The DAIS 2026 Speaker Spotlight is a series where we hand the mic to the speakers heading to Data + AI Summit and let them answer five short questions โ in their own voice, no press-release polish.
Below, Tushar Madan on apps, agents, and what really gets in the way at enterprise scale. Lightly edited for length โ otherwise, the words are his.
โ
A year ago the hard part of a data app was building it. Today, AI-assisted coding has flipped that โ building is fast. The hard part is everything after.
โ Tushar Madan
The topic
What is your talk about, and who is it for?
For anyone trying to ship a data app or agent to real users at enterprise scale: what actually stops these projects at the prototype stage, and the small set of decisions that get them through.
Why this, why now
What's changed in the last 6โ12 months that makes this topic urgent right now?
A year ago the hard part of a data app was building it. Today, AI-assisted coding has flipped that โ building is fast. The hard part is everything after: authentication, the data path, scale, the operational layer most teams discover only when something breaks at 2am. Most companies have shipped a first app and stalled on the second. The platform pieces that actually fix that just landed. This is the right moment for this conversation.
The personal stake
Why are you the person giving this talk?
I came to product from the field. In my last role I led customers through their first real data wins โ and apps were the surface that closed the loop. They were the thing that pulled ingestion, models, and activation into a single workflow that got the right decision to the right person at the right time. This talk comes from watching that work, repeatedly, across very different industries.
What you'll leave with
What will someone be able to do on Monday morning that they couldn't do before?
You'll leave knowing the most common ways app and agent prototypes fall apart on the way to production โ authentication trouble, data path fragility, the operational layer no one planned for, the second-app stall. None of these are rare. All of them are avoidable. If you're trying to get something past the demo phase and into the hands of real users, this is the session to come to.
The bigger picture
How does this fit into where Databricks โ and data and AI more broadly โ is heading?
Every interesting thing happening in data and AI right now ends with a person using something โ a dashboard, an agent, an app. That's where value actually gets delivered, and it's the layer most companies have under-invested in. The substrate to fix that is finally landing: Apps for the interface, Lakebase for the data, agents that hand off across the platform. This talk is about being ready when the rest of the stack catches up.
Part of the DAIS 2026 Speaker Spotlight series โ more voices dropping in the weeks ahead. Got a DAIS speaker you'd love to hear from next? Mention them in the comments โ we're always listening.
