AI has officially moved beyond experiments and into real-world production — making decisions, taking actions, and reshaping how businesses operate.

But there’s a problem.

Even as adoption accelerates, most organizations still haven’t fixed the one thing AI depends on: their data.

In this episode of Today in Tech, Keith Shaw speaks with Gaurav Pathak, Senior Vice President of Product Management for Metadata and AI at Salesforce, about the growing “trust paradox” — where employees increasingly trust AI outputs, while leaders know their data foundations aren’t ready.

They break down:

Why AI adoption jumped from 48% to 69% in just one year

How agentic AI is spreading faster than any enterprise tech before it

Why data quality is still the #1 blocker to real AI deployment

The risks of “shadow AI” inside organizations

What leaders must fix now before AI mistakes hit customers and revenue

As AI systems become more persuasive — and more autonomous — the cost of bad data is no longer theoretical.