2026-08-18 01:55:48
This post is part of the SWD + AI series—practical guidance for using AI as a thought partner across the various stages of your data storytelling work. If you’re just joining us, start with the first two installments: start with context and craft a story. Explore all of our AI resources.
Now that we have the story planned, it’s time to start developing the content that will support our message and narrative. When data is part of that, a good first step is choosing a visual that aids in comprehension. The right graph makes your point immediately clear. The wrong one makes your audience spend their mental energy decoding the graph instead of understanding your message.
This is where people sometimes stumble. They use the first chart that comes to mind—or simply carry forward the one they used during exploratory analysis. But a graph that works for exploring data isn’t necessarily the best for communicating it. Your audience and takeaway should drive the choice. By this point, you’ve already done that work: you know your audience, you’ve planned your story, and you’ve written takeaway titles that tell you exactly what each graph needs to show. Let those sentences guide your design.
A handful of graph types meet most everyday needs—here are the ones we use most at SWD (from storytelling with data: before & after, Wiley 2025):
Bar charts compare across categories. Dot plots and slopegraphs emphasize change between two points. Line graphs show change over time. When in doubt, familiar works: your audience shouldn’t have to learn how to read a graph before they can understand your message. Use something less familiar only when it reveals an insight that would otherwise be difficult to see. (For more on when to use these and other common visuals, check out the SWD chart guide.)
Choosing an effective visual is rarely a straight line from first attempt to finished graph. You try one form, realize it emphasizes the wrong thing, try another, get closer. This iterative process—experimenting with different views of the same data to find the one that best serves your message—is one of the most valuable things AI can accelerate. The outputs are often rough and may take some back-and-forth to get right. Still, rather than spending time building a chart only to realize it doesn’t work, you can prototype options quickly, evaluate them against your takeaway, and commit to a direction before investing in the final build.
If you’re following this series in order, you already know what you want to show—your takeaway titles from the storyboarding step outline this. If you’re coming to this post independently, or you’re still in exploratory mode, AI can help you figure out what’s worth visualizing first. Share your data and ask it to surface patterns or trends worth highlighting. Once something interesting emerges, pause to articulate the takeaway before moving into prototyping. Either way, that articulation step is key—it’s what will turn a prototype into a purposeful visual.
From there, the workflow is straightforward: tell AI what you want to communicate, share your data, and ask it to suggest options. You don’t need to clean or aggregate the data first—that’s part of what AI can handle. For each prototype, ask AI to explain the tradeoffs: what each option makes easy to see and what it obscures. You’re not looking for AI to make the decision; you’re using it to quickly surface options you can evaluate with your own judgment and knowledge of your audience and goals.
Once you’ve chosen a direction, build the chart in your tool of choice—either yourself or with the help of AI features within your tool. We’ll explore the latter in more detail in the next post in this series. Either way, it’s important that you remain responsible for the accuracy of what’s shown.
Before getting to the prompt and example, let’s review some potential pitfalls.
AI optimizes for the data, not the message—without clear direction, AI may try to visualize everything you give it or suggest a chart that represents the data accurately but doesn’t communicate your point. Lead with your takeaway, and be explicit about which data matters to the story.
AI may suggest unfamiliar chart types—it may recommend something technically interesting but unfamiliar to a general business audience. Push back if it suggests something your audience is unlikely to recognize immediately.
AI-generated prototypes aren’t finished work—use them to evaluate direction and spark ideas, then build the final chart yourself. When you build it from your own data, you also control the accuracy of what’s shown.
Be mindful of what you share—avoid including sensitive data or personally identifying information in your prompt. If your raw data contains details you’d rather not share, aggregate into a summary table or anonymize it before passing it to AI.
A note on tools: chart rendering capability currently varies significantly across toolsand account tiers. For the planning and thinking steps in this series, any tool can assist you. For visual prototyping, however, you’ll get better results with tools that can render actual chart images. Even then, it sometimes takes multiple prompts to get actual images in the output, rather than text-based approximations of the charts.
In my testing for this article, Copilot and the free version of ChatGPT struggled to render chart images for visual comparison, while the free versions of Gemini and Claude generally produced stronger results. These capabilities are evolving rapidly, so your experience may differ as the tools improve.
The paid tiers generally produced the strongest results overall, so if you have access to one, it’s worth using for this step. If not, the free version of Gemini is currently your best bet among the tools we tested. If you want to maximize your options, you could even copy and paste your prompt across multiple tools to generate a set of approaches to choose from or iterate upon.
If you’re continuing in the same AI conversation from one of the previous steps (start with context, craft a story), your context is already established and you can jump straight to the prompt below. If you’re starting a new conversation, take a moment to briefly orient AI: describe your audience and note the key message you’re trying to communicate visually. A few sentences should suffice.
I’m working on a data visualization for a presentation and want to explore chart options. I’ll share the context and my data. Act as a thought partner to help me identify and prototype effective visual options.
My audience is: [briefly describe]
I want to show: [state your takeaway in a single sentence]
How it will be used: [describe the context—for example, a single slide in a live presentation, a standalone graph in a report or email—and the goal, such as informing, persuading, or prompting a specific action]
Here is my data: [paste your data, aggregated table, or describe the dataset]
Please generate 2–3 charts that could work well for this message and audience. Render the graphs so I can evaluate them visually. For each, briefly explain what it makes easy to see, and what it might obscure.
Before making suggestions, ask me any questions that would help you give better input.
Note: if the output looks code-like or uses text and symbols to approximate the charts rather than rendering them visually, follow up with: Can you render the graphs as actual images so I can compare them visually?
Let’s look at an example.
