Everyone is talking about AI governance. Almost nobody is talking about the part that actually decides whether it works. I had a blast chatting with Gaurav Bhandari, AVP and Head of Data and Analytics consulting at Infosys, on The Ravit Show at Data Citizens on the Road by Collibra. One line stuck with me. Roughly 80% of AI governance is just governing the data that feeds your models. We have been here before. Data governance started as a compliance and privacy problem in regulated industries. Then data became the asset everyone wanted to mine for value. Now AI has raised the stakes again, because a model is only as good as the context behind it.
Gaurav broke that context down into five things every enterprise has to get right:
- Trust. Can you rely on the output.
- Ethics. Even when you trust it, is it the right answer to put in front of people.
- Regulations. Are you staying compliant as the rules keep shifting.
- Privacy. Do people still control their own data.
- Security. Is everything safe once it sits inside your workflow.
Miss one of these and your AI agents are running on shaky ground.
What stood out to me was how the Infosys and Collibra partnership fits this moment. Ten plus years working together, and not just in finance. Retail, manufacturing, life sciences too. Collibra brings the platform. Infosys weaves the policies, controls, and structure into one governance story instead of a pile of disconnected tools.
His advice for the next 12 months was refreshingly simple. Stop thinking about data governance. Start building data plus AI governance.
The companies that treat these as one problem will move faster than the ones still treating them as two.
Full interview is live now.
Follow The Ravit Show for more conversations from across the Data and AI world, and subscribe to the newsletter to stay ahead.
#data #ai #collibra #governance #infosys #api #datacitizen #theravitshow
