Latest / The Series A Show with Fexingo: Early-Stage Funding, Pitch Decks, and Startup Milestones / Startups Are Using AI to Personalize Pitch Decks
Transcript
- Lucas: So there's this startup, let's call it FinFlow — a small fintech out of Austin — that raised its Series A back in March. But the interesting part isn't the round itself. It's how they got there. Luna: I think I know where this is going. They used AI to tailor their pitch deck for each investor, didn't they? Lucas: Exactly. And they're not alone. I've been tracking a growing trend where startups are using large language models and investor CRM data to generate a unique deck for every single investor they meet. FinFlow's founder told me they created over 80 different versions of their deck. Luna: Eighty versions? That's a lot of time. But I guess if AI is doing the heavy lifting, it's not that crazy. Lucas: Right. The AI takes the core content — the problem, solution, market size — and then reorders it, rephrases case studies, even tweaks financial projections based on the investor's portfolio. If an investor loves SaaS, you lead with your subscription metrics. If they're big on AI, you highlight your machine learning pipeline. Luna: So it's like a smart template that adapts. But how do they get the investor data? Are they scraping public info? Lucas: Some are. But the smarter ones use a CRM — something like Affinity or even a custom Notion database — where they log every conversation with an investor, their published interviews, and their LinkedIn activity. Then the LLM pulls from that. FinFlow told me they increased their meeting-to-term-sheet conversion by about 40 percent. Luna: Forty percent is massive. But I have to wonder — does the deck still feel like the founder? If the AI writes it, doesn't it lose that authentic spark investors look for? Lucas: That's the tension. The best founders I've seen using this treat the ai generated version as a first draft. They then go through and inject their voice, their stories. The personalization is more about structuring the narrative than dictating every word. Luna: Okay, so it's a tool, not a crutch. But what about the data privacy angle? Are investors comfortable with founders scraping their public profiles and feeding them into an LLM to craft a pitch? Lucas: It's a gray area. Public information is public, but using it to manipulate a pitch feels different. Some investors have started asking founders outright, 'Did an AI write this deck?' And if they get a yes without nuance, it can backfire. Luna: I can see that. There's a trust element. If I'm an investor and I feel like the deck is too tailored, I might think the founder is hiding something or trying to game me. Lucas: Exactly. So the smart approach is transparency. One founder I spoke with includes a small disclaimer at the bottom: 'This deck was personalized with the help of AI to highlight aspects most relevant to your fund.' That honesty actually improved rapport. Luna: Interesting. So it's not about tricking anyone — it's about efficiency and relevance. But is this only for Series A and later? What about seed-stage founders with limited budgets? Lucas: Seed-stage founders are actually the biggest users. They don't have a network of warm intros, so they're sending cold decks to dozens of investors. Personalization can be the difference between a meeting and a pass. There are now tools like DeckMatch and PitchAI that cost under $100 a month. Luna: That's accessible. And the numbers back it up. I saw that the ARKK fund is up 6 percent in the last five days — investors are clearly hungry for innovation. But they're also being more selective. A personalized deck might be the edge. Lucas: Right, and look at Palantir, trading around $128. The market rewards companies that understand data-driven decision-making. So it's no surprise that startups are applying that same philosophy to their fundraising. Luna: But let's talk about the risks beyond authenticity. What if the AI hallucinates? Generates a financial projection that doesn't add up, or references a case study that's fabricated? Lucas: That's a real danger. Every founder I've talked to who uses this has a human-in-the-loop review step. They check every number, every claim. The AI is a drafting tool, not a final editor. If a founder sends out a deck with a hallucinated number, they lose credibility instantly. Luna: So the rule is: always proofread. But there's also the issue of version control. If you have 80 versions, how do you keep track of which investor saw which slide? You don't want to accidentally send a deck praising a competitor's thesis to an investor who hates that space. Lucas: Great point. Some teams use a naming convention — like 'FinFlow_Sequoia_v3' — and store everything in a shared drive with a log. Others use dedicated software that automatically tracks versions per investor. It's a new operational challenge, but manageable. Luna: It sounds like the startups that do this well treat it like a systematic process, not a one-time hack. So what's the bottom line for a founder listening right now? Should they jump on this trend? Lucas: I'd say start small. Pick your top five target investors. Research them manually, feed the data into an AI tool, and generate a custom deck for each. See if the response rate improves. If it does, scale up. But never skip the human review, and never lose your voice. Luna: And maybe add that disclaimer. Trust is currency in fundraising. Lucas: Exactly. Speaking of trust, I want to mention something quickly. If these conversations are useful for what you're building or running, we keep this show ad-free thanks to listener support. If today's episode was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. No pressure, just a small way to keep the insights flowing. Luna: Yeah, it really helps us keep doing the deep dives without any sponsor noise. Back to the topic — one more thing on AI decks: I've heard that some founders are now using AI to also personalize their verbal pitch, not just the slides. They have a chatbot that suggests talking points based on the investor's background during the meeting. Lucas: That's next-level. But I worry about over-automation. If a founder is reading off a screen during a pitch, they lose the human connection. The deck should be the map, not the script. Luna: Agreed. So the key takeaway: AI personalization is a powerful tool for early-stage fundraising, but it works best when it amplifies the founder's story, not replaces it. Lucas: Right. And as the market evolves — with funds like ARKK showing renewed risk appetite — founders who can efficiently demonstrate relevance will stand out. Just make sure the data is right, the voice is real, and the ethics are solid.