AI Writes the Deck. Who Writes the Creative Strategy?
The deck is the easy part now
Ask an AI tool to build a pitch deck and it will. Type a prompt and clean slides come out in minutes — layout, charts, a cover. For the deck itself, the problem is solved.
So open one and notice what's missing. The slides are there, filled with text, but there's no argument running through them. You can't say why this slide comes before that one, or what the deck is actually arguing. Each slide is individually plausible and collectively says nothing.
That's not a knock on the tools. It's a sign they solve a different problem than the one that makes a new-business pitch hard. The slow, expensive part of a pitch was never the slide design. It's the creative strategy — the angle a client actually buys. That part is still waiting.
Agencies already know where AI stops
This isn't a theory. It's how US agencies are already using AI, and where they've drawn the line in practice.
Agencies have moved AI deep into the early work of a pitch — research, audience profiling, competitive analysis, first-draft briefs. Havas built it into its Converge platform against client data; Golin and The Brooklyn Brothers use Claude for initial synthesis before narrowing to company profiles. The appeal is speed: new business is the sweet spot, where a team often has about four weeks to turn a pitch around. (How agencies are using AI for strategy and client pitches, Digiday)
But listen to how agency strategists describe the limit. Golin's chief strategy officer frames it as a shift "from efficiency ... to efficacy." Translation: AI handles the efficiency — compressing hours of research and formatting — while the efficacy, the judgment about what the work should actually argue, stays with the strategist. AI gets used for scale and speed. It does not get handed the core idea.
That's the same line, said a dozen different ways across the industry: use AI for the deck and the digging, not for the point of view.
The container versus what goes in it
Here's why the line falls exactly there. Any persuasive pitch has the same spine — start from the background, name the problem, work through an interpretation, land on a conclusion. The slides are the container that holds that line of reasoning. They are not the reasoning.
AI deck tools automate the container. What goes in it — which angle, in which order, from which point of view — is still a person staring at a blank screen. And in advertising that part carries more weight than anywhere else. A consulting deck persuades with data and analysis. An advertising pitch persuades with a point of view: the same market, the same target, the same brand can yield completely different directions depending on the angle you choose. Choosing and shaping that angle is the whole job — and it's the one thing a tool that averages everything it has read can't do.
Even in consulting — where decks lean on data, not a point of view — the gap shows up. An ex-McKinsey consultant ran five AI presentation tools through the real process of building a deck, broken into six steps: define the audience and outcome, structure the storyline, sketch slides and headlines, fill in content and data, edit, and polish. The tools helped with the early scaffolding. On the content, they failed — the generated slides, he found, take more time to fix than building them from scratch, so you keep the skeleton and throw out everything it filled in, because the real content has to be built against the facts of the actual project. (An ex-McKinsey consultant's 5-tool experiment, slideworks.io)
If that's true where the work is data-driven, it's truer in advertising, where the whole pitch *is* the point of view. The deck got faster. The thinking didn't.
Why asking ChatGPT doesn't close the gap
If the deck tools only do the container, the obvious move is to ask a general-purpose AI for the content — the strategy itself. Agencies try it, and it comes back thin.
A general model predicts the most likely next words; it doesn't reason toward a creative direction. It has no stake in the brand and no angle to commit to, so it hands you something that sounds strategic and argues nothing. You can feel it the moment you try to take it into a room: the pieces are there, but there's no through-line holding them together.
What gets compressed, and what doesn't
A new-business pitch can cost an agency weeks of senior planner time — and most of it isn't spent on slides. It's spent finding the angle: the one reading of the brief the whole pitch will stand on. That's the expensive part, and it's the part that decides whether you win.
That's the gap Synapsis works in. You enter four fields from the brief — brand, objective, background, target — and it designs three strategic directions, each with its own angle, its own line of argument, and the campaign message it lands on. It isn't making the slides. It's making the creative strategy the slides will carry — so a planner spends the saved hours sharpening the best direction instead of starting from a blank page. Which direction you take, and the final call on the idea, stays yours. That's the line agencies already draw; Synapsis is built to sit on the right side of it.
Slides are already fast enough. It's the strategy's turn to get faster.
Read more
- How the engine designs that line of reasoning: Can AI Write a Creative Strategy for a Pitch?
- Why it stops at strategy and doesn't hand you an execution plan: You need a strategic direction — so why do you get an execution plan?
- The same experiment, run against a general-purpose AI: Same brief, same AI. Why were the results different?
Frequently asked questions
Can AI write a creative strategy for a pitch?
Not the way most people ask it to. A general-purpose AI averages everything it has read and has no angle to commit to, so the result sounds strategic and argues nothing. Deck tools, meanwhile, only build the slides. Designing the creative strategy — the angle the pitch stands on — is a separate problem, and the one Synapsis is built for.
How do ad agencies actually use AI in a new-business pitch?
Mostly for the deck and the digging: layout, formatting, first-draft slides, research, audience profiling. The strategy — the angle you'll actually pitch — still gets built by a planner or creative strategist. Agency strategists describe it as using AI for efficiency, not efficacy: speed on the inputs, human judgment on the idea.
What's the slowest part of building a pitch?
Not the slide design. It's finding and shaping the angle — the creative strategy the whole pitch stands on. That's where the senior hours go, and it's the part AI deck tools leave untouched.
How is a strategy tool different from an AI deck generator?
A deck generator produces the container: formatted slides. A strategy tool produces what goes in it: the creative strategy that gets a client to the conclusion. Synapsis reads a campaign brief and designs three strategic directions with different angles — it designs the thinking, not the layout.