Build vs Buy, Acquisition Integration, and Launching Agentic AI in Construction | Rajitha Chaparala

Rajitha Chaparala has spent 20 years building data and AI products in some of the world's most heavily regulated industries. She founded Intralinks' AI Center of Excellence, scaled the data platform at ZoomInfo through two simultaneous acquisitions delivering eight figures of incremental revenue in year one, and has spent the last four years as VP of Product for Data and AI at Procore, bringing AI into one of the most document-heavy, safety-critical industries in the world.In this episode, we cover:How she built AI for M&A due diligence when she couldn't access actual customer documents, and why crowdsourcing training data on Upwork was the only path forwardThe redaction product that emerged as a happy accident from the same NLP foundation, and what that taught her about building on existing platformsThe five principles she used to run two massive parallel workstreams during ZoomInfo's acquisition integration, and why clarity of vision matters more than process when you're moving fastThe build vs buy decision at every company she's been at, and why the answer has been different every timeWhy human-in-the-loop isn't a limitation in construction AI, it's the right design for where accuracy and governance currently standWhy she hires for adaptability above everything else, and what breaks when you don'tA conversation for data and AI product leaders who want a grounded view of what it takes to build AI products in industries that don't forgive mistakes.