Building dashboards with AI

1. Building dashboards with AI

You've spent the last two chapters cleaning, enriching, and finding insights in your data. The next step is making sure those insights actually reach the people who can act on them — through dashboards, and through data stories. Let's start with dashboards.

2. What is a dashboard?

A dashboard is a visual display of data used to monitor conditions and/or facilitate understanding. It's an intentionally wide definition —

3. What is a dashboard?

dashboards come in many shapes, from patient monitoring to race-team performance to the Economist's Big Mac index. Whatever form they take, the work of building one well has always meant slow, hands-on iteration with users. AI is transforming that.

4. Where AI plays... and where it doesn't

A complete dashboard lifecycle has four phases: discovery, prototyping, development, and deployment. AI is genuinely useful in the first three — accelerating the parts that always depended on manual iteration with users. Deployment is different. It still needs a BI platform for versioning, data pipelines, and governance. The rest of this lesson zooms in on each of those first three — discovery, prototyping, and development.

5. Discovery: from interviews to user stories

Let's say we've just finished a round of user interviews. We've got recordings, transcripts, and a stack of notes. The old way: read through everything, write up user stories — tot-up each group's pain points, goals, and success metrics. The AI way: we can upload the transcripts into Claude and ask it to extract those user stories for us. A few minutes later, we've got a structured user requirements document we can refine and share.

6. From user stories to working prototypes

Now we've got user stories. Time to turn them into something users can react to. We give Claude the user stories and our dataset and ask for a low-fidelity wireframe — just enough to help your users to agree on data appl structure. The layout it produces flows from headline KPIs at the top down to granular detail on demand. Once that lands, we ask for a higher-fidelity interactive prototype — one users can actually click through, before we commit to any production work.

7. Iterate in hours, not weeks

AI really earns its place here. Show a wireframe to a user, get feedback, ask AI to revise it, show it again. That cycle used to take a week of design iteration, and now can run in just an afternoon. The dashboard you finally build is one your users have actually shaped — not the one you blindly hoped would land.

8. AI designs, BI runs

Once you have a validated prototype, deployment moves to your BI platform. That's where versioning, data pipelines, role-based access, and corporate governance live. AI handles the design; your BI platform handles the running.

9. Let's practice!

Time to take some interview notes and turn them into something you can show a user.

Create Your Free Account

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.