What happened at Prompt 2026

“We should do a conference this year” is somehow the most energizing and terrifying sentence.

With a brand as distinct as Hex’s, we knew we needed to build a space to explore the next age of data and analytics that felt nothing like every other conference, and everything like Hex.

It all went down at the historic San Francisco Mint on Oct. 27 with 250 Hex customers. The result was earnest, useful, and engaging content. The storyline throughout the entire day boiled down to this: Humans are still very much the hero in this AI + Data era. We can’t shy away from this; we must seize the opportunity in front of us all.

Here’s what went down.

Introducing Hex 4.0

At Prompt, we introduced Hex 4.0, a reimagined platform for humans and agents to work on data together. Ask questions and build apps from Hex or from the agents you already use, like Claude, Codex, or Cursor, then publish them securely on your trusted data. Click any number in an app to see exactly where it came from. And our all-new execution engine runs apps up to 12x faster on our own internal dashboards.

Hex 4.0 is in private beta today, with broader availability rolling out over the coming weeks. Your existing projects will keep working. Read the launch post →

“There’s never been a better time to be a data person”

Meghana Reddy and Nova Wang from StubHub kicked off the day with the bold claim that there’s never been a better time to be a data person. The highs and lows over the past years, as AI began doing more and more in their workflows, meant owning a new role in the data work they love so much.

Ultimately, they tell us we should focus on being excellent at systems thinking and translating our knowledge and skills. Not to give away what makes us special, but to scale what we do and know.

Strong agree. And it’s a sentiment we heard over and over again from our speakers. Jordan Farrer, Dori Wilson, and Connor Macdonald from Chime® emphasized the role of humans in creating meaningful context for agents. It’s created an entirely new role at Chime® and connected data teams with business users more than ever.

We’re breaking all the tradeoffs

These times are changing, and several of our speakers discussed trade-offs they no longer have to make. They’ve created new operating models to reposition themselves in this new era. Mariam Melikadze at The New York Times shared her decision framework for managing the onslaught of new requests and deciding when AI should lead vs. a human. I appreciated her matrix that weighs how reversible something is against how impactful it would be. She’s sending non-reversible, highly impactful decisions straight to her team.

Oliver Gomes’ story of rebuilding the data team at the NHL to reduce the noise from 4 hand-offs to one cohesive data pod is another way teams are reorienting themselves around AI. This allowed his team to develop an entire operating model that other teams in the league can adopt. More data in more places answering deeper questions. We love to see it.

Other speakers like Ian Macomber at Ramp, Sean Hunter at Fox, and Devon Mittow at Lyft spoke more about the tech stack you have to build to enable this rapid pace. Because expectations aren’t slowing down, it takes new systems to enable fast and accurate answers.

We loved hosting round table and workshop breakouts

We hosted two roundtables that allowed our customers to discuss topics without an agreed-upon “right” answer. The “Directives without direction” roundtable compared notes on what happens after leadership tells everyone to automate everything. Most attendees said access and budget aren’t the blockers anymore. The real friction is knowledge stuck in team silos, and they are figuring out how to scale that with AI.

The second roundtable paired business users with data practitioners to explore what the next version of their partnership looks like in the age of AI. The biggest theme from our roundtable: business stakeholders and data teams want to work more closely, but nobody wants to become the bottleneck or give up control. Everyone agreed it should be simpler; nobody has cracked it yet, and we’re taking that as a challenge.

Focused workshops on evals and Generative Apps

We also ran two hands-on workshops. In the evals workshop, teams used the Hex CLI to build and run their own eval suites (tests that score how well the agent answers questions), then improved their context each round to raise those scores.

In the generative apps workshop, we handed teams a deliberately boring data app and asked them to use the Hex agent to redesign it and add a new workflow, because not everything needs to be a dashboard. It was, unsurprisingly, a lot of fun.

The future is bright

The day comes down to one word: optimism. The stories shared during the panel discussion led by Katie Bauer, Hex’s Head of Data, painted a picture of what a data team can and should look like when self-service finally clicks.

Can’t wait to see you next year!

Till then,

Rachel Herrera 🔮