Aravind Srinivas, cofounder and CEO of Perplexity, says he stopped using traditional pitch decks when raising money. Instead, he writes a memo, invites investors to ask questions, and lets Perplexity itself answer many of them
He admits that the only time he put together a deck was for Perplexity’s Series A. Now, for subsequent rounds, he handles questions via memo and interactive sessions. If investors send detailed follow-ups, he copies their emails into Perplexity and asks it to reply “like Aravind” then sends the AI-generated response to the investor Often, that suffices This method is unusual in Silicon Valley but well suited to Srinivas’s world. Perplexity is an AI search engine backed by top investors like SoftBank, Nvidia, and Jeff Bezos. Its product is tied closely to knowledge retrieval and automated Q&A using it to pitch itself becomes a symbol as much as a tactic
From a journalism standpoint, this raises interesting issues
It blurs the line between product demonstration and fundraising. The pitch becomes both statement and tool
It’s a confidence bet. Srinivas assumes the AI is reliable, articulate, and aligned enough to represent the business correctly. If the AI errs in tone, metrics, or nuance, the misstep could cost investor trust
It signals a shift in investor expectations. If AI can answer deep diligence questions, the emphasis may shift from polished slide decks to real-time interaction, model credibility, and foundational data
Risk of overreliance. The more you let AI mediate your investor communication, the less direct human narrative you control. That may work while things go well but when challenges arise, human judgment and voice matter more
Srinivas is making a strong statement about how AI should change even the culture of startups and investing. Whether this becomes a trend depends partly on how Perplexity fares and partly on whether investors at scale accept “AI-mediated pitch” as credible. If it works, we might see more founders dropping slides and relying on memos plus AI interfaces. If it fails, it will be a cautionary experiment in letting tech speak for the business