AI Strategy: Why Perplexity's CEO Replaces Pitch Decks with AI Talks
Pitch decks are theater. Aravind Srinivas, Co-Founder and CEO of Perplexity, replaced them with something more accountable: a short memo, a live product demo, and an extended Q&A where investors can ask anything and see the metrics pulled in real time.
His stance is simple: the only time he built a pitch deck was for the Series A. After that, he moved to direct conversations supported by the product itself.
The shift: from slides to signal
Instead of weeks lost to slide polish, Srinivas writes a memo, blocks two hours, and invites investors to interrogate the business. Metrics come up live. Follow-ups get answered on the spot.
If a question goes beyond what's on hand, he points them to Perplexity and says: ask it. In his words, "it already knows everything."
AI as co-presenter
On a recent deal, an investor sent a dense thread of follow-ups. Srinivas pasted the email into Perplexity with the instruction: "answer it like Aravind."
Because the platform has ingested years of his interviews and posts, the response reflected his thinking clearly and consistently. He reviewed it and concluded he wouldn't have done a better job by hand.
Traction that backs the method
- Active users: 22-30 million
- Growth: ~20% month-over-month
- Scale target: "a billion queries a week" if the pace holds
- Day one (2022): 3,000 queries in a single day
Perplexity positions itself as an answer engine. Instead of sending people down link trails, it delivers direct responses and sources. The company is also prioritizing features for business customers while usage scales.
Operator principles you can apply
- Replace decks with proof: Share a memo, show the product, and let Q&A do the rest.
- Turn AI into your second brain: Centralize interviews, posts, docs, and let AI synthesize consistent, context-rich answers.
- Instrument everything: Be ready to pull metrics live. If you can't show it, it likely doesn't exist.
- Invite scrutiny: Encourage investors and customers to test assumptions in real time.
- Publish your thinking: The more public signal you create, the better your AI can reflect your stance under pressure.
- Bias to enterprise value: Ship features that matter to business users and prove ROI quickly.
From researcher to CEO
Srinivas moved from research roles at OpenAI, DeepMind, and Google Brain to building a company that operates on first principles. He references a simple fork often attributed to Larry Page: do impactful research or build a company that drives impact at scale. He chose the latter, without the slides.
What this means for your strategy
Make your product the narrative. Replace presentation polish with observable truth. Train your team to use AI as a real-time explainer, not a novelty. Publish more, so your AI has a public memory to work with. And hold yourself to metrics you can surface on demand.
If you're building an executive capability around AI for sales, investor relations, or ops, consider structured training for your team to shorten the learning curve. See our curated tracks for leaders here: AI courses by job.
Further reading
- Berkeley Haas - context for the conversation format and leadership approach
- Bloomberg Tech Summit - remarks on growth trajectory and product scale
Your membership also unlocks:
