We’ve all sat through slide shows that look stunning and say very little. That disconnect is the real problem. My view is simple: AI can speed the work, but only strong human editing makes a deck persuasive. Watching Jeff Su break down his process showed me how to turn pretty slides into sharp arguments.
Jeff’s core message matches what I see every week with brand leaders and teams. Drafts are easy. Iterations win. As he puts it, the real job is what happens after the first version. I agree—and I’ll go further: AI should be your assistant, not your author.
The Case for Human-First Slides
Jeff draws a clear line between text models that argue well and slide tools that look great. The trick is combining both. He nails it with this north star:
“What you actually want is up here. Slides that are analytically sharp and visually polished.”
I’ve watched too many teams outsource thinking to tools. That’s how you end up with decks that impress at a glance and fall apart on questions. Jeff’s walkthrough hits the right nerve: iteration, not generation, creates trust.
What Jeff Gets Right
Jeff treats AI as a power tool for speed. But he never dodges the hard part: judgment. Three moments stood out.
“The initial draft is never the hard part.”
Yes. The hard part is deciding what stays, what goes, and what moves. When legal says risk belongs up front, it moves. When a claim lacks a source, it gets fixed. When a chart inflates numbers, it gets edited. That’s craft.
“Moving slides around is something we can do manually in much less time.”
Exactly. Don’t waste prompts on tasks your hands can do faster. Save AI for heavy lifts like sourcing data, drafting charts, or compressing dense copy.
“The agent [is] designed for speed, not reasoning.”
That’s the line every leader should memorize. AI will happily “combine” two calls-to-action by cramming both on a slide. It won’t apply taste unless you tell it how. Strong prompts help, but taste and narrative are on you.
Proof That Precision Matters
Jeff’s examples show how precision changes outcomes.
- He moves risk mitigation right after the executive summary to match leadership priorities.
- He asks the tool to find current model market share, then turns it into a table and a chart.
- He updates an inflated usage stat to match official figures and adds a footnote.
Those steps look small. They signal credibility. They also show the right use of AI: fast research, clean visuals, and less grunt work.
- He directs layout to make the key column dominant, then manually adjusts type size and spacing.
- He combines two “act now” slides, limits to three points, and removes the image for clarity.
- He translates the deck, then runs a bilingual polish step with another tool.
This is the playbook: speed from AI, quality from you.
My Playbook For Decks That Win Fans
As someone obsessed with creating superfans, I care most about clarity and action. Here’s how I apply Jeff’s approach with my clients.
- Write titles first. If your titles tell the story alone, you’re on track.
- Trim body text with AI, then read it out loud. If it sounds stiff, it is.
- Ask for one fix per prompt: source this stat, chart this table, translate this page.
- Prioritize what leaders care about. Put risk, cost, and timing up front.
- Make the punch line impossible to miss. Use size and color to guide the eye.
- Cite everything that could be challenged in a meeting.
Do those steps and your deck stops being a slide show. It becomes a decision tool.
Final Thought
AI won’t replace the skills that matter most: critical thinking, clean structure, and brave editing. Jeff shows how to cut noise and keep control. I’m convinced the winners will treat AI as a speed boost, then double down on taste and truth.
My challenge to you: ship a “version one” in minutes, then spend your energy on three passes—story, proof, and design. Ask your tool to go fast. Ask yourself to go deep. Your audience will feel the difference.
