Has the Indian IT industry structurally shifted from high growth to mid-growth? Are we entering a new cycle driven by AI, comparable to cloud adoption earlier?

With each evolution in technology, the services industry caught hold of the new technology and grew. What has changed with AI?

The AI technology stack includes the infrastructure layer -- the data centers. The play for the services industry is less here since it requires different competencies. But on top of the large-language models (LLMs) are platforms that AI companies have built. There’s a lot of noise around these players. But can enterprises simply plug and play these tools? No. Looking at that architecture visualized in the AI operating model, at each layer, there needs to be functional and domain context. This lies with the services companies, because we have been working with customers for a long time. We have vertically aligned structures, by and large. That’s a big advantage.

Secondly, the data within an enterprise has to be aligned, whether structured or unstructured, with the AI services being built. The services industry has to do this work. And on top of it, one can build micro-applications or agents to deliver services. Again, the services industry will enable this. All these areas are beyond the traditional IT budget and are a huge direct cost for customers.

But we are going through a transitional phase. One is that the services industry is pivoting its capabilities from a traditional goal to an AI-oriented one. The faster one pivots, the faster they reach the destination and can grow from there.

Some companies have different challenges. Example, they may have more legacy businesses, and hence, be subject to more deflationary pressure while pivoting. Hence, the pivot may take longer.

The core services are shrinking and will continue to for two reasons. One is that the price point will go down, so thereby, for the same services, one gets charged less. This is the case even with an AI infusion. Plus, the market is not expanding. The value share is shifting from one vendor to another as part of the vendor concentration.

Hence, companies like us can still grow. One can take someone else’s market share to grow because core services are a huge market. But the unit prices are going down. A select few vendors will remain in the end.

Simultaneously, a new opportunity is emerging with agentic services on the direct cost of business, which is bigger than the IT budget. When one pivots, they should move to the new addressable market.

Today, Fortune 2000 enterprises have hundreds of local, regional vendors, and specialised vendors in the services sector. All of this will be consolidated. Most enterprises, over time, may have 5-6 vendors in core services. That’s still a big market opportunity, to capture which one must be competitive. Pricing models must be reinvented, with costs coming down. It has to be a different price range for the company to consolidate with you. Hence, there is an opportunity to reimagine the core services.

There is a bigger opportunity in the agentic solutions or AI-native services. When these two come together, there will be double-digit growth.

In the next five years, we would like to double LTM’s revenue as part of the ‘Lakshya 31’ strategy. To do this, while inorganic will help, we still have to grow closer to double digits. If we can get our playbook right on the core services by re-imagining and doing it faster on the newer services, we can get that growth.

What percentage of your revenues today can be said as AI-led rather than AI-labeled?

We decided not to go down that path because we had gone on a similar topic when the cloud came in. Eventually, everything became cloud-oriented. A similar case with digital revenue.

The company’s mindset is to become an AI-centric organization. Anything that we do for ourselves and for our customers has to be AI-driven. It’s about whether we are pivoting on our attributes to become an AI-centric organization. We can be evaluated based on how we are doing on the talent transformation, investments in our IPs, patents, and platforms, and the industry solutions we’re building.

What percentage of your workforce is actually deployable on AI work today? Attrition is down, utilisation is soft—is that stability or early warning of weak demand?

Talent always wants to be associated with growth because it provides career progression, positivity, and mindset. Our attrition has come down, and it’s flat from across the quarters because we are a growth-oriented company. We don’t have narratives about layoffs or large-scale restructuring. Our employees are focused on delivering growth and value to the customer.

Skilling is a work in progress. In Q4, 90 per cent of the employees have done the intermediate AI course, and more than 50 per cent have done advanced courses.

We announced a partnership this quarter with IIT Kharagpur to develop a joint curriculum, not just for our workforce but also for freshers from different colleges. We spend time with them and plan with the academic institutions so that it’s part of the curriculum. We also announced partnerships to offer MIT Open Learning’s Universal AI to our workforce. More AI courses are underway for our employees.

