Source: How one designer led an AI revolution at Pendo: The paternity leave epiphany

Brian Greenbaum (product designer at Pendo, a product analytics company) talks about driving AI adoption across their org. What caught my attention: he claims clear rules increase usage, not decrease it.

The argument goes like this. People avoid AI tools when they don’t know what’s allowed. Can I paste this code? What about customer data? The uncertainty creates friction. So he worked with legal and security to create an internal wiki with approved tools and data handling rules.

I find this plausible, though I’d want to see the actual numbers. It matches what I’ve noticed with other “permission” problems: people often want to do the right thing, they just don’t know what it is.

The other piece was visibility. He ran hands-on workshops every two weeks and set up a public Slack channel for sharing experiments. The idea being that secret AI usage creates weird dynamics where people either hoard knowledge or feel embarrassed.

To convince skeptics in leadership, he built an MCP server (a way for AI to connect to external data sources) that could query company data with natural language. Showing beats telling, I suppose.

I should note: this is one company’s story. Pendo is a tech company with tech-savvy employees. The “just create a wiki” approach assumes people actually read documentation (they often don’t). The workshops require someone with bandwidth to run them. And in fear-based cultures, public Slack channels become performative rather than authentic.