If you’ve been following this series, you’ll recognize the scenario. I’m a People Analytics Manager at a mid-sized consulting firm. My team has completed a thorough analysis of the company’s hybrid work policy—examining performance ratings, in-office attendance patterns, collaboration network data, and attrition trends. We have a recommendation: move from the current three-days-in-office, two-days-remote policy for all employees to a differentiated approach based on role and team type.
In the first two posts, I worked through the context and story planning stages. I identified my audience and what’s at stake, formed a Big Idea, built a storyboard, and developed a narrative arc with takeaway titles for each planned slide. Now it’s time to start creating the actual content—and for several of the slides, that means choosing and building effective graphs.
I’ll work through two of those graphs here, each supporting a different point in the story. For the first—showing how early-tenure attrition has increased since we implemented the hybrid policy—I used Gemini. As mentioned, this tool had the best output across the free tiers that I tested (the others were Copilot, ChatGPT, and Claude). For the second—showing how the policy is affecting different employee groups in opposite ways—I used ChatGPT Plus. I followed the same prompting approach across both. In practice, you would likely continue with the same tool; I’m varying which I partner with here to give you a general sense of the output and how different tools handle this task.
I gave Gemini the general prompt shared earlier, with the following specifics:
My audience is: a leadership team with divided opinions and stakes in the outcome
I want to show: early-tenure attrition has spiked since we implemented the hybrid work policy
How it will be used: this will be a single graph that is part of a larger live presentation; the ultimate goal is to persuade the leadership team to move to a new differentiated policy
I shared a data table that summarized attrition rates pre- and post-hybrid policy by role type and tenure.
Gemini asked a few clarifying questions before proceeding—useful for orienting the tool, though by this point in the process I already had clear answers to all of them. It posed questions about the proposed differentiated policy (whether I’d be advocating flexibility based primarily on tenure, role type, or both), where the expected pushback would come from, and whether the leadership team prefers traditional graphs or if something less familiar would be acceptable.
After answering the questions, Gemini gave me three options: a slopegraph, grouped bar chart, and dumbbell graph. It explained that it chose gray to signify the Before Policy baseline and the Mid and Senior tenure segments where no dramatic change occurred, and red to highlight the Early tenure attrition post-policy, where the change was most dramatic.
These were accompanied by a table that explained for each why it’s powerful (what it makes easy to see) and what it may obscure (trade-offs). For example, for the slopegraph it stated, “It shows that all early-tenure groups were impacted, but highlights that the Collaborative group had the most extreme shift. The ‘story’ is instantly visible.” For trade-offs, it shared that the grouped bar chart “design can become very visually heavy and busy with 18 bars,” and that “if the audience is not familiar with dumbbell plots, they might need a moment to understand that the dots represent ‘Before/After’ points.”
Gemini recommended the slopegraph for this situation. It also shared the following to help think through how each might work in a live setting:
Gemini gave me some decent options here. I prefer the first two; the dumbbell, though interesting, feels unnecessarily complicated, and would take a lot of explaining before we could focus on what the data is showing. I agree with Gemini, that the slopegraph in particular makes the change between before and after the policy easiest to see (both where things have been stable in the higher tenured groups and where it clearly has not for the early tenure employees). The familiarity point about the grouped bar chart is worth keeping in mind.
If I needed a quick and dirty view for my own use, either of these would give me a useful starting point. Given the high stakes in my situation, I’ll want to recreate and customize the design for my audience. As I think ahead to the live presentation, I can imagine starting with overall attrition by tenure in a familiar grouped bar chart. From there, I could transition to a slopegraph, using the movement from one form to the other to make sure my audience knows how to read the slightly less familiar visual. Once that structure is established, I can move to the panel of three slopegraphs showing the breakdown by role type. I’ll recreate the visuals for this so I can have full control over the design details. I’ll use AI to help with it in the next post.
In the meantime, let’s look at some options for another important visual in my presentation.
Next, I worked with ChatGPT Plus. I gave it the general prompt with the following specifics:
My audience is: a leadership team with divided opinions and stakes in the outcome
I want to show: the hybrid policy is hurting the employees who need support most
How it will be used: this will be a single graph that is part of a larger live presentation; the ultimate goal is to persuade the leadership team to move to a new differentiated policy
I shared a data table that summarized performance ratings before and after the hybrid policy by role type and tenure.
ChatGPT’s questions were more analytically focused than Gemini’s—asking about sample sizes, confidence intervals, and cohort definitions. It also asked how much to editorialize in the graphic itself (not yet: I’ll do that in a later step). I answered the questions and asked it to proceed.
Here are the options ChatGPT suggested and the tradeoffs for each:
ChatGPT also offered “One additional idea I’d seriously prototype—a 2x3 small-multiple slopegraph.” It originally shared a prototype that was a little messy (see below). Note the narrative it outlines to accompany it—I found this a useful reminder of how the graph fits into the broader story, even if the visual itself needed work.
When I asked it to clean up the image, it confirmed I was okay with a mockup rather than a data-perfect chart, then shared the following:
Like it did for attrition, the slopegraph makes the areas of change stand out among the mostly flat lines. Looking back at ChatGPT’s initial suggestions, I had thought the diverging bar chart showing change from baseline would be workable—but this small-multiple slopegraph is clearer. We get the absolute numbers in addition to the change, whereas the diverging bars only showed the latter. Using the same graph type as the attrition chart has another advantage: by this point, my audience will already know how to read it, reducing the cognitive load for interpreting this data.
