Investors receive thousands of pitch decks every year, and even those with the best of intentions will never be able to give feedback on all of them.

While founders can now turn to ChatGPT for feedback, it will likely tell them they have the best idea ever — even if they did ask the chatbot to be brutally honest — and won’t have investor-level knowledge baked into the answers.

Ada Ventures, a UK-based pre-seed investment fund backing inclusive entrepreneurs, wanted a tool to help all founders get instant, actionable feedback on their decks, before they even start pitching. They’ve launched that tool.

“Deck Genius” aims to give all early stage founders free access to VC-level expertise and insights, almost instantly.

“I think [Deck Genius] is 10x better than what most VCs would give you in terms of feedback,” says Michael Tefula, Ada Ventures principal and head of product. “And that’s because it’s running this analysis process multiple times.”

According to Tefula, who developed the product, there are 20 to 50 LLM calls running in the background at any one time to methodically review the pitch deck across three key areas: narrative, design and venture fundamentals.

Does it tell a good story? Is it visually appealing? And, most importantly, does it hit all the venture necessities? That encompasses product, team, market, go to market and everything else investors are looking for in a deck.

“Imagine having three very experienced VCs surgically break down your pitch deck,” Tefula says. 

Where it might take investors days or weeks to respond with a few lines of feedback, the tool gives founders “VC-quality feedback” in 30 to 90 seconds, with both slide level analysis and a holistic assessment of the deck.

The tool covers everything from the very granular — it might say “you’ve got a competitor’s slide, but actually you haven’t provided a thorough explanation of how you are planning to develop your moats over the long term” — to the more abstract — it might point out that something on the first slide contradicts something said on the fifth slide.

Deck Genius started out as an experiment in the form of an “AdaGPT” experiment about a year ago, which Ada Ventures released in part to see whether founders were open to engaging with AI-powered tools.

As it turned out, they were. More than 200 founders used AdaGPT every month, and one startup even ended up getting a term sheet offer from Ada Ventures off the back of it.

But the team decided it was too Ada-centric, and the goal with Deck Genius, which it hopes will process more than 500 pitch decks per month, is to be useful to all founders, not just those interested in pitching to Ada.

“We think that if you are participating in the tech ecosystem as an investor or entrepreneur, you’ve got to give lots rather than take, take, take from the ecosystem,” Tefula says. “So we want to set as many founders up for success as possible because we just think we’ll just have a much better tech ecosystem.”

“One of the founding principles of Ada Ventures is that we are an inclusive firm,” he continues. “We don’t want to be gatekeeping. Best practice on how you can build an amazing pitch deck should be out and open to everyone to use.” 

Tefula built the review tool alongside the Ada team and the help of AI coding tools such as Cursor, Claude and Codex, training it not on the decks uploaded, but on existing public material.

Founders have complete privacy when they upload their material via PDF to Deck Genius. Everything is immediately deleted after analysis unless founders opt-in for a retrieval code so they can access the information later.

The platform has been live in alpha for about a month now and is open to feedback from founders as it enters beta today. It has already taken on feedback from AdaGPT to try to eliminate bias in the tool, which has been trained on public data about startups and entrepreneurs and therefore is biased towards white men.

One user, for example, put her pitch deck through AdaGPT, and the models were very critical of her financial slides. After changing her team to say it was all men rather than all women, it was much less critical.

“That just taught us that these models are very biased,” Tefula says, “so you need to have controls in place to control for that.”

They are protecting against biases mostly through prompt engineering and ensuring the prompts think through the answers rather than immediately churning out replies.

“We found that’s really helpful,” he says. “It doesn’t eliminate the bias completely, but certainly it does a much better job than if you were not controlling for it in that way.”

Ada Ventures closed its £63m Fund II in 2024, and has invested in startups including Jack & Jill, Toothfairy, Flexa and BlakBear.

The goal is always to make the ecosystem more inclusive and, with this tool, to make the tech startup world more accessible, especially for founders without any VC connections or knowledge.

“Before you send your pitch deck out to investors, run it through Deck Genius, get a very detailed assessment of the deck, and then you can take the feedback from that, improve the pitch deck and send it off to investors such that you can actually increase your chances of engaging the people that you’re trying to raise money from,” Tefula says. 

Plus, it doesn’t matter whether those investors are at Ada Ventures or not.