Building India’s R&D Backbone

India’s chemicals industry stands at an inflection point. As global supply chains fracture and geopolitical realignments accelerate, the pressure on Indian manufacturers to build genuine, homegrown R&D capability has never been greater. For decades, the sector operated on a familiar plan: skip the messy, expensive pilot stage, jump from laboratory to full-scale production and hope the gaps close themselves. They rarely did. The result was an industry that imported both technology and dependency in equal measure.

Xytel was built to close the existing gaps. India’s first ‘pilot plant’ company, it constructs – as one might guess – mini working models of large-scale chemical plants. These models serve as a critical bridge between laboratory discovery and industrial process. Xytel operates across oil & gas, pharmaceuticals, fine chemicals, fertilisers and defence. In a country still building the institutional muscle for serious R&D, Xytel sits at a rare nexus: part engineering firm, part innovation partner, part national infrastructure.

Rajeev Hallur, the firm’s Managing Director, has spent thirty years in this space, including 8 years leading Xytel. In this conversation, he speaks with uncommon candour about how his company decides which innovation bets to make, what a costly product failure taught him about market excitement and engineering discipline, how geopolitical turbulence has rewritten the rules of supply chain and why AI’s biggest challenge in Indian manufacturing is not the capital — it’s culture

Innovation is a word that gets thrown around a lot. At Xytel, what does it mean and how do you decide which bets are worth making?

For us, innovation is not a department or a strategy add-on, it is the business. Every single project we take on is an innovation in itself. Our pilot plants are the mechanism through which innovation happens for the entire Indian chemical industry.

Right now, we’re seeing enormous activity in green energy (specifically green hydrogen), and in what’s become the latest buzzword: Sustainable Aviation Fuel or SAF. We receive at least one inquiry every day in this space alone.

If there are more than three segments in a project where we only have a Plan A and nothing beyond, that’s a no-go. We will not touch it. Beyond that threshold, we run a classic cost-benefit analysis, checking whether our innovation direction is in sync with where the industry is heading, and then we decide how to move forward.

Xytel sits at the R&D nexus of Indian chemical industry. Has the culture around R&D and intellectual property in India changed in the last few years?

Dramatically and for the better. 10 or 15 years ago, if we sent out 100 proposals to the private sector, one or two would materialise. The culture was to skip the pilot stage entirely, jump from laboratory to full-scale production and hope for the best. There were two drivers: impatience, and a very real fear of IP leakage. Without strong intellectual property protections, companies were reluctant to expose their processes even to a trusted R&D partner like us.

That has genuinely changed. There is now a much deeper respect for IP in Indian industry and a corresponding willingness to go through the full process - lab, pilot, scale-up – properly, without cutting corners. Companies understand that the pilot stage is where you earn the technology. Skipping it is skipping the learning. We’ve also seen our client mix shift. We used to do almost 100% of our business with public sector oil companies. That percentage has come down significantly as private sector clients have grown more sophisticated in their R&D approach.

Have you ever had to shut down an innovative project to protect the business? What happened and what did you take away from it?

Yes, and it’s a story I think about often. About ten years ago, we launched into the analyser business. We had what we genuinely believed was a breakthrough product: it worked on the principle of dry chemistry, as opposed to the wet chemistry that dominated the market at the time. Results came in half the time, making this significantly more cost-effective. The market’s reaction was quite literally electric: orders poured in the moment we started the launch process.

But there were niggles. Small problems erupted but we told ourselves we’d solve on the go. We thought the excitement justified moving fast. We were wrong.

Once we had multiple installations in the field, we realised just how maintenance-intensive the product was. We had to have an engineer physically present at sites almost constantly. This was unsustainable, both operationally and economically. We ultimately had to recall the products and shut down the entire vertical.

The lesson: let the full lifecycle testing run its course before you launch. Market excitement is not a substitute for engineering rigour. We learned that the hard way.

Many leaders talk about innovation with accountability as if it’s obvious. What does this actually look like inside Xytel?

Growth needs direction and direction requires monitoring. That’s the foundation. At Xytel, we run what we call ‘vision meetings’ every month. Every business leader presents the story of their innovation using a fixed template, no exceptions. The template comprises of the following:

• What is the vision for this innovation?

• How aligned is it with the organisation’s vision?

• What specific objectives will realise this vision?

• What initiatives are being pursued to hit those objectives?

• What KPIs are tracking progress on each initiative?

The template creates a common language across very different types of innovation. I’m not always looking for equal progress across the board, since some projects move faster than others by nature. What I’m looking for is sustained progress. That’s the signal I need.

How has geopolitical turbulence changed the way you think about supply chains, both in terms of efficiency and resilience?

This is something that is changing every day. The first time this really hit us hard was the semiconductor crisis. Every product, every sector was affected simultaneously in ways we simply had not anticipated. There was a running joke internally: if a delivery was late, the reason was ‘the chocolate chip problem.’ That’s how pervasive it had become. That crisis fundamentally rewired how we think. Before, our supply chain mantra was simple: best price, best quality, good lead time. Resilience wasn’t even part of the conversation. Now it’s a fourth pillar and in some ways, it overrides the first three.

Have you consciously accepted either higher costs or lower efficiency to reduce geopolitical exposure?

On cost — yes. On efficiency — not so much, at least not yet. Developing new vendors costs real money. You must qualify them, build the relationship, test them. But we’ve made a deliberate decision to expand our vendor base significantly, even at the cost of margin. That’s the price of resilience.

What risks feel most acute today compared to 5 years ago?

Two stand out. The first is currency risk. We import and we export, which means we’re exposed in both directions simultaneously. Managing that requires very tight coordination between our supply chain function and our treasury: constant hedging, careful payment timing, relationship management with counterparties.

The second – and this one caught us off-guard – is logistics. About three years ago, a critical shipment got stuck at a port due to a disruption we had not modelled. For a product like ours, extended time at a seaport can cause real physical damage. That incident changed our approach entirely. Every logistics decision now includes a geopolitical risk assessment: is there any active disturbance along this route?

Where has AI delivered real value for Xytel, beyond pilots and proof-of-concepts?

In our industry, the most relevant form of AI is simulation software. For years, there’s been a claim that simulations could eventually replace pilot plants entirely; that you could model the chemistry computationally and skip the physical step. Fortunately for us, that has not happened. And the reason is interesting: the behaviour of chemical processes at smaller scales isn’t understood well enough to be reliably simulated. Our physical pilot plants remain indispensable precisely because the small-scale dynamics are too complex and unpredictable for current AI to model accurately. At the larger industry scale, simulation software is in wide use. But at our scale – the pilot scale, the messy middle between lab and full production – it’s not there yet.

As AI moves deeper into decision-making, how do you ensure human judgment and ethics remain central?

AI is remarkably good at removing human bias from decisions. But it can also be brutally inhuman. It does not naturally weigh environmental considerations, social consequences or the texture of a relationship. Those things don’t show up in the training data the way hard metrics do. Whenever we implement an AI initiative, I stay in close contact with the people affected by it. I watch for the moments where the algorithm would make a decision that a thoughtful human wouldn’t, and I ensure there’s a mechanism for intervention at exactly those points.

How do Indian MSMEs compete with MNCs on AI adoption when the capital requirements are so different?

I think the gap is less about capital than it is about mindset, and that’s actually both a problem and an opportunity. In Western manufacturing contexts, impermanence is built into the employment relationship. An employee signs on knowing they might leave in a year. That creates a certain psychological readiness to embrace change, including AI-driven change.

In India, especially in manufacturing, people join with the intention to stay. That loyalty is a genuine strength. Most of our employees have been with us for 10-15 years. But it also creates a reluctance to embrace disruption. When I ask line managers about AI implementation, their instinct is to protect the people working under them. That’s not cynicism, it’s loyalty. It’s the same instinct that makes them great managers.

Quote – ‘Rather than mandate AI from the top, Xytel invests heavily in re-skilling workers whose roles are affected by automation — creating a pathway that reduces fear and builds genuine acceptance before implementation.’

So, we do a lot of work on re-skilling. We create genuine alternative pathways for people whose roles evolve. And it works — we’ve seen it work for 30 years. When computers arrived, we had rooms full of secretaries and documentation staff. That world changed completely and our people adapted. The same will happen with AI. The fear is real, but so is the human capacity to adapt.

The chemical industry spans wildly different sub-sectors. How do you manage the talent dimension of that complexity?

It’s genuinely one of our hardest challenges. The skills required for, say, lubrication oil work don’t transfer cleanly to another petroleum application as the specificity is real and we can’t paper over it. We navigate it through deep investment in cross-training, through deliberate project rotation where possible and through acknowledging honestly where a particular expertise requires dedicated focus.

The advantage we have is that our people stay. Long tenures mean accumulated expertise that doesn’t walk out the door. The challenge is keeping that expertise current in a space that’s evolving as fast as the policy environment and global supply chains are moving.

Uncertainty has become the baseline condition for CEOs. How has that changed the way you lead?

I take guidance from my spiritual teacher on this. He describes the phase between two periods of growth as a ‘transition phase’ and he says that all stress is essentially our inability to adapt to transitions. Once the transition resolves, you find yourself back in a comfortable state. That’s always been true.

What’s different now is the frequency of change. Earlier, transitions used to come every few years. Now they come constantly. And here’s what I’ve realised: the ability to navigate transitions is a skill. Like any skill, it atrophies if you don’t use it. A prolonged period of stability would worry me now as I’d be concerned that we were losing the muscle.

If you were advising other CEOs scaling their businesses today, what’s one leadership instinct to strengthen, one to unlearn?

On strengthening: most leaders have a role model, a successful CEO or company they admire. And they study what that leader is doing right now. That’s the mistake. What you should be studying is what they did when they were at your size. What changes did they make to their thinking, their systems, their processes, their strategies at that stage of the journey? That’s the relevant information. By the time a leader has earned their stripes, their whole modus operandi has changed. What they do now is the output of a journey, not a template you can paste onto your own.

On unlearning: let go of the one-person show. When you’re small, wearing every hat is not just acceptable, it’s necessary. But at scale, the roles of vision setter and vision implementer must be separated. You cannot be both. The moment you try, you lose the feedback loops that tell you whether your vision is working in the field. Decentralise. Hire strong people. Trust them with real responsibility.