More broadly, this exercise reinforced a few things I’ve learned about working with AI. AI prototyping for visual choice is still imperfect. Getting usable chart images sometimes takes more back-and-forth than you’d like. This should improve with time. Even today, though, it’s often faster than sketching by hand or iterating directly in a graphing tool, and it surfaces concepts you might not have considered—particularly if you’re still building your repertoire of visual approaches. A few things worth trying: ask for more options if you want a broader set to evaluate, share a rough sketch with your tool if you have a specific idea in mind, or bounce between tools to see how different ones handle the same data.
The visuals I’ve chosen here are starting points, not finished graphs. In the next post—designing effective graphs and slides—I’ll return to some of these and work through how AI can help refine them: cleaning up clutter, focusing attention, and making them presentation-ready.
In the meantime, if you want to go deeper on using AI for data storytelling, watch the recording of our recent live event, where Simon and I share additional tips and examples.
2026-08-12 00:12:22
One of our top tips for explanatory communications is to use color sparingly and purposefully to help your audience understand your data and message. Color should be an explicit choice, not something your software applies by default, whether that's a graphing tool or an AI assistant generating your first draft. These tools can build a chart in seconds. They might even add highlighting on their own. But they don't know which data matters most to your audience. That call is still yours. Used thoughtfully, color is often one of the quickest ways to improve a graph.
If you struggle with how to do this in Excel (or with charts embedded in Microsoft Word or PowerPoint), this post will walk you through the steps.
Let’s consider an Excel graph, which shows year-over-year (YoY) change in sales for cat food brands from a pet food manufacturer. This example comes from exercise 4.2 of storytelling with data: Let’s Practice! You can download the data to follow along.
A good way to start is with shades of grey and then use color only where you want your audience to focus. A greyscale chart gives you a neutral foundation and serves as a reminder that more work needs to be done before it's presented, making it easier to be deliberate about where color belongs.
Begin by removing Excel’s default colors and replacing them with a shade of grey. Right-click on the chart and choose Format Data Series.
In the Format pane on the right, go to Fill & Line (the first tab with a paint bucket icon) and set the fill to a neutral grey.
If your graph had multiple series, you’d repeat as needed for the other bars until the full chart is in grey.
Tip: once you’ve created a clean greyscale version, save it as a reusable template so you don’t have to repeat those steps.
Let’s assume we want to call attention to the brands that declined year over year. To focus on these brands, we could choose to only color those bars. For accessibility purposes, we’ll avoid red and instead choose orange, a hue that reinforces the negative sentiment.
To color only the declining bars in Excel, click once on the bar you want to color. Then, click again to select just that specific bar (this avoids coloring all the bars at once). Right-click and choose Format Data Point. In the Format pane, go to Fill & Line (paint bucket icon). Under Fill, choose Solid fill and select the desired color.
Repeat for each declining brand, leaving the increasing brands in grey. Or, as a shortcut, you can simply click on the next bar and use Cmd+Y on Mac or Ctrl+Y on Windows to repeat formatting.
Coloring just the negative values tells your audience exactly where to look. Now imagine that within the declining brands, we want to specifically highlight the two brands that decreased the most: Fran’s Recipe and Wholesome Goodness. We could make those orange and everything else grey, but that would lose the signal for brands with lower sales this year versus last year. Instead, let’s vary the color's intensity to draw extra attention to the two biggest drops.
There are at least three ways to do this:
Color the two bars a darker orange and the others a lighter orange.
Change the transparency for the bars you want to deemphasize.
Overlay a transparent white shape on the other bars to create a stronger visual emphasis on the two bars at the top.
To manually vary the color in Excel—the first option—follow the same steps above but select a lighter or paler orange for the declining brands that are not Fran’s Recipe and Wholesome Goodness. Alternatively, you could keep the same color hue and adjust the transparency within the Fill menu (second option).
For the transparent shape, or the third option, insert a rectangle object (this shape works well for bar charts) over the other decreasing brands.
Then format the rectangle with a white fill, no outline, and the desired transparency.
By varying the color intensity, we can highlight all the declining brands while bringing greater focus to the largest negative changes.
Imagine we need to share this graph for an audience to consume independently. If we are not able to walk through the information with them, we can pair the designed chart with explanatory text to make the point clear. A text annotation next to the data, with key words bolded and colored the same orange, visually ties the words to the data, making the takeaway quick to see and understand.
Color should support your message, not distract from the data. The best charts use a restrained palette, purposeful highlighting, and clear words to guide the audience to the intended takeaway.
For more Excel how-to’s, check out our tutorial library. And if you want to see these same concepts applied to a line chart, look at our related post on emphasizing a data point.
2026-08-05 23:30:30
I’m proud to call the Pacific Northwest of the United States my home. Here we have rocky beaches, hidden waterways, uninhabited islands, and mountains as far as the eye can see. Which, admittedly, isn't very far through the fog most times of the year. I hike in Washington’s mountain ranges to take advantage of the summer weather. Hikes with a waterfall destination are among my favorites, and I live in the right spot for it: Washington is home to over 2,900 waterfalls, including the tallest waterfall in the continental US!
Being in nature is inspiring, and a recent waterfall hike got me thinking about one of my favorite types of charts: the waterfall chart. These are such powerful tools for storytelling, because they show where you started, what changed along the way, and where you ended up.
We challenged our community with waterfall charts many moons ago, but I’d love a fresh look. I have historically used waterfalls for sales breakdowns or cause-of-change analyses, and they also work for tracking HR headcounts, subscription data, or even score changes over the course of a sports match.
Waterfall charts work especially well for live walkthroughs, where you can spend time discussing important bars in the middle section of the chart, but I also love the white-space that they provide in written communications. I’m curious what all the different use cases are for waterfall charts, so that is your challenge this month.
