When Notion AI launched just weeks after ChatGPT hit the scene, it wasn’t a fluke. It was the result of deliberate choices, scrappy execution, and a brutally honest understanding of what AI could—and couldn’t—do.
In this episode of Next Gen Builders, host joins —former Head of Data at and —for a conversation about how Notion AI came to life in record time, what the role of a modern data leader really looks like, and common traps to avoid when building with AI.
Listen to the episode on , , or .
Being a modern data leader
Data work often falls into two categories: enabling the business to answer key questions and using data to power product experiences. The first often prompts data leaders to ask whether their teams should operate as centralized service centers—with ticketing systems, intake forms, and request queues.
Daniel doesn’t think so. He embeds his teams within product and marketing, so they can partner closely with stakeholders to build context, trust, and faster feedback loops. That said, he’s clear-eyed about the nature of his role as the data leader, which often means acting in service of other teams.
“As a data leader, I’ve come to terms with the fact that this is a service role. My success depends on whether my stakeholders have what they need to make better decisions.”
Building Notion AI
Notion’s AI journey didn’t start with a mandated strategy—it started with curiosity. In the summer of 2022, one of the company’s co-founders began experimenting with early LLMs (large language models). Following a company retreat, the founders built a working demo and shared it with the whole team.
That initial excitement quickly turned into a formal tiger team with members borrowed from across the organization. The goal was to move fast with a small, scrappy team that could be disruptive and figure things out.
“Notion prides itself as a company on being fast to market,” Daniel says. “It’s the CEO’s job to ask, ‘Why can’t we ship faster?’ And our job is to answer that with tradeoffs—not excuses.”
Notion launched a public waitlist shortly after ChatGPT’s release. By the time Notion AI became generally available, millions had signed up and a paid add-on was quickly generating revenue.
Staying ahead of the competition
While many companies were still figuring out their AI strategy, Notion was already working on their next set of AI features. That speed didn’t come from perfect infrastructure—it came from pragmatism.
Daniel and team built with the belief that the product experience is paramount, and everything else can come later.
So they focused on:
- Prioritizing product value over technical purity
- Using the product internally to test the experience
- Piloting with customers to gather early feedback
- Launching fast with core use cases
- Learning and iterating based on the results
“Being first to market or really early to market in AI is an advantage for two reasons. One, you build buzz. Two, you get more shots on goal in terms of understanding what works for users and what doesn’t.”
Navigating the AI hype
“You should push technology to the frontier of what it’s capable of,” Daniel says. “But you need to be realistic about what that frontier is.” Knowing the difference between eventually possible and actually doable now is one of Daniel’s biggest lessons from building with AI.
He recalls overestimating what features like RAG (retrieval-augmented generation) could deliver in the early days, which can create a perception of snake oil if you’re not careful. That’s why Daniel recommends testing feasibility before setting launch timelines.
“Get a rough internal version to a really good spot before you promise marketing a ship date and especially before you invest a lot of product engineering time on it.”
Another key learning: AI doesn’t replace structured data. It needs it. Building effective AI features still depends on understanding your data, defining metrics, and building pipelines that make sense for your business.
Balancing AI zealots and skeptics
With any emerging tech, there’s a risk of creating a divide between the AI zealots who believe in everything and the skeptics who trust nothing.
Daniel saw his role as staying in the middle. “You need to maintain a good balance,” he says, “so that even the people who are a bit skeptical can feel safe to voice skepticism and get excited about a cool capability.”
That balance matters, especially to ensure AI is a company-wide mindset instead of the responsibility of one team.
Growing through uncertainty
Daniel closes the conversation with a moment of personal reflection—a time early in his leadership journey when things weren’t going as planned. A key hire backed out. Another team member resigned. And he was about to head out on parental leave with a team of one and no clear path forward.
“It was the first moment since working in tech where I realized things weren’t always going to be up and to the right … But these pivot points are really helpful in forcing you to reframe your approach.”
Instead of powering through with momentum, Daniel hit pause, got honest with his team, and focused on what mattered most: hiring the right people. He reprioritized, leaned into recruiting, and rebuilt the team from scratch.
Because in fast-moving environments like AI and data, clarity doesn’t always come from charging ahead. Sometimes, it comes from stepping back, refocusing, and building again with intention.
Tune in to Daniel’s story
Notion’s AI journey is all about moving fast without losing focus. From staying grounded in what’s technically possible to building with small, scrappy teams, Daniel’s insights offer a real-world look at what it takes to scale AI effectively. Make sure to listen to the wherever you get your podcasts.
Want to make sure you don’t miss an episode of Next Gen Builders? .