How do you see the impact of tools like Anthropic’s Co-work? Many believe this could be a survival moment for the industry.
There is no doubt that models, such as Claude Co-work, have delivered significant improvements in productivity and performance. These tools are evolving rapidly and have moved beyond being mere coding assistants. They are beginning to orchestrate entire workflows and knowledge-intensive tasks. That shift must be acknowledged.
The pace at which these systems can now produce outputs is remarkable. There will undoubtedly be meaningful productivity gains.
However, when we look at traditional IT services environments, where most services firms operate, the reality is very different. These are highly complex, deeply embedded systems. Client environments typically involve legacy infrastructure, regulatory obligations, fragmented databases, cybersecurity requirements, and integrations across multiple systems.
This is precisely where services companies add value by integrating, modernising, building micro services, and stitching together disparate systems.
To assume that newage tools, such as Claude Co-work or similar agentic systems, can simply “plug and play” and replace this complexity is misplaced. While these tools are powerful, they cannot fully substitute for the intricate enterprise context in which services firms operate.
The technology services industry remains structurally strong. The complexity of enterprise systems, and the contextual knowledge required — whether in a pharmaceutical company or a bank — cannot be easily replaced. Services firms understand how to navigate these environments and bring systems together. That is why we should not be overly concerned; the reaction has been exaggerated.
Have we seen a similar technologic al shift in the past that raised comparable concerns but ultimately became part of the industry?
The closest parallel would be the ERP (enterprise resource planning) wave, when platforms, such as Oracle and SAP, became mainstream. At the time, there was a strong belief that full-blown ERP systems — essentially the early phase of the SaaS era — would eliminate custom coding altogether and make services companies redundant. In reality, the opposite happened.
Demand for systems integrators increased fivefold. Services companies generated far greater value through integration, customisation, and maintenance than ERP vendors themselves did in a typical implementation.
I am not suggesting that today’s AI wave is identical, but it is a useful analogy. This cycle may be different because generative and agentic AI can directly produce code and potentially replace certain traditional development platforms. So yes, the dynamics will evolve.
I am not downplaying the impact of agentic AI. It is foundational and powerful. The real question is our ability to harness that power within enterprise systems. In fact, this could be more of a tailwind than a headwind — though we must acknowledge that not every company will adapt successfully.
The role of services firms will evolve. They will move beyond being implementation partners to becoming AI orchestration partners. They will integrate external AI frameworks — whether Claude or others — into enterprise environments, embed these tools into workflows, manage governance, and ensure integration with legacy systems.
You cannot simply ask an AI tool to build a complete enterprise system overnight.
If AI tools become deeply embedded and deliver sustained productivity gains, won’t the long-term need for system integrators decline?
I don’t believe the need disappears — it changes. Enterprises will need SI partners to do very different things from what they do today.
Take data readiness. Many enterprise databases have evolved over decades. Preparing that data for AI systems — cleaning it, structuring it, and ensuring governance and compliance — is a massive task. It does not happen automatically simply because an AI tool exists.
Companies that have built strong data engineering, AI, and platform capabilities, particularly those with partnerships with hyperscalers, will continue to thrive. Someone still has to evaluate which AI frameworks to adopt, how to integrate them with existing environments, and how to align those decisions with business objectives and governance standards.
Enterprise architectures are deeply embedded and interconnected. Interface contracts, regulatory requirements, and operational dependencies make them extremely complex. That complexity sustains the relevance of services organisations.
Over the next 10-15 years, what we are likely to see is gradual evolution, not abrupt displacement. The expertise within services firms will remain critical, especially in helping enterprises meaningfully leverage AI within their broader technology landscape.
Two structural shifts are visible: pressure on entry-level hiring and concerns about growth. Are sustained double-digit growth rates unlikely?
You will need to wait for our Strategic Review next week — on February 24 in Mumbai — where we will share more concrete numbers and forward-looking commentary. I think you will find some of the projections interesting.
One structural reality already visible is the divergence between revenue growth and employee growth. That decoupling is real, and the industry will have to come to terms with it. It implies a stronger focus on efficiency. Many companies have already begun that journey, and we will outline in greater detail how we see it playing out.
We look into the future in two phases. The next two years will be about navigating and stabilising through this transformation — adjusting operating models and embedding new technologies. The following three years could open up a new growth trajectory.