Share a waterfall chart. You can use your own data, publicly available data, or simply practice with a classic SWD example. As always, be sure to anonymize any data that can’t be shared publicly.
For the uninitiated, I’ve gathered resources on how to build a waterfall chart and some design considerations in the related resources section below. Your waterfall communication can be serious or for fun, tell whatever story inspires you!
Share your waterfall chart story here by August 31st at 5PM PT. If there is any specific feedback or input that you would find helpful, include that detail in your commentary.
Here are some additional resources to help you build a waterfall chart.
SWD chart guide: what is a waterfall?
SWD video tutorial: Make waterfall charts in Excel (the EASY way!)
SWD video: waterfall presentation example
Datawrapper article: What to consider when creating waterfall charts
Microsoft PowerBI article: Waterfall charts in PowerBI
Analyst Academy video: When to use a waterfall chart (and how to make one)
2026-07-23 23:59:58
“Can you send me your slides before the meeting?”
This is a totally reasonable request if it happens often in your organization and you’ve planned for it. However, it might trigger mild panic if you were only planning to present the material live. A slide designed to support a spoken presentation is fundamentally different from one created for independent reading. Yet we don’t always make this distinction, which can lead to trouble.
In a recent team training, a client shared a slide intended to do three distinct jobs:
Serve as a pre-read for finance leadership before the meeting,
Support their live presentation, and
Work as a leave-behind reference after the meeting.
While it can feel efficient to design your communication to meet multiple needs like this, it can easily backfire. When you try to make a single slide work for different situations, you’re forced into compromises that mean it doesn’t fully satisfy any of them. The result? Communications that are too dense to present live, yet too sparse to stand on their own. When we know the scenario we are designing for, we can tailor the materials specifically to that setting.
Let’s look at the client’s original slide (anonymized for confidentiality), which tried to satisfy all three needs at once. Then we’ll explore ways to improve it.
As a standalone slide, the commentary provides helpful context, but the core message isn't clear. The reader has to work to identify what matters and why. A strong pre-read should help leaders arrive informed and ready to discuss, rather than leaving them scanning through tables before the meeting.
In a live setting, this slide creates immediate friction. The busy tables and bulleted text force people to split their attention between reading the screen and listening to the presenter. Even if the audience read it in advance, they'll spend time hunting for details as you speak rather than engaging with the message. A live delivery is more effective when attention is deliberately guided by revealing one idea at a time instead of showing everything at once.
Being specific about the situation helps to improve design. Ask what the slide needs to do, who will use it, and in what setting. Once those are determined, it becomes much easier to shape the story and choose the right format.
The overall message the client wanted to communicate in the original slide was that, although the year-to-date performance is ahead of plan, a lot of work remained to achieve the forecast. The main action was to reiterate the priorities to ensure the team remained focused on the three workstreams and to quickly escalate any blockers to achieving the projected savings.
Articulating the goal helps determine what stays, what goes, and how to structure the narrative for each situation.
Let’s consider the live delivery first. When presenting in person, you don't need to cram every detail onto a single view because you control the pacing. Instead of showing everything at once, break the content into a sequence where each slide supports a specific point.
Here is how we could progress through the same information for a live audience:
Because the story unfolds in a controlled manner, the presenter can guide the room through the information instead of hoping people find the right detail at the right moment. Takeaway titles frame the key message for each slide so the audience never has to guess. Supporting visuals reinforce the narrative rather than competing with it.
If leadership has already reviewed the pre-read, a live meeting shouldn't rehash every data point—it should focus on implications and decisions. But what about the materials that will be sent around before or after the meeting?
For a document meant to be read independently, incorporate the necessary depth without overwhelming the reader. Rather than leaving the data in dense tables, organize the content into a clear summary designed for solo reading.
Single-slide communications often contain a lot of information, which can make it difficult to digest. By moving to a structured two-panel layout, the document establishes clear visual hierarchy. Bold takeaway headers immediately signal where to look first, while complete sentences directly beneath that provide the necessary detail and context for independent reading. Color applied intentionally and sparingly ties the text to the data it describes.
The original table views can then move to an appendix as a reference for anyone who needs the precise values for each workstream.
To recap, circulated slides and presentation decks are related, but they are not the same thing. When one communication is asked to do multiple jobs, the result is almost always compromise.
The next time you build a communication, pause and consider the medium:
For live presentations: Keep slides sparse so the audience focuses on you and your spoken narrative.
For written consumption: Build structured, detailed documents that provide complete context for independent reading.
Designing intentionally for the scenario at hand ensures your audience gets the right level of detail—and your message lands.
For more examples of visual transformations, check out the before-and-afters in our makeover gallery.
2026-07-09 20:34:54
Like many professionals right now, I’m part of the real-world experiment with AI in the workplace. Almost every day, I’m figuring out where it helps and where it gets in the way. As I experiment, one question keeps coming up: when do I actually want to use AI?
There are tedious parts of the data storytelling process that I genuinely enjoy, like tinkering with alignment and color, and sketching ideas before committing. While these steps might feel unnecessary, they almost always result in new ideas and better output. By contrast, building each chart from scratch or manipulating Excel data to produce a non-standard chart rarely improves my thinking. This is the part I’d happily hand off, as long as it doesn’t cost me control over the final design. Every minute I don't spend manually creating charts is another minute I can spend thinking about the underlying story and design.
I'm not looking for AI to redesign the slide for me, but I am open to using it to produce a starting point that I can refine and polish. I recently put this collaborative approach to the test, and the results are shifting how I work.
A client project gave me the perfect testing ground. Below is the original slide. (The data and scenario were altered to protect confidentiality.)
Vyrenta's early revenue lead narrows by 2028 as Alunis expands outside UCAN. Figures in USD bn. Sample data, not actual financials.
My initial thought: this stacked bar chart is trying to do too much. The goal is to show how the company, Alunis, will close the revenue gap with a top competitor, Vyrenta, by focusing on regional expansion over time. This chart needs to communicate change over time, compare two companies, and show each company's regional composition. That’s a lot of heavy lifting for one visual!
Side note: if you’ve checked out our latest book, before & after: practical makeovers for powerful data stories, you may recall the chapter on five common data visualization mistakes to avoid. The first most common mistake: adding too much to one graph.
Step one for this makeover is to avoid the mistake by splitting the graph into multiple charts. Here’s where things start to get interesting.
Since I like to sketch to iterate on my ideas, I made a rough draft of how I thought this redesign of multiple charts could look.
Hand-drawn sketch of the chart layout showing Vyrenta leads on total revenue and in UCAN, while Alunis grows faster and pulls ahead across APAC, EMEA, and LATAM.
I shifted from bars to lines, which have a lighter visual footprint—an advantage for a slide with so much data. I divided the graphs by location: one graph for total revenue and one for each region. Instead of sizing them consistently, I made the total view larger, and arranged the four regional charts into a small grid occupying the same amount of space as the total chart. This is my subtle way of visualizing the data’s part-to-whole nature.
Pre-AI, I would have created each one of these graphs in PowerPoint. This creation step would take a lot of time, patience, and mental energy.
This time around, I shared my sketch with Claude. For a few months, I’ve been testing Claude for PowerPoint, an add-in that was released earlier this year. While I’m not attached to any specific AI tool, this PowerPoint add-in has been the most useful for my workflow because it creates editable charts directly in PowerPoint, my preferred slide software.
That editability is what makes this collaborative experiment possible. It’s worth mentioning that CoPilot and ChatGPT also create editable charts. (I’ve had less success with the output—specifically CoPilot.) I’ve also learned that Gemini offers a similar capability for Google Slides, so this workflow can be adapted to other common tools.
Below is my prompt and the information I shared with Claude. Please ignore any typos, as I cannot be bothered to proof my chat messages.
The Claude assistant is asked to rebuild the chart from the the uploaded data and hand-drawn layout sketch.
Here is the output from the above prompt. It’s decent. In fact, it’s better than that. This is a marked improvement from the original stacked bars.
Claude rebuilt the sketch as fully editable native PowerPoint charts using the same small-multiples layout, real axes and labels.
Now, for most people, simply rewriting the slide title and adding additional descriptive language above the charts would be more than enough to call this makeover done.
For me, however, it’s not enough. There are a few easy tweaks that could take this to the next level. I’ll describe the top three.
Shade the areas: Since the comparison is between the two companies, I’ll make this more obvious by shading the space between the lines and adjusting the color to indicate the lead. At a glance you’ll see the purple overtaking the gray for three of the four regional charts—a design choice that reinforces the main takeaway.
Adjust axis ranges: Best practice states that you shouldn’t have multiple small charts near one another with different scales. When you deliberately break a best practice, it's worth understanding why the rule exists in the first place: to prevent someone from missing the scale difference and misreading the data. In an ideal world, I’d make all of the scales consistent, but if I do that here, it will be impossible to see the change at the regional level. To design around this constraint, I’ll make the ranges somewhat comparable by adjusting the maximum Claude chose, only label the minimum and maximum values, and add data labels for clarity and emphasis. I realize it’s not perfect.
Format the years: Space is limited on this slide so I want to preserve every bit of it that I can. For the year labels on the right, I’ll replace the four-digit years with two-digit abbreviations. I’ll also choose to only label the bottom charts.
My final design looks like the following.
The same charts, now ready to share. Changes made: shaded areas between the two lines, direct end-of-line labels with values, per-panel axis ranges scaled to each region, solid-versus-dashed lines to split actuals from projections, and compact year formatting on the smaller panels. Sample data, not actual financials.
I’m quite pleased with the redesign, not just for the final outcome, but also for the impressive partnership between humans and AI throughout this makeover process. This result isn’t always the case in my many AI experiments, and it comes down to one thing: editable output.
A big hurdle for many practitioners was that AI-generated charts were images, and for whatever reason some of the image processing lacked accuracy. That makes it critical for a data visualization designer to make adjustments to the final output. Now that AI tools can generate solid editable charts and slides, they're finally useful additions to an existing data storytelling workflow.
The implications extend beyond experienced practitioners. If you're new to data storytelling or aren’t sure how to build a particular chart in, AI can help you overcome that initial learning curve.
It doesn't eliminate the need to learn the foundational principles that guide the output. For this AI-generated makeover, I still had to know that a stacked bar chart wasn’t ideal, and why. I still had to determine a simple layout for five charts on one slide. And I still had to recognize that breaking the "consistent axis scales" convention wasn't a mistake, but a deliberate design decision worth reinforcing. I was fully in the driver’s seat.
That’s the value we continue to bring at the moment. We provide judgment and direction, while AI accelerates the execution. And if you ask me, that’s the most enjoyable part of data storytelling, anyway!
Want to learn more about how to partner with AI? Watch the recording of our live event. Cole Nussbaumer Knaflic, CEO & founder of SWD, and fellow data storyteller, Simon Rowe will share practical ways to maximize your data storytelling workflow with AI.
2026-06-24 22:00:08
This post is part of the SWD + AI series—practical guidance for using AI as a thought partner across the various stages of your data storytelling work. Explore all of our AI resources.