We are getting great support from our partner ecosystem — the hyperscalers and the new AI companies who are working with us to bring their course content and learnings to our employees. It’s a work in progress, but we have made great progress with 90 per cent of our workforce trained with intermediate AI skill sets.

You’ve talked about efficiency—will AI-led productivity gains stay with you or get passed back to clients?

That is more about how the contracts are structured for the existing book of business and how they will be structured for the new set of business. For the existing contracts, we pass on the productivity benefit to the customer, but they are backed by genuine implementation. They’re not discount models. So, if we’re getting that productivity benefit, we’re reducing the cost as well. For us, it has to be margin neutral if not margin upside. All our productivity engagements have been margin-neutral because we’ve proportionately reduced the cost when we passed on the benefit. Many clients love it because they know that if they take a discount, it’s not sustainable. They also want to see the speed of response, time to market, and the operational benefit, which is important for the business. Coming to the new contracts, we price them with a new benchmark of productivity. It is in line with our standard operating margins for the new contracts. That’s why across sectors, margin has not been an issue. It has been the revenue deflation, and it’s the same story with us.

What verticals are still experiencing some softness? Are clients still delaying discretionary spends, or are you seeing a real pickup?

We work in an environment where, within a sector, there are 10-20 focus accounts. So any headwind we face usually is due to one or two accounts going through transformations. In BFSI, except for one client, we have grown double-digit. We have also seen double-digit growth in verticals like consumer and manufacturing. For instance, one of our top accounts in the tech vertical was going through productivity headwind. And yet, the vertical has grown sequentially double-digit. But for the full year, it saw flattish growth because we started the year in that account by passing on the productivity benefit. Every sector has its own pressure points. Because of all the geopolitical tensions, consumption is impacted; oil prices are high, and the travel vertical is impacted. Retail is struggling with supply chain costs. What we look forward to is whether there are any headwinds in the set of accounts in those sectors. We are focusing on a mitigation strategy if that happens. With BFSI, we expect a recovery. But it will not be at the same speed as the deflation occurred.

Are you seeing more cost-takeout deals than transformation deals?

It’s both. Vendor consolidation is essentially cost takeout deals. AI-led agentic services, reimagining business processes, modernisation of data infrastructure applications to keep it AI-ready, are all transformational spend. Any savings coming out of the vendor consolidation flow back into transformation. The reality is that no customer is sitting on a ton of new budget. You have to find money somewhere. So vendor consideration is a great opportunity to find some money so they can achieve both the transformation mandate and the cost-saving mandate. Clients are allocating a dedicated budget for the AI transformation. However, the speed at which they spend depends on the geopolitical situation.

Are clients pushing for renegotiation or price cuts in renewals or large deals?

Historically, every renewal has been an opportunity to renegotiate. This has changed. It appears in the form of a consolidation, because clients also understand there is a benefit of economies of scale. When they are negotiating one contract with you, they’re trying to see if, for instance, if they give us the other contract, could we give them better benefits? If they consolidate three vendors, can we give them more benefits? A deal we announced in Q2 saw a similar arrangement. We were already doing some work and took on more by giving a discount on the existing piece of work, as well as the overall scope. So it’s a win-win arrangement.

Which certain skill sets or niches be better pursued through inorganic moves rather than organic growth?

We are looking at inorganic growth two-dimensionally. If any companies can give us faster access to a skill set, whether on the engineering side, or if they can bring a certain culture and mindset, we may consider them. But currently, not many such companies are available; most are traditional companies positioning in that narrative. Going forward, we may look at start-ups or early growth companies. We already invested in Voicing.AI since we felt the contact centres are ripe for disruption, which turned out to be reality. Many of our existing customers are adopting Voicing.AI.

The second dimension is that the AI will take a sovereign angle much faster than anticipated. There will be a push to localise the models, platforms, data centres, infrastructure, or the data. We are actively pursuing partnerships, organic development, and offerings to see how we can help our customers, especially those in Europe and in the Middle East, to adopt a sovereign-based AI solution. To develop a sovereign AI solution, the most important part is the domain understanding. You also need access to the people who are near shore or on-site, which complements our offshore delivery centres. With that combination, we can start building sovereign solutions.