrete use cases together, it became clear that the original programme design would have missed several practical constraints and high-impact opportunities.
When AI strategy and delivery are separated in this way, several problems appear:
- AI and automation use cases are defined without a grounded view of how work really flows through service desks, application operations or SRE teams in India.
- Change is pushed onto delivery teams rather than built with them, which slows adoption and undermines trust.
- The people closest to the systems and data are not involved when impact and risk are assessed
The root issue is simple. If the power centre and AI strategy sit in one geography, and the delivery centre sits in another, and those two are not tightly integrated, gaps appear. Those gaps show up as missed use cases, brittle implementations and avoidable delivery friction.
What good looks like when GCCs are empowered
Closing the gap between AI strategy and delivery is not only about changing reporting lines. It is about giving GCCs the tools, access and responsibility to shape how AI shows up in day-to-day work.
A few practical characteristics stand out in organisations that are doing this well:
- Democratised access to AI tools – Engineers, SREs, service desk agents and administrators in India have direct access to enterprise AI platforms, rather than relying on a small central team to mediate every interaction.
- Local ownership of agents and automation – Delivery teams can design and build agents and automations that reflect the repeatable tasks they handle every day, instead of waiting for distant product teams to prioritise their use cases.
- Tighter feedback loops – The same people who run production operations are involved in evaluating AI behaviour, refining prompts and models, and deciding what is safe to automate.
In organisations that are leaning into this shift, a two-step pattern is emerging. First, they make their AI platforms broadly available so that teams in India and globally can use them the way they might use any familiar assistant: to answer questions, accelerate analysis and support daily work.
The next step is more important. They enable people across roles – from developers and SREs to service desk agents and system administrators – to build their own agents that automate repeatable, high-volume tasks in their specific context.
When a service desk analyst in India can design an agent to handle the routine steps they take dozens of times a day, or an SRE can build an agent that codifies how they investigate and remediate a recurring incident pattern, the gap between AI strategy and execution narrows. AI stops being something that is done to the delivery centre and becomes something that is built and owned within it.
Across the industry, some organisations are providing Copilot-style licences at scale to GCC teams, while others are giving broad access to cloud AI services and tooling. The common pattern is that India based delivery centres are no longer passive recipients of AI initiatives. They are active participants in shaping how AI is applied to real operations.
What enterprise clients should be asking
For enterprise clients running complex AI transformation programmes, India’s evolution changes what they should reasonably expect from their technology partners. Partners with serious India operations can now offer AI-literate delivery capacity at a scale that simply did not exist five years ago. The relevant question is whether those partners have actually restructured how their India teams work, or whether they have only changed the language around the same underlying model.
A few practical questions are becoming increasingly important in RFPs and executive reviews:
- Where do your AI and data decision makers sit, and what do they actually own in terms of design, governance and running?
- How are India-based teams integrated into your risk, compliance and CSAT framework, not just into your staffing plan?
- What does your career and capability path for AI talent in India look like, and how does it reduce delivery risk for us as a client?
- How have you adjusted your operating model and accountability structure in the last three years to reflect India’s role in AI-led work, not just traditional IT services?
Delivery strategies built on older assumptions about where AI expertise lives will produce the outcomes those assumptions deserve: slower programmes, brittle operations and missed opportunities to differentiate.
Organisations that act early to align strategy with execution and embed capability where it truly exists will be better positioned to build scalable, resilient and high-impact AI programmes.