In the previous post, we used AI as a thought partner to develop your Big Idea. You know your audience, you understand what’s at stake, and you have a clear point of view. Now comes the time to turn all of that into a story.
A common instinct at this point is to open your laptop and start building slides. Resist it! Before touching any tool, the most important thing you can do is go low-tech. Get out some sticky notes. Brainstorm content that might belong in your communication—one idea per sticky note—without filtering or judging. Then edit ruthlessly: what stays, what goes, what order serves your message? The low-fidelity format matters. It keeps you from getting attached to ideas before they’ve earned their place, and rearranging is effortless when nothing is committed to a slide.
This is storyboarding—and it’s one of the most powerful planning tools we teach at SWD. The process I’ll walk through took about 30 minutes, from initial sticky notes to a tested narrative and draft takeaway titles. That’s time invested upfront so I’m not spending hours later reworking slides to fix structural problems. But brainstorming and arranging content is only half of it. The other half is finding the right structure to hold it together.
The default flow for most business presentations is a linear path: you start at the beginning of the project, walk through what you did, and end with findings and recommendations. It’s logical. It’s familiar. And it tends to be forgettable, because it mirrors the analyst’s journey rather than the audience’s needs.
An alternative—the structure we find incredibly effective for business storytelling—is the narrative arc. Rather than following the chronology of the work, the narrative arc organizes content around tension and resolution. It has a deliberate shape: plot that establishes context, rising action that builds towards the central conflict, a climax where the tension peaks, falling action that moves toward resolution, and a clear resolution—the call to action.
Making the shift from the linear path to the narrative arc isn’t always intuitive. Many people find it straightforward to brainstorm content and arrange it chronologically. But restructuring around tension and resolution requires a different kind of thinking. This is one place where AI can be particularly helpful: not to generate your story, but to help you find the arc that’s latent in your material.
The key ingredient that makes it work is tension. This isn’t the tension that matters to you as the analyst or the communicator of information, rather it’s the tension that matters to your audience. Recall what we assessed in the first installment of this series: What is at stake for them? What is the gap between where things are and where they need to be? When you identify that tension and build your story around it, you stop reporting findings and start telling a story that moves people to act.
The storyboarding process is deliberately human. Brainstorming, editing, and arranging your content is work that should happen away from the screen and away from AI. As we’ve discussed, the analog nature of it is a feature, not a limitation.
That said, AI can be a useful thought partner at two different moments in this process. The first is after brainstorming, when you have an array of sticky notes but haven’t yet found a structure. At this point, AI can help you identify a narrative arc from your raw material: finding the tension, suggesting an order, helping you see the story that’s embedded in what you’ve generated. The second is after you’ve arranged your content (with or without AI’s assistance)—once you have a draft structure. Here, AI can pressure-test whether the arc holds: evaluating the tension, identifying gaps, and flagging places where the story loses momentum.
Either way, the goal is the same: to surface blind spots and help you assess and refine before you invest time building anything.
Once your low-tech plan is solid, AI can also help you turn your sticky note topics into draft takeaway titles for your eventual slides. This is a useful bridge between planning and building. We’ll see how it works soon through an example.
Before getting to the prompt and example, let’s review some potential pitfalls of working with AI at this stage of the data storytelling process.
AI may flatten the tension—effective stories need a moment of discomfort before resolution. AI often smooths this out in favor of a more neutral, balanced narrative. Watch for responses that sand down the edges of your story.
AI will want to add, not subtract—it tends to suggest more content, more context, more caveats. A tighter story is almost always a stronger story. Use AI to find out what to cut, not what to add.
AI doesn’t know your audience—it can evaluate logical flow, but only you know what your specific audience needs to hear, in what order, to be moved to act.
Be mindful of what you share—avoid including sensitive data, personally identifying information, or confidential business details in your prompt.
If you’re continuing in the same AI conversation from the previous step, your context is already established and you can jump straight to the prompt below. If you’re starting a new conversation, take a moment to briefly reorient AI: share your Big Idea, describe your audience, and note what’s at stake and the key tension you aim to build the story around. A few sentences should suffice. Then use the prompt below, modifying as needed to meet your needs.
I’m working through the storytelling process using storyboarding and the narrative arc framework from storytelling with you by Cole Nussbaumer Knaflic (Chapters 3 and 4). I’ll share my work and ask you to act as a thought partner to help me find or pressure-test the narrative arc. Your role is to identify gaps, inconsistencies, or places where the story loses momentum. Don’t rewrite it for me and don’t suggest adding more content. Focus on whether what I have can be shaped into—or already follows—a clear narrative arc that builds tension and moves toward resolution.
Start with the option that fits where you are in the process:
Option A: after brainstorming
I’ve generated ideas for potential content but haven’t structured it yet. Help me identify a narrative arc from this raw material. [share a photo of your sticky notes or a list of your ideas]
Option B: after arranging
I’ve arranged my content into a sequence. Help me evaluate whether it follows a narrative arc or whether it could be restructured more effectively. [share a photo of your storyboard or a list of planned content]
Review and tell me:
Does the content map to a clear narrative arc—is there identifiable tension, clear climax, and fitting resolution?
What isn’t earning its place—is there anything that feels redundant, out of order, or likely to lose the audience?
Are there any logical gaps—places where I’m assuming the audience will connect dots they may not connect?
Before providing feedback, ask me questions that would help you give better input.
To see how this comes together in practice, let’s walk through an example.
If you read the first installment in this series, you’ll recognize the scenario. I’m a People Analytics Manager at a mid-sized consulting firm. My team has completed a thorough analysis of the company’s hybrid work policy—examining performance ratings, in-office attendance patterns, collaboration network data, and attrition trends. We have a recommendation: move from the current three-days-in-office, two-days-remote policy for all employees to a differentiated approach based on role and team type.
