If I could only pick one stock from this entire list, this would probably be it. Persistent has quietly become India’s most impressive AI company — not because of flashy announcements, but because of what’s actually happening in their order books. They’ve trained over 16,000 AI specialists and are winning GenAI contracts in healthcare and banking that their larger rivals are struggling to compete for. A 76.6% stock CAGR over five years isn’t luck — it’s a company that keeps delivering. The growth is real, the margins are holding, and I think there’s still a long runway ahead.

Tata Elxsi is one of those companies that most people outside the tech world haven’t heard of — which is honestly part of the opportunity. They do the kind of deep, specialised AI work that goes into the software running inside your next car, the algorithms helping doctors read MRI scans, and the systems powering broadcast television. A ROCE of 46.6% is frankly extraordinary — most companies would kill for half that. And they carry zero debt. In a world where AI companies often burn cash to grow, Tata Elxsi is profitable, disciplined, and quietly world-class.

Here’s a thought experiment: every time a bank somewhere in the world upgrades its fraud detection, improves its risk models, or modernises its digital banking platform, there’s a decent chance OFSS software is involved. They’re not glamorous — they don’t make headlines — but they print money. A 40.6% EBITDA margin is almost unheard of in enterprise software. They pay a solid dividend. They owe nothing to anyone. I think of this one as the boring, brilliant stock that quietly outperforms while everyone else is chasing the shiny things.

I know what you’re thinking — HCL is too big to be a multibagger. And you might be right in the short term. But here’s why I still have it on this list: when a Fortune 500 company decides to overhaul how its 50,000 employees use AI, they don’t call a startup. They call HCL. The sheer scale of enterprise AI adoption that’s coming over the next three to five years is almost hard to wrap your head around — and HCL is one of a very small number of companies globally that can actually execute on contracts of that size. Steady, dependable, and bigger than most investors give it credit for.

Infosys gets a lot of flak — some of it deserved — but I think the market has been too harsh on what they’ve built with their Topaz AI suite. This isn’t a company bolting “AI” onto old products as a marketing exercise. They’ve built a genuine AI platform that sits on top of Microsoft Azure and AWS, and their global client relationships mean they get called first when enterprises want to move fast on AI. The stock may feel slow, but don’t underestimate the compounding effect of thousands of AI contracts renewing and expanding year after year.

Every time you see an ad on your phone that feels suspiciously relevant to something you were thinking about — there’s a reasonable chance Affle’s tech is behind it. They’ve built something genuinely rare: an AI platform that connects brands to consumers across India’s fragmented mobile internet landscape, with real proof of performance. What I really like though is the financial discipline. No promoter pledge, net cash on the books, and they’ve been quietly buying up small AI startups in Southeast Asia. It’s a growth story with a solid foundation underneath it.

Zensar is the kind of stock that doesn’t show up in most people’s screeners because it’s not a headline-grabber. But that’s exactly why I find it interesting. They’ve carved out a tight focus in insurance and manufacturing — two sectors where AI adoption has barely scratched the surface — and they’ve built delivery expertise that’s genuinely hard to replicate. The stock trades at a more reasonable valuation than most of its peers, which means if they land a couple of big deals, the re-rating could be sharp and fast.

Happiest Minds was built differently from day one — no legacy IT baggage, no decade-old mainframe contracts to protect. The founders deliberately built it to be AI and digital-first, and that shows in the kind of work they win. They focus on product companies — the kind of software businesses that need fast, intelligent engineering — and they’re small enough that one genuinely transformational client can change their trajectory. It’s a higher-risk pick, I’ll be straight with you about that, but the potential upside if they keep executing is significant.

I love Bosch as an AI pick for a simple reason: most people don’t think of it as an AI company at all. But walk into a modern car factory and you’ll find Bosch systems predicting equipment failures before they happen. Look at the software stack in a new electric vehicle and Bosch’s fingerprints are all over it. As India’s manufacturing sector finally starts to modernise — and it is, faster than most people realise — the demand for intelligent industrial systems is going to explode. Bosch is positioned perfectly for that, and the backing of its German parent gives it engineering depth that no Indian startup can match.

Saksoft is my wildcard, and I’ll tell you exactly why I included it. Everyone is obsessed with the shiny AI models and the large language models and the generative AI applications. But almost nobody is talking about the unglamorous, absolutely essential work of getting data clean and organised enough for AI to actually use. That’s what Saksoft does. And as more Indian companies try to implement AI and discover their data is a complete mess, the companies that can fix that problem become incredibly valuable. Small cap, under the radar, and potentially sitting on a very big opportunity.