Summary
In this episode of the Data Reliability Experience, we are joined by Ulas Kumar, co-founder and CEO of Gritworks, along with Nachiket (Lead Sales Engineer at iceDQ) and Adnan (Director of Sales and Go-to-Market at iceDQ), for a deep dive into the unstructured data goldmine that enterprises are sitting on, and what it actually takes to make it safe, usable, and trustworthy.
Drawing from Ulas's experience building Gritworks to solve the complex, sensitive data challenges facing regulated enterprises in banking, insurance, and healthcare, and Nachiket and Adnan's front-line view of how enterprises manage data at scale, the discussion explores why most organizations are only using a fraction of what they've collected, and why that has to change as AI moves into production.
Key Topics Covered
- Why unstructured data (contracts, scanned documents, call recordings, emails, PDFs) remains the largest untapped asset inside most enterprises
- The five steps required to use unstructured data well: access it safely, understand it, legally use it, trust it, and scale it
- Why converting a PDF to text does not make data "structured" or usable
- How structured data gives you the "what," while unstructured data gives you the "why" behind every decision
- The real compliance risk of getting this wrong, including how a single HIPAA breach can cost $10 million
- Why sensitive data needs to be sanitized at the source, not downstream
- The "teenager's bedroom" analogy for the two types of unstructured data organizations typically encounter, and how to bring order to each
- Why industries under the most regulatory pressure (banking, insurance, healthcare) are furthest ahead on unstructured data maturity
- Why AI agents need the "why," not just the "what," to make trustworthy decisions, and what that means for data governance
- Why building a "trust pipeline" for unstructured data, not just an access pipeline, is the next frontier for data engineering teams
About the Guest – Ulas Kumar
Ulas Kumar is the co-founder and CEO of Gritworks, where he helps enterprises make sensitive and unstructured data, including documents, images, audio, and other complex data types, safe and usable for AI, testing, and analytics. He brings over 16 years of experience across technology, strategy, and operations roles at Google, Infosys, and LEAD Group, where he led large-scale operations spanning more than 3,000 schools. Earlier in his career, Ulas was part of the founding team at Meghshala, a company recognized with the National Award for Social Innovation from the Prime Minister of India. Gritworks' focus is squarely on the data challenges faced by regulated enterprises, particularly around privacy, secure data access, and reliability as AI moves into production.