Our Big Idea is: It’s time to shift from our current three-days-in-office policy to a differentiated approach based on role and team type—one that meaningfully reduces costs and enables people to perform better and stay longer.
Now the work shifts. Context is set. The Big Idea is clear. It’s time to figure out how to tell the story.
I started—as I always do—with sticky notes. I wrote one idea per note, without filtering. I thought about the project from the perspective of each person who will be in the eventual meeting room: what will Diana need to feel confident championing this recommendation? What would make Robert a supporter? What might Priya push back on? I also thought about the data: what we learned and which findings were essential, versus interesting but not necessary.
After about 10 minutes, I had twenty-five ideas in front of me. I’ll list them here so you can get a sense of the breadth:
Why we undertook the analysis
Current hybrid policy
History/timeline of policy
Methods: data we collected
Benchmark data: comparison to peers
Employee survey data: work preferences
Performance ratings by role type & tenure before/after
Collaboration network data before/after
Office utilization: actual vs. assumption
Office utilization: breakdown by office
Attrition rates before/after
Attrition rate: breakdown by office
Travel/OOO breakdown by role
Manager feedback on hybrid policy
Cost of attrition by role type & tenure
Real estate cost analysis & projections
Finding: impact varies by role type & tenure
Finding: early-tenure employees struggle most
Finding: independent/analytical employees thriving
Finding: office utilization lower than expected
Cost of status quo
Options considered
Recommendation: differentiated approach by role and team type
What differentiated policy would look like in practice
Next steps and implementation timeline
Then I edited. I put aside anything that served my process rather than my audience’s needs—the methodology, policy history and timeline, the benchmarking data, options considered. I also set aside anything unnecessarily granular for my audience or that didn’t lead to actionable output: attrition by location, manager feedback, geographic utilization patterns.
As I edited and arranged, I also found myself rewriting my sticky notes. What had started as descriptive—“attrition rates before/after”—became more pointed: “early-tenure attrition has spiked, and it’s costing us.” The process of deciding what to keep and how to sequence it was already pushing me to think like a storyteller rather than an analyst.
That left me with twelve items. I arranged them in a sequence—not yet a narrative arc, just the order that felt logical to me at the time, which turned out to be fairly chronological: policy overview, data findings, recommendation, call to action.
I had a solid foundation but wasn’t yet confident in the structure. The data findings in particular felt like a list of things I’d learned rather than a story with momentum—and I wasn’t sure the tension was apparent for my audience. This felt like a good point to bring in AI to help. I turned to Gemini.
Quick note on tools: I used Claude in the first post in this series and Gemini here. In practice, it will usually make sense to stick with a single tool throughout a project so context can accumulate over time. I’m intentionally varying the tools I use throughout this series so I can see how well this workflow transfer across models.
Given that I haven’t used Gemini yet for this communication, I shared initial context along with my storyboard and prompt:
The following is Gemini’s initial response.
Gemini’s response was more useful than I initially realized. At first glance, it seemed to confirm the order I already had. But three observations stood out on closer reading.
The first was the question about sticky 7, “one policy can’t serve everyone equally.” Gemini flagged that stating this finding too early might give away the punchline before the data has had a chance to build the case. That’s a fair point, and this caused me to consider another reason to hold it back. For an audience member like Priya, who championed the current policy and is likely to be defensive about any change, naming the conclusion too early invites resistance before the evidence has had a chance to land. A stronger approach in this instance could be to let the individual findings accumulate—the attrition spike, the performance decline, the role-type differences—so that by the time we state “one policy can’t serve everyone equally,” the audience has likely already arrived there themselves. Given this, I decided to move this note later in the sequence.
The second was the suggestion to bring the financial figures closer together. Currently, I had the early-tenure attrition costs ($2.4M) in the rising action and the real estate savings ($3.2M) much later. Gemini pointed out that combining them creates a single, harder-to-ignore financial argument—roughly $5.6M in total impact. For Robert (the CFO) in particular, that framing is more compelling than two separate numbers spread across the presentation.
The third observation—about the climax—is where I disagreed. Gemini placed the real estate opportunity at the peak of tension. But for this audience, the emotional peak isn’t the financial case. It’s the human one: early-tenure employees are leaving at rates far above our historical baseline, and the current policy is actively making things worse. That’s the moment I want the room to feel before I offer the resolution. The financial argument is powerful supporting evidence, not the climax.
This is a good example of why AI is a thought partner, not a decision-maker. Its structural suggestions were useful. But I’m the one who knows what will resonate with this specific group. AI can help you think through audience needs, but it can’t replace real audience knowledge. When that knowledge is incomplete, talking with stakeholders or people closer to the audience is often more valuable than asking AI to fill in the gaps.
I arranged my stickies along the general arc Gemini recommended, incorporating the ideas I raised above, and continuing to rearrange some things to come up with a narrative flow that made sense to me (for example, leading from traveling employees to the lower office utilization to the real estate savings, since these ideas build on each other). I explained this change to Gemini, which responded with the following.
While I’m certainly feeling good about things now, I’ve learned to be cautious about AI enthusiasm. “You have a rock-solid narrative arc” is encouraging—but encouragement isn’t the same as rigorous pressure-testing. AI tends toward affirmation, and a confident-sounding endorsement can give a false sense of security. This is exactly the pitfall I flagged earlier: AI may flatten the tension and smooth over the edges of your story precisely when you most need rigorous push back.
Before moving on, I wanted to put the arc under more scrutiny. So I shared the revised sequence (below) with Gemini and asked it to pressure-test the narrative more specifically—this time focusing on whether the tension was genuinely meaningful to my specific audience, whether anything was losing momentum, and whether there were logical gaps my audience might not bridge on their own (Option B of the potential prompt shared earlier in this post).
Here’s Gemini’s response:
Before diving into feedback, Gemini asked three clarifying questions—a direct result of the “ask me questions first” instruction in the prompt. I hadn’t explicitly shared my audience context in this exchange, and rather than plowing ahead with generic feedback, Gemini flagged that gap. That’s the prompt doing its job—and a good reminder that the quality of AI’s feedback is directly tied to the specificity of what you give it.
The questions caused me to reflect. The first pushed me to be explicit about something I’d been holding implicitly: Diana is my primary audience, and her bias is cultural—doing right by employees while presenting a position she can defend with data. But Robert’s financial bias matters too, and Marcus’s operational one. A story that lands for Diana needs to at least speak to the others.
The second question had an obvious answer for me: Priya. She championed the current policy and is likely to see any change as a threat. The most important objection to anticipate is hers.
The third identified a potential gap I hadn’t explicitly thought through: the verbal bridge between the real estate finding and the early-tenure performance findings. These two sit on opposite sides of the arc, and how you move an audience from one to the other matters. I hadn’t worked that out yet, and it’s exactly the kind of thing that’s much easier to address now than after I’ve built the slides
I’ll also note the opener. This kind of enthusiasm is AI’s default mode: charming, but meaningless. Watch for it, and don’t let it substitute for the substantive feedback that follows.
I answered Gemini’s questions and asked it to proceed with the pressure-test. Here was the response:
A few things stood out. The observation about stickies 8 and 9 was practical and correct: office utilization and real estate savings are one idea, not two, and presenting them separately risks stalling the momentum right before the climax. I’ll combine them.
The observation about sticky 7—“one policy can’t serve everyone equally”—was the most useful structural insight. Gemini suggested treating it as falling action rather than the climax itself. This doesn’t change where it sits in the sequence, but does impact how I frame it. Stated as the peak of tension it sounds like an accusation directed at the current policy—and at Priya. Let the data accumulate first: the attrition spike, the performance decline, the role-type differences. By the time we state the conclusion, the audience will already have arrived there themselves.
The two gap identifications were also valuable. The first—bridging the early-tenure performance decline to the attrition spike—isn’t only about narrative flow. It’s anticipating Marcus’s objection before he can make it. The second—a proactive nod to Priya’s perspective on hybrid flexibility—is the kind of thing I need to do but might have left implicit. Having it named as a structural gap rather than just a presentation nicety is useful.
The quality of this feedback was directly tied to the audience context I provided. Gemini’s earlier response was useful but generic; once it knew that Priya had championed the current policy, it could surface the “Priya defense” gap. In hindsight, I could have accelerated this process by providing that audience context from the beginning. Had I uploaded my completed Big Idea worksheet, Gemini would have had access to those details from the start.
Rather than rearranging my stickies again, I noted these refinements to carry into the building phase. With the arc solid and the structural gaps identified, I had one more thing to do before closing out the planning stage: turn my sticky note topics into draft takeaway titles.
I gave the following simple prompt: I’m happy with the overall structure. Can you now help me turn each point in my storyboard into a draft takeaway title for the eventual slide? Each title should be a single, pithy statement that tells the audience what to notice or understand from that slide—not a label for the content, but a clear point of view.
Gemini initially gave me much more than I asked for. Rather than parse it myself, I followed up with: Thank you—can you give me just the draft titles as a simple numbered list, without the sticky note references, arc labels, or explanatory notes in parentheses?
Here is the response:
This is a good starting point. Read top to bottom, they tell a cohesive story. Each title pushes the narrative forward and the financial figures feel integrated naturally. “The true cost of this mismatch is human” is strong as a climax title and will appeal to my audience.
That said, several are too long for slide titles. A takeaway title needs to make the point succinctly—the supporting detail lives in the content of the slide itself, not the title. With that in mind, I refined to the following:
Our hybrid policy treats all roles the same—but they aren’t
We looked beyond office attendance: perf, collaboration, & more
Independent and analytical teams thrive remotely
Office mandates work against client-facing teams
Current policy doesn’t match how people actually work (and we’re paying for it!)
The real cost is human: new hires are set up to struggle
Early-tenure attrition has spiked—and it’s costing us
One policy can’t serve everyone equally
A differentiated approach recaptures $5M+
Not removing flexibility—making it work better
Approve the change: stop the drain
These eleven titles give me a clear plan for building the deck. They aren’t final—as I build out the actual slides with my team, some will sharpen, some will shift, and a few may get trimmed or expanded as the content takes shape. But having them now means I won’t ever have to start from a blank slide. The story is mapped. The tension is clear. The call to action is explicit. Everything that follows is building on a foundation rather than figuring it out as I go.
One reason I remain such a strong advocate for storyboarding is that it saves time. The work described here may look involved when written out step-by-step, but the actual process was fast—roughly 30 minutes from initial sticky notes to a tested narrative and draft takeaway titles. That’s time invested upfront so you’re not solving structural problems after you’ve already build slides.
For this high-stakes communication, working through the storyboarding and narrative arc process with AI as a thought partner got me to a place I wouldn’t have reached as quickly on my own. What I have is stronger for it: more robust, more audience-aware, and more intentional than if I’d worked through it alone. That’s pretty incredible.
In the next post in this series, we’ll move to the third core SWD skill: choosing an effective visual. In the meantime, watch our live event, where Simon and I explore how to use AI for better data storytelling—including diving deeper into ideas from this series.