Message from co-founders
We believe the ability to solve these problems will distinguish the companies that create real value from AI. The last year gave us greater confidence in our ability to do exactly that.
Our results reflected this progress.
to ₹32,997 million, organically
to ₹4,899 million
Profit after tax growth
to ₹2,868 million
Our Net Promoter Score was 78 for the year and reached 81 in the fourth quarter, one of the highest we have recorded. We increased our investment in R&D to 6.4% of revenue.
These numbers are important. The trust behind them matters even more.
More than 100 of the world's largest enterprises work with Fractal. Our ambition is to become increasingly central to how they use AI to run and transform their businesses.
Our R&D is producing AI with increasingly advanced capabilities. Vaidya.ai (Cogentiq Health), our healthcare AI, can reason through complex medical questions and has demonstrated world-leading performance on difficult medical reasoning. Fathom Deep Research can investigate complex questions, reason across large amounts of information and produce evidence-backed answers. PiEvolve, our AI for data science and machine learning, can autonomously test and improve solutions to difficult machine-learning problems.
We are also turning these capabilities into products that solve real business problems. Asper.ai is helping consumer goods companies bring pricing, promotions and demand planning together. Flyfish.ai uses a team of AI agents to perform sales work.
Underlying much of what we are building is a simple principle: advances across frontier and open models should make our products and solutions better as well. Our role is to make these models useful inside an enterprise by connecting them with the company's knowledge and data, giving them the tools and controls to act, and making them reliable enough to run real business processes.
The stronger the models become, the more we can build on top of them.
There is another change underway that we believe will shape the next decade.
For several years, the economics of AI have been discussed largely in terms of cost, particularly the cost of running AI models. That cost has been falling rapidly. As it falls, the more interesting question is the value AI can create. The conversation is shifting from cost per token to value per token.
The first wave of enterprise AI has focused heavily on efficiency: using AI to perform existing work faster and at lower cost. There is enormous value in that.
The larger opportunity comes from being able to do things that were previously too expensive, too slow or simply impractical. As the cost of intelligence falls, companies can apply it to many more decisions, more often and at far greater scale.
This is why we believe AI can reduce the effort required for existing work while dramatically expanding the amount of work worth doing.
Some tasks will take a fraction of the effort they take today. Some existing work will disappear. At the same time, thousands of new applications of intelligence will become economically viable.
We are already seeing this in our conversations with clients.
AI has moved into core enterprise budgets. Programs are getting broader. Clients increasingly want to redesign complete business processes. They want greater control over their AI and their proprietary knowledge. They want the flexibility to use different models. They increasingly want partners who will stand behind business outcomes.
