How the best data teams use the Notebook Agent

Three months ago, we launched the Notebook Agent. Since then, tens of thousands of people have used the agent to build analyses and transform insights into influential artifacts — with thousands of messages exchanged every single day.

The feedback has been incredible. Agentic analytics has become a core part of many data team’s workflows, and we're seeing some of the world's most innovative companies work with data in new ways.

We want to share the magic we're seeing, so keep reading for an inside look at how Ramp, Figma, and Notion are using AI agents in Hex to supercharge analytics.

Ramp: Using data to make the business better every day

The team at Ramp lives by the mentality: “AI or die.” When a PM asked Ricky Meyers, an Analytics Engineering Manager at Ramp, to track the performance of their auto-routing feature, he spun up a dashboard in minutes with his AI analytics tool of choice: the Notebook Agent.

Instead of bouncing between Hex and ChatGPT or Claude, Ricky stayed in his Hex notebook, tagged a relevant SQL cell, and asked the agent to plan and build the analysis.

The agent returned a clear step-by-step analysis framework and a list of success metrics to track. Ricky then asked it to visualize those metrics, offloading the tedious dashboarding work to the agent. In under a minute, he got polished charts without touching a chart builder!

Distribution is just as important as analysis, and the Notebook Agent gives Ricky a leg up by giving him a summary and interpretation of the results he can easily share with his stakeholders.

As Ramp’s CFO once said, “Product wants data to tell them what to do more and faster.” At Ramp, data’s role is to make the business and product 1% better every day — and the Notebook Agent helps make that possible.

Figma: Building a world-class product insights program

At Figma, the Notebook Agent is empowering teams outside of the core data org to build and scale reporting.

When AI Insights Researcher, Rie McGwier, joined, they needed to transform millions of survey responses into insights the entire company could use. Over the last year, they’ve built world-class NPS and Product Health programs — all in Hex, turning messy raw data into clean, actionable insights.

Self-taught in SQL and Python, Rie used the Notebook Agent to level up quickly, cutting a weeks-long learning curve down to hours. With less time spent debugging syntax, they could focus on building a strong, scalable program.

As the survey program matured, the data became more complex. Simple fields like signup date became calculated metrics that needed consistent logic across teams. Rie used the agent to build reusable components — one of which now powers more than ten dashboards across the org.

This foundation also helped Senior Researcher, Molly Jane Nicolas, speed up report development. With the agent handling tedious tasks like writing queries and building charts, she focuses on the more strategic (and fun) parts, like storytelling. In just 20 minutes, she can launch a survey, analyze feedback, and build dashboards — a process that used to take days.

Molly's reports are now essential to Figma’s business reviews. And the team's ambitions don’t stop there; next up is enabling conversational self-service for less technical stakeholders with Threads, while Rie and Molly keep pushing what's possible with the Notebook Agent.

Notion: AI as a collaborator and accelerant for analytics

The Notion data team is no stranger to AI; they regularly use a rolodex of tools like Claude and Cursor. But when it comes to analyzing data and building shareable reports, Hex is their go-to AI analytics solution.

Aks, a Lead Data Scientist at Notion, spends most of his day in Hex working on sales analytics — tracking financial metrics, revenue reporting, and predictive modeling to understand performance and keep customers happy.

He leans on the Notebook Agent to query data, build models, and create visualizations — all within the same workspace where their dashboards and analyses already live.

For Aks, the agent is a collaborator. When he wanted to predict customer expansion, churn, and renewal, the Notebook Agent combined internal and customer behavior data to help him build and iterate on classification models. He often assigns the agent a task, shifts gears to something else, then returns once the work is done.

When he's working outside his usual wheelhouse, the agent helps him explore ideas, suggest approaches, and quickly assess what’s possible.

Notion is always shipping — which means new product metrics are popping up all the time. Aks uses the agent to see what’s out there, so he can dive into the latest product usage data without waiting on teammates to point the way.

Compared to past AI features, Aks says the Notebook Agent feels more intentional. Even his teammates outside of the data team have shared glowing feedback about the way it explains its thinking, shows its work, and delivers answers they can trust.

AI is reshaping how data teams work — the questions they ask, the tools they rely on, and the speed at which they deliver value. At Hex, we can barely remember what life was like before the Notebook Agent. It's become second nature in our own workflows, helping us move faster and focus on higher-leverage work.

Every day, more teams are using the agent to shift from gut-feel decisions to data-driven insights at scale. And this is just the beginning — we can’t wait to see what you build next!