As professionals increasingly turn to artificial intelligence to streamline their workflows, a new frontier is emerging in data visualization and presentation design. The latest installment in the ongoing SWD + AI series offers practical guidance for using AI as a thought partner across the various stages of data storytelling work. Following previous explorations into starting with context, crafting a narrative story, and choosing appropriate visuals, the focus now shifts to the critical phase of actually designing those visuals.

Once visuals have been chosen, the real work begins. Designers must ensure their slides communicate clearly, direct the audience’s attention to what matters, and hold together as part of a cohesive presentation. This is precisely where many people spend the bulk of their time, and where the gap between a design that is merely "good enough" and one that is genuinely effective becomes most visible. Whether a graph originates from a traditional charting tool or an AI generator, the first version is rarely ready to present. It almost always requires decluttering, focused attention, and careful wording to make it meaningful.

A graph without a takeaway title forces an audience to guess at the underlying point, while a slide crowded with raw data makes them work far too hard to understand it. Getting this step right is ultimately what separates a presentation that simply informs from one that actively moves people to take action. Fortunately, this is also a stage where artificial intelligence—provided it receives the right context and direction—can shoulder much of the heavy lifting. Rather than building each slide from scratch, professionals can describe what they need, share their data, and let AI produce an initial version to react to. While this first draft will rarely be perfect, it serves as a valuable starting point, allowing creators to reach a working draft in minutes rather than hours and fundamentally changing what is possible in tight reporting cycles.

Unlike earlier phases of data storytelling where AI might be used to evaluate standalone prototypes or static images, current workflows allow for AI integration directly into presentation software such as PowerPoint. This distinction is crucial because the resulting output consists of fully editable slides rather than unchangeable images. Creators are no longer just reviewing concepts; they are building the real deliverable, with AI helping to refine elements directly in place.

Effective graph and slide design relies on a few essential principles: removing what does not belong, creating visual order through alignment, actively directing attention to what matters, and using precise wording to clarify the message.

The first major principle is decluttering. Every single element on a graph or slide contributes to cognitive load, which is the mental effort an audience must spend processing visual information. Anything that fails to add informative value constitutes clutter, and clutter inherently makes visuals feel far more complicated than they actually are. The objective is not minimalism for the sake of aesthetics, but rather the reduction of extraneous details so that essential insights can stand out. In practice, this frequently means stripping away chart borders, heavy gridlines, unnecessary tick marks, redundant labels, and default formatting that software tools often add simply because they are built-in templates rather than audience-serving design choices. When unnecessary elements are removed, the remaining data points grow stronger.

Alignment represents the second core pillar. When elements across a slide are misaligned—such as text starting at varying points, graphs floating at inconsistent positions, or titles failing to anchor the content positioned directly below them—the overall slide feels fundamentally unintentional, even if viewers struggle to articulate why. Consistent alignment establishes a reliable visual structure that an audience can navigate intuitively without conscious thought. Thoughtful design fades gracefully into the background, whereas poor design places a heavy cognitive burden on the viewer.

Focusing attention is the third vital step. Even a clean, well-aligned visual still leaves the audience to decide for themselves what matters most, and there is no guarantee they will focus on the intended takeaway. Creators must actively guide where the eye travels. The single most powerful tool for this objective is color used sparingly. When every element shares the same color, or conversely, when every element is a different color, nothing receives emphasis. However, when a single element is highlighted against a neutral field, the audience looks there immediately. Size, position, and contrast operate on similar visual principles; making important data points larger, placing them where the eye lands first, and leveraging visual weight helps establish a clear hierarchy. The goal is to make the right takeaway entirely obvious without needing to explicitly command the audience where to look.

Using words wisely is the fourth principle. Even the finest graph requires supporting text to make its point crystal clear. Every graph needs a descriptive title that informs the audience precisely what they are viewing. Every slide requires a strong takeaway title—a complete sentence that answers the "so what?" question before the audience is forced to ask it. Axes require clear labeling, and key data points benefit from concise annotations where they add genuine meaning. Nevertheless, more words do not equate to better communication; every word that fails to earn its place merely adds noise. The discipline required here mirrors every other aspect of data storytelling: include what serves the audience and ruthlessly cut what does not.

A final practical consideration involves adhering to organizational standards. If a company maintains a standard template or formal brand guidelines, professionals should always start there. Colors, fonts, and layouts aligned with established brand identities are far more than aesthetic choices; they signal professional credibility and consistency to the audience. Before applying broader design principles, creators must ensure they are working within their corporate visual identity, consulting internal communications or design teams if any uncertainty remains.

These core principles function harmoniously when applied with the ultimate audience in mind. Creators must constantly evaluate what their audience needs to see and what potential obstacles might get in their way. When slide design succeeds, it becomes entirely invisible, allowing the audience to focus exclusively on the message rather than the mechanics of the presentation.

Working with AI to Design Effective Graphs and Slides

Having a shared foundation in proven data storytelling principles becomes especially useful when collaborating with artificial intelligence. While earlier stages of work relied on providing AI with just a few sentences of context, experts recommend going further by sharing the specialized SWD + AI primer with the tool before beginning design work. This foundational resource grounds the AI in structured design principles, yielding noticeably higher-quality results than prompting without background context.

The modern workflow relies heavily on AI integrated directly into presentation applications rather than functioning as a standalone chat interface. This integration ensures that outputs manifest as editable slides rather than static graphics. Tools like Copilot in PowerPoint, alongside dedicated AI assistants integrated into presentation environments, fully support this streamlined workflow.

Professionals can approach this phase with a draft slide they have already constructed, asking the AI to help refine it, or they can describe their goals and let the AI generate an initial version to react to. Either approach typically bridges ninety percent of the gap quickly, producing clean, structured, and targeted slides that would have demanded significantly more time if built entirely from scratch. The remaining ten percent relies entirely on human judgment—the final adjustments to color, emphasis, wording, and alignment that transform a standard slide into an exceptional one. Some of these refinements can be executed by directing the AI, while others are best handled directly by the creator.

Before initiating prompts, users are advised to share the primer and instruct the AI to adopt those specific design principles. The typical prompting framework involves sharing the draft slide or describing the desired visual alongside the underlying data, briefly defining the target audience, stating the core takeaway in a single sentence, and clarifying how the slide will be utilized within a live presentation or standalone report.

In Practice: Bringing Design Principles to Life

For professionals following the progression of this ongoing series, the recurring scenario centers on a People Analytics Manager at a mid-sized consulting firm who is preparing a leadership presentation to recommend a structural shift in hybrid work policy. Across previous installments, this manager successfully identified the target audience and formulated a central Big Idea, mapped out a compelling narrative arc, and selected the appropriate visuals to construct. The current phase centers entirely on translating those choices into polished slide designs.

In this practical application, the manager utilized Copilot integrated directly into PowerPoint, ensuring the resulting output consisted of editable native slides rather than images requiring manual rebuilding. Before diving into generation, the user shared the foundational SWD + AI primer with the assistant, establishing a firm grounding in data communication principles. Once the AI confirmed its readiness, the manager provided structured data summaries alongside specific prompts defining the audience, the primary takeaway, and the functional context of the presentation, successfully bridging the gap between raw data and professional visual communication.