Bootstrapping AI innovation: navigating the trade-offs

In August 2024 Mohit Modi, Head of Strategy and Corporate Development at Crisil, sent his weekly brief to the CEO with mixed emotions. Crisil’s first Gen AI pilot was officially a success and was ready for deployment. Months of effort and hard work had paid off; the system was live, reliable and built entirely in house. The CEO had closely followed the team’s entrepreneurial journey and immediately responded: “This is a great start, Mohit, congratulations to the entire team! The question now is how do we pace up our innovation efforts as we want to take on more AI opportunities?”At the next functional review the CEO put this question more formally to Chief Technology Officer Zak Murad and Head of AI services YD Singh: “Can you prepare a plan to increase the velocity of AI / Gen AI innovation across the company?” Crisil had been using AI/ML in its internal tools for many years, but in a calibrated manner. The trust of its stakeholders had been central to its reputation and to its purpose of making markets function better. The success of Gen AI in credit memo had validated one aspiration, proving that Gen AI innovation was possible within Crisil’s disciplined environment. But it also set a new challenge: making it faster.