If you've been reading tech headlines lately, you'd think AI agents are everywhere, automating workflows and revolutionizing enterprise operations. But according to MongoDB Inc. (MDB) CEO Chirantan Desai, the reality is a lot more mundane: mostly pilots, not much production.

The Gap Between Hype and Reality

Speaking during MongoDB's third-quarter earnings call on Monday, Desai—who took the CEO role less than a month ago—offered a reality check on agentic AI. Yes, everyone's talking about it. No, it's not actually running at scale yet.

After conversations with more than 30 customers spanning large enterprises and AI-native startups, Desai said he hasn't seen "AI agents running in production that fundamentally transform the business or serve customers better." Instead, what he's seeing are lots of pilots—experiments that haven't graduated to full commitment.

The hesitation is particularly pronounced in heavily regulated industries like financial services, healthcare, and the public sector, where companies are experimenting with new AI models but holding back on deployment. "We are fairly early," Desai acknowledged, though he emphasized that MongoDB is positioned to capture that workload once enterprises are ready to make the leap from pilot to production.

Strong Quarter Despite AI's Slow Rollout

While the AI agent revolution may still be waiting in the wings, MongoDB's business isn't. The database platform delivered a impressive third quarter on Monday, posting revenue of $628.31 million—up 19% year-over-year and well above the consensus estimate of $591.52 million. Earnings came in even stronger at $1.32 per share, crushing analyst expectations of $0.80.

The market loved it. After closing down 1.05% on Monday at $328.87, the stock surged 22.84% in pre-market trading following the earnings release. MongoDB scores high on momentum and growth metrics, with favorable price trends across short, medium, and long-term timeframes.

So here's the setup: MongoDB has the infrastructure ready for when enterprises finally move beyond pilots and start deploying AI agents at scale. The question is when, not if, that transition happens. For now, Desai seems content to let the hype cycle run its course while his company builds the foundation for what comes next.