Latest / Tech IPO Conversations with Fexingo: Public Software Companies and Capital Markets / How Post-IPO Companies Use AI Agents for Investor Relations
Transcript
- Lucas: Mark Zuckerberg predicted this week that within five years, billions of people will have personal AI agents. Most of the coverage focused on consumers — but there's a quieter shift happening inside public companies that I think is just as significant. Luna: You're talking about investor relations teams using AI agents, right? I've seen a few announcements from IR software vendors. Lucas: Exactly. A growing number of post ipo software companies are experimenting with ai powered IR assistants — tools that can handle routine investor queries, summarize analyst questions, draft responses, and even track sentiment across the shareholder base. It's still early, but the trajectory is clear. Luna: So it's like a chatbot for investors? But that raises obvious compliance red flags — selective disclosure, material non-public information, that whole world. Lucas: You're right to flag that. The SEC's Regulation Fair Disclosure — Reg FD — says if you share material information with one investor, you have to share it with everyone. So these AI tools are being built with strict guardrails. They're typically trained only on public information — earnings call transcripts, investor