The Perfect Elevator Pitch in the Age of AI

Every startup founder needs to have a good elevator pitch ready. The elevator pitch sums up your company’s purpose and starts a conversation with investors, clients, customers or anyone else you want to do business with.

The idea of the elevator pitch harkens back to a time when founders had to introduce their company to someone they serendipitously found standing next to them in an elevator—with only a minute or two to make a great impression before the doors opened. But nowadays, you’re just as likely to present a pitch online or at a competition.

The tools used in crafting a pitch are also changing. It often makes sense now to draw on AI for inspiration or to generate multiple versions of a pitch.

What goes into the perfect pitch?

An elevator pitch should be concise and catchy, with an opening sentence that grabs listeners’ attention. Humans love listening to stories, so the ideal pitch is structured as a short narrative with a conflict and a happily-ever-after resolution.

A winning elevator pitch needs at least two components:

The problem you’re trying to solve. Whoever’s listening should be convinced that this problem is painful and needs to be fixed—even if it doesn’t affect them personally. If it does affect them, they should see that you understand what they’re dealing with.
Your solution. You want to show that your company has the answer and is the right one to implement it. This might mean highlighting a unique approach or advantages that differentiate you from competitors.

If you’re addressing an investor, you also want to prove that there’s a market for your product or service.

AI can help startups escape the “operational time suck” of customizing pitches.

Leveraging AI to craft a pitch

To work on a pitch with AI, you’ll need to enter questions or commands known as prompts. Examples include, “Write a one-paragraph pitch for my company’s product based on this document” or, “Here’s a description of what my company does. How would you word this as a pitch to investors?”

Then, chat with the AI to refine the pitch.

George Burgess, founder and CEO of Modern Day Talent, says that AI has been especially useful in writing pitches. That’s because George is a solo founder. “I often treat AI almost as a second person, and I can then brainstorm with it. I can go back and forth and think about different ideas and try different things really quickly,” he says.

Burgess has gotten the best results when he asks the AI for multiple options, then tells it which ones he likes or dislikes and why. “I'll say, ‘Come up with ten ways I could describe this in a succinct manner.’ It'll give me ten, and I might say, ‘Okay, try ten that are more creative.’ Or, ‘I like this angle,’” he says.

Burgess sometimes rewrites a pitch himself, then asks the AI to improve on his wording. For example, an AI-generated pitch to customers included the phrasing:

"We connect you with South Africa's top 0.1% of talent, rigorously selected to build remote teams for Europe's fastest-growing companies. Enjoy the benefits of superior skills, English fluency, and alignment with European time zones."

Burgess rewrote that in a more conversational style:

“We build remote teams in South Africa for Europe’s fastest-growing companies, saving them 30-40% on payroll. We hire just 0.1% of applicants, ensuring you get the very best. Why South Africa? It’s English-speaking, and in a European time zone.”

He then asked the AI to polish it and got this output, which emphasizes how the customer benefits:

“We assemble elite remote teams in South Africa for Europe's fastest-growing companies, offering a 30-40% payroll saving. By hiring only the top 0.1% of applicants, we guarantee exceptional talent for your business. Why South Africa? With its English-speaking workforce and convenient European time zone, it's an ideal choice.”

Oliver King-Smith, CEO of smartR.AI, recommends feeding the AI lots of documentation on your company so it can help you distill that material into a pitch. “More information is good, and you don't need to worry about being too wordy,” he says.

Try out a few AI tools to see which ones you prefer. Available platforms include:

ChatGPT. OpenAI’s chatbot took the world by storm in late 2022, and regular updates have expanded its capabilities with different formats like audio and images.
Gemini. Google’s AI model works with a variety of inputs (including images or handwriting) and is designed to be able to explain its reasoning.
Claude. This AI assistant from Anthropic touts its ease of chat and safeguards against harmful content.
Perplexity. This search engine draws from current information online to answer research queries and presents the results in prose format, with citations.

Personalizing the pitch

To be effective, a pitch should be tailored to the person you’re talking to. While it’s hard for one founder or a small team to make time for extensive preparation before each pitch, an AI can speed up the process.

Yuri Cataldo, co-founder and general partner at Athenian Capital, suggests asking AI to analyze an investor’s past moves. “You can have it do some research on the individuals and say, ‘I'm pitching to this person.’ Let's just say it's Reid Hoffman. ‘What types of startups does he have a tendency of investing in? And so what words should I bring up?’” he says.

King-Smith gives an AI model information about prospective clients, then asks it to incorporate the details into a LinkedIn outreach message. “We want to get it relatively short, but super specific to what their needs are,” he says. When prospects see a personalized pitch instead of a generic greeting, they’re more likely to agree to a call and continue the conversation.

Bob Bilbruck, CEO of Captjur, says AI can help startups escape the “operational time suck” of customizing pitches. Presenting an individualized pitch is especially crucial when approaching investors, because they all evaluate startups from different angles. “It has to speak to the levels of investments that they may be making,” he says. “If a VC is a pre-seed to early-stage investor company, then the elevator pitch has to speak to that. If it's a family office, it might be a totally different pitch because a family office is usually involved in investing in longer-term type investments.”

Man standing in front of people sitting beside table with laptop computers

Learning from dynamic feedback

You can use AI to practice delivering an elevator pitch or to get ready for a presentation.

One option is to ask AI to analyze your pitch before an event. Cataldo once worked with a startup that was preparing for a pitch competition focused on sustainability. “When I read their pitch deck, I was like, ‘Listen, you've totally missed the sustainability part about this. So you need to rewrite this.’ And so we used some AI tools to help them take their existing information and then point out the sustainability element,” he says.

AI can also play the part of devil’s advocate and help you spot weaknesses in a presentation. “You could actually do some prompts or inputs within most AI systems that say, ‘Hey, find the holes or find the problems with this pitch where investors might throw up objections,’ and then give alternative approaches and positioning to overcome these objections,” Bilbruck says.

And you can ask an AI model to roleplay with you. You can input your pitch, then have the AI respond like a skeptical audience member while you chat with it. Running this simulation a few times can help you see different ways the conversation might go.

Mistakes to watch out for

As you start using AI for pitching, you want to steer clear of some common pitfalls.

First, don’t give up just because an AI model spits out mediocre verbiage on the first try. To get a useful pitch, you need to engage with the program and iterate.

“I work with a lot of people and introduce them to some of these tools, and the number one issue I find is they'll say, ‘Oh, well I asked it to write my elevator pitch and I didn't really like the results.’ But they gave it no feedback. They didn't tell it what they didn't like,” Burgess says.

You also can’t count on AI to do all the heavy lifting. You still need to have a solid understanding of your product and customers. Your knowledge should inform the prompts you’re entering, and you should be able to edit the AI’s work with the perspective of an expert.

“I've seen people who don't really understand what they're doing; they'll just rely on AI to fill in the gaps for them. And it really shows that they don't actually know their market,” Cataldo says.

Watch out for “hallucinations,” or fibs that the AI may insert into otherwise flawless prose. AI models usually sound 100% confident in their assertions, so you have to approach their output with a sense of skepticism.

If a claim sounds too good to be true, verify it with a non-AI source. “Don't believe it until you've checked it yourself personally. You could end up with serious embarrassment if you're unlucky on that one,” King-Smith says.

And be careful about sharing proprietary data with a public-facing AI, which might reveal it to the next user. This is something to watch out for if you’ve signed an NDA with a client.

AI can’t do all the work of crafting an elevator pitch, but it can help with ideation, research, phrasing and more. Experiment to see where it makes the process easier, and always add a human touch to your pitch.