6 ways to save time with AI in Flourish

A practical look at the Flourish Assistant and Connector — from data wrangling to custom popup design

Most of the time you spend on a data visualization isn’t spent on the chart. It’s spent on everything around it: cleaning data before you can import it, rebuilding from scratch when the chart type doesn’t fit, manually updating three embeds when new numbers land, wrestling with settings to get a tooltip to look right. The Flourish Assistant and Connector are built to handle all of that for you.

The Flourish Connector builds charts directly from conversations in Claude and other AI tools. The Flourish Assistant refines them from inside the Flourish editor.

In this post we’ll show you what that looks like across six real workflows. From a single prompt that wrangles messy data and builds a chart at the same time, to a one-sentence instruction that designs a popup with custom HTML and CSS.

Flourish Connector

1. Build a branded chart from messy data in one prompt

Imagine you have revenue data for Deloitte, PwC, EY, and KPMG, but each firm uses different names for the same service lines. What Deloitte calls “Audit & Assurance”, PwC calls “Assurance”, and KPMG simply calls “Audit”. The tax and advisory lines have the same problem. Before you can chart anything, you need to standardize the naming across all four firms. What used to mean a detour into a spreadsheet, can now be solved in one prompt.

I have revenue data for Deloitte, PwC, EY, and KPMG but each firm uses different names for their service lines. Standardize them into four categories — Audit & Assurance, Tax & Legal, Advisory, and Consulting — then create a Flourish grouped column chart.

What used to take two tools and twenty minutes — clean the data, then build the chart — now happens in a single conversation and one prompt.

Once your data is standardized, the Connector can just as easily turn it into a full sequence of charts — see our guide to building an interactive data story in minutes.

Flourish Assistant

2. Find the one story hiding in fifteen lines

You have a chart covering European inflation, spanning 15 countries and seven years of data. The story is in there — you’re just not sure anyone will find it in a chart with 15 overlapping lines in similar colors.

Instead of searching for the right settings to make your chart more legible, you type one prompt into the Assistant.

Highlight Poland's line. Set all other country lines to gray.


A chart that could have taken several rounds of manual edits to get right is ready to publish from a six-word instruction.

For more on making deliberate choices like this, see our guide on how to improve your chart colors.

Flourish Connector

3. Generate multiple chart options from one prompt

The client presentation is tomorrow. You have data on digital ad spend and internet penetration for 30 countries, and you know a scatter plot, a world map, and a bar chart would each tell a slightly different story. The problem is you don’t know which one lands best until you see them — and you don’t have time to build all three from scratch.

You drop the dataset into Claude and ask the Connector to build them all.

I have a dataset of digital ad spend and internet penetration for 30 countries. Create three Flourish charts: a scatter plot with internet penetration on the x-axis and ad spend on the y-axis, a world choropleth shading countries by ad spend, and a bar chart ranking the top 15 countries by ad spend.
A Claude conversation confirming three Flourish charts have been built from one dataset — a scatter plot, a world choropleth map, and a ranked bar chart — with links to each

Three chart types compared and a decision made in the time it used to take to build one.

For more on comparing chart types this way, see our guide to exploring chart options with AI and Flourish.

Flourish Assistant

4. Add annotations in one prompt

A chart can show what happened. But without context, it can’t explain why. Below, we see an area chart of UK energy bills from 2019 to 2024. The spike in 2022 is unmissable — but a reader without context won’t know what caused it.

To fix this, all you need to do is open the chart in Flourish and type one prompt into the Assistant.

Add a vertical line highlight at Q1 2022 labeled 'Russia invades Ukraine — wholesale gas prices surge' in #808080


The annotation takes a few seconds to add. The context it provides would have taken a paragraph to write.

See our guide to highlighting key data moments for more ways to add line highlights, range highlights, and per-facet annotations.

Flourish Connector

5. Update three charts in one conversation

Imagine you publish a monthly client report with three embedded charts: brand awareness, channel reach, and share of voice. Every month, updating them means opening each chart separately and re-uploading the data one by one. It’s not hard, but it takes time you don’t always have.

The new numbers are in. Instead of going into every chart manually, you paste all three chart IDs into Claude together with the new data.

Here is the January 2025 data for this month's report. Please update all three charts: brand health trends [ID], channel reach [ID], share of voice [ID].

Three charts updated in the time it used to take to update one.

Flourish Assistant

6. Design a custom popup without writing code

Your Premier League scatter plot is ready — goals versus assists for 49 players. Hover any dot and you already get a popup detailing player stats. But it’s the default Flourish layout: unformatted, no hierarchy, no sense of which numbers matter most. Making it look like a proper player card with name and club prominent, goals and assists as the headline stats, and supporting metrics below used to mean writing custom HTML and CSS yourself.

Now, you type one prompt into the Flourish Assistant.

Design a custom popup showing player name and club at the top, with goals and assists as the main highlighted stats (like in a dashboard), and shots on target, goals per 90, and minutes played as supporting stats below. Maintain visual hierarchy and use colors that are not used in the visualization.

The data was always there. Now it’s presented in a way that’s actually easy to read — and a popup that would have needed a developer to design properly takes one sentence.

See our guide to designing your ideal popup in 4 prompts for more ways to reshape the default layout.

Try it today

Your workflow will look different from the ones in this post. You might use the Connector for the build and the Assistant for the popup, or skip straight to updating a chart you made six months ago. What stays the same is the principle: describe what you want in plain language, and Flourish handles the rest.

Both are ready when you are. The Flourish Assistant is live in the editor. To try the Connector, find out how to connect your favorite AI tool such as Claude or ChatGPT.