Latest / The Series A Show with Fexingo: Early-Stage Funding, Pitch Decks, and Startup Milestones / How Startups Build AI Agents to Automate Investor Outreach
Transcript
- Lucas: So it's mid-2026 and we are seeing a very specific fundraising trend that I think is going to define the next couple of years. Startups are building AI agents to automate investor outreach — not just sending a blast of generic emails, but actually having an AI system research each VC, draft a personalized note, and even follow up at the right time. Luna: Right, and I've heard some founders are seeing reply rates that are way above what they used to get from manual outreach. What's the real story here? Lucas: The real story is a SaaS company called BeaconLayer — not a household name, but they raised a $3 million seed round last month, and their founder told me they used a custom AI agent to send out 380 personalized investor emails. They got a 40 percent reply rate. That is absurdly high. Luna: Forty percent? I mean, most cold email campaigns for fundraising get maybe 5 to 10 percent. So what were they doing differently? Lucas: The AI didn't just swap in a name. It read the VC's recent blog posts, portfolio companies, and even their Twitter activity. Then it generated a three-sentence opener that was genuinely specific. One investor told me they got an email that referenced a talk they gave at a conference three years ago. That grabbed attention. Luna: But doesn't that feel a little… creepy? Like, if I'm a VC and I realize a bot researched my entire digital footprint, I might be weirded out. Lucas: That's the tension, and it's real. Some VCs have started flagging ai generated outreach as a red flag. But the founders who are doing it well are transparent about it. They'll say something like, 'My AI assistant pulled up your work on supply chain AI — that's why I'm reaching out.' It's honest, and it saves time. Luna: So the key is disclosure. I've also seen some founders using tools like Clay or Apollo to layer in data enrichment, then running it through GPT-4o or Claude to write the draft. But BeaconLayer built their own agent? Lucas: They used a combination. They started with a tool called Relevance AI to build a custom workflow, then used OpenAI's API for the language model. The founder said the total cost was about $200 for the outreach campaign. Compare that to the hours they would have spent manually. Luna: And the result was a $3 million round. So the ROI is clear. But I wonder — does this work for later stages? Like, Series A or B? Or is it mostly for seed where you're cold emailing more? Lucas: I think it's most powerful at seed, because you have a longer list of potential investors and less warm intros. At later stages, you're usually going through warm intros from existing investors. But even then, I've seen Series A startups use AI to personalize their investor update newsletters, not just cold outreach. Luna: Speaking of newsletters, there's a related trend I saw — startups using AI to segment their investor update list and send different messages to angels versus VCs versus strategic investors. That seems like a no-brainer. Lucas: Exactly. And let's tie this to the market environment. If you look at the ARKK ETF, it's up about 6 percent in the last five days. That signals renewed appetite for risk and growth. So if you're a founder, now might be a good time to ramp up outreach. But you need to do it efficiently. Luna: Right, and the efficiency gain is real. But there's a risk of over-automation. If every founder sends ai generated emails, VCs will start tuning them out. So the differentiation might shift to how good your AI prompt is. Lucas: That's exactly what's happening. The founders who treat the AI as a junior researcher — not a writer — are winning. They have the AI gather intel, then they personally add the final touch. One founder told me he always reads the AI draft and changes at least one sentence to make it sound like him. Luna: So the human touch still matters. I think that's the key takeaway. But let's talk about a concrete example of what the AI actually writes. Can you give me a sample opener that BeaconLayer might have used? Lucas: Sure. So imagine a VC who invested in a robotics startup last year. The AI might write: 'I saw your investment in Dexterity Robotics and noticed your focus on warehouse automation. We're building a similar stack for inventory tracking, but with a computer vision twist. Would love to share our beta with you.' It's specific, it's relevant, and it's not generic. Luna: That's good. And it's not long. I think VCs appreciate brevity. So the AI also helps keep emails short. Lucas: Absolutely. And the data backs that up. According to a study by DocSend, the optimal pitch deck length is 10 slides. Similarly, the optimal cold email length is under 100 words. AI can enforce that discipline. Luna: Now, what about the downside? I've heard stories of AI sending emails to the wrong person, or getting the tone completely off. Like, being too casual with a very formal VC. Lucas: Yeah, that happens. BeaconLayer's founder said they had to iterate on the prompt several times to avoid overly enthusiastic language. The AI initially wrote things like 'I'm obsessed with your work' — which is fine for some VCs, but weird for others. So they added a tone parameter: 'professional but warm.' Luna: So it's a learning process. I also wonder about the legal side — is there any issue with scraping VC data? Like, if an AI reads their LinkedIn profile, is that okay? Lucas: That's a grey area. Publicly available information is generally fair game, but some VCs have started adding clauses to their websites prohibiting scraping. So founders need to be careful. The smart ones use data from public sources like Crunchbase or PitchBook, which are meant for this use. Luna: Good point. So let's look forward. Do you think this trend will become standard practice within a year? Lucas: I think it's already becoming standard. I'd estimate that 30 to 40 percent of seed-stage startups are using some form of AI for outreach today. By next year, it'll be the norm. The question is how VCs adapt. Some might create AI filters of their own. Luna: So we might see an AI arms race between founders and VCs. That's fascinating. But for now, the founders who use it thoughtfully have a clear edge. Lucas: Look, I think it's worth mentioning — and this ties back to why we can spend time on these trends — that this show is ad-free and listener-supported. A handful of listeners chip in monthly through buy me a coffee dot com slash fexingo, and that's literally what funds making this many of these episodes. If you find these conversations useful for what you're building or running, that's where you can help keep it going. Luna: Yeah, it's a small group that makes it possible. And we really appreciate it. It keeps us independent and focused on what matters to founders. Lucas: Exactly. So back to the AI outreach trend — one more data point. I spoke with a VC at a top firm who said they've seen a 300 percent increase in ai generated cold emails this year. They're not ignoring them, but they're more skeptical. So the bar for quality is rising. Luna: That makes sense. So the advice is: don't just press 'send' on an ai generated campaign. You have to review, customize, and make it human. The AI is a tool, not a replacement. Lucas: Right. And the best founders are using AI to do the research that they would have done anyway, just faster. Then they add their personal story. That combination is hard to beat. Luna: Alright, so to wrap up: AI agent for investor outreach — use it, but use it wisely. Keep it personal, keep it short, and be transparent. Any final thought? Lucas: One more thing. Don't forget that fundraising is still about relationships. AI can get you the meeting, but it can't close the deal. That's still on you. Luna: Well said. That's a wrap for this episode. See you next time.