It’s not because it’s AI-generated. It’s because it misses the whole point of a pitch deck.
It was pitch day yesterday for two of my organizations. I listened to 15 pitches. At least 10 of the pitch decks were created with AI. And we immediately rejected all of them.
The pitches weren’t rejected because they were AI. They were rejected because they were bad pitches. Although they were nicely formatted, they failed to engage investors.
The AI pitches were full of information, far too much information in fact, but not the right information that investors need to assess an early-stage startup. They were full of graphics, icons, and splashes of color, but not graphics that explained the business and often made the story more confusing.
My simple advice: if you want investment — don’t use AI to generate your pitch deck.
Don’t get me wrong — I love AI. I built an AI coding startup a decade ago and have invested in many AI startups. I use AI daily for translation (extremely helpful), for startup diligence (life changing), for proving my wife wrong (yeah, don’t do that unless you like sleeping on the sofa). AI is a gamechanger for so many activities. But, at least for now, not for creating a pitch deck.
I get it. Writing pitch decks is hard. It’s time consuming when you need the time to build the product and talk to customers. But there are no easy shortcuts. The job of creating the pitch deck is the job of explaining your business, of distilling all the things you can do to a few short sentences that resonate with investors. That can’t be handed off to a graphics designer or a pitch deck writer on Upwork. And it can’t be handed off to Claude.
Good, But Not Good Enough
Your best friends Claude and Chatty make pitch decks that actually aren’t bad. In fact, they’re probably better than 50% of the human-created pitch decks I see. But…
That’s not good enough. We don’t invest in 50% of startups. We invest in the top 2%. And AI won’t get you there. Not even close.
Here are the specific reasons why AI-generated pitch decks don’t cut it.
Data Dump Instead of Story
If you give AI a data dump about your startup, it will produce a pitch deck that’s a nicely formatted data dump. And that’s not what investors need.
First rule for a good pitch deck is to tell a story. Not a story of how great your product is but how an investment in your company will lead to a 100x return for investors. It’s a sales pitch to convince investors to buy stock in your business.
I don’t know why AI can’t understand that, but it doesn’t. It can output problem, solution, market size, competition slides, etc. but it can’t generate a message of why this is a good investment. Instead, it outputs lots of words that say lots of things but doesn’t come together into a coherent story of generating return for investors.
Granted, most early-stage pitches do a poor job of that, too, but those pitches don’t get investment either.
Lack of Clarity
I don’t know why Chatty and Claude can’t read my many articles (plus those by many other pitch deck experts) that advise you to limit text to 30 words or less per slide. But they don’t.
You have only 10 minutes to present or 1 minute for a reader to skim the deck. The more we have to read, the less we listen to you. The more we’re squinting at fine print, the less we’re focused on the headlines. The more text on the page, the less we see of the key takeaways.
You can go into gory detail in the 2-hour deep dive call, but first you have to get there. At the pitch stage, the story needs to be simple, clear, and succinct. Each slide should have one key takeaway and 2–3 bullet points with graphics that specifically reinforce that point.
Thirty words per slide max, 11 slides, 330 words total to tell your story. That’s not easy. As Mark Twain famously said, “I didn’t have time to write a short pitch deck, so I wrote a long one instead.”
That process of distilling your message is the key. Cutting 3000 words down to 300 forces you to decide what is important and what isn’t. Yes, there’s a dozen things your product can do and ten different ways to sell it. You have to decide which is the most important and tell us about that, because you only have time for one. AI can’t make those judgements for you. If you say you have 12 different markets, AI will make a pitch deck that describes all 12 of them in a confusing mess of capabilities instead of focusing on one critical solution to a specific problem.
The key to a good pitch deck is to polish, polish, polish until every single word serves a purpose in crafting a clear story. AI can’t do that. What it outputs is grammatically correct, but doesn’t resonate with humans. It’s great for a report, horrible for a sales presentation.
To get there, you need to try different approaches and see what works. That may be different for different audiences. The founder needs to know that. That’s why pitch decks are time-consuming and difficult, and the process of improving is never-ending.
It takes many, many iterations. Try the first draft on your friends. Try it on your parents. Try it on startup mentors and advisors. What are people understanding and where are they getting stuck? Where are you including too much detail, where does it not have enough? What’s resonating with the audience and what’s boring them?
AI can’t do that for you, and if you’re relying on AI, you aren’t thinking through all the ways you could describe your advantages and your killer advantages. You have a finished output and you won’t have a clue why it isn’t working or how to fix it.
We’re Investing in You
It’s a truism of early-stage investing that we bet on the jockey, not the horse. The 3 most important things in deciding whether to invest are team, team, and yes — team!
That doesn’t mean your resume or your LinkedIn profile. It means you. Your personality. Do we trust you? Do we believe in you? Do we feel your passion? Are we convinced you’re willing to spend the next decade battling windmills until you finally find a way to knock them over?
If you send us an AI pitch deck, then read the text to us, a well-trained monkey could do that. And we’re not investing in monkeys, no matter how well-trained they are.
What we want to hear in your pitch deck is you. Your voice. Your plans. Your resilience. Your story. Because that’s what we’re investing in. All the other details in an early-stage pitch deck will change when you pivot three times before finding product-market fit.
An AI pitch deck tells me you have no confidence in yourself, and frankly, no clue what you’re doing. You’ve thrown a nicely formatted data dump at me full of too much information about your startup, but lacking the most important element — you. Pass.
The Sameness of AI
If you read one article on Medium each week, you might not notice an article is AI-generated. In fact, you might even prefer it. If you read 10 articles a day, you’ll quickly come to hate AI slop.
It’s not that it’s bad. It’s that it all sounds the same. Every article uses the same words and phrases, the same sentence constructions, the same weird attempts to sound human. Pitch decks are the same.
If you see one, you might think, oh that’s not bad. If you see ten, they all look the same. In fact, if you see ten on a specific topic, like batteries or cures for cancer, you’ll see the same images, hear the exact same phrases. The decks are interchangeable. It’s hard to get excited that your startup will be the one success out of an ocean of sameness.
Just Don’t Do It
It’s tempting to use AI to generate a pitch deck. You think it will save you time and money, and maybe even produce a deck that’s better than you can create on your own.
AI is advancing fast and this article might be out of date in a year. But at least for now, AI generates crappy pitch decks that will kill your chances of attracting investors. So put the tokens away and open PowerPoint and sketch out the 300 words that will convince me your startup will be a huge success.
Want to make a good deck that resonates with investors? Follow my advice here: https://pitchingangels.com/pitch-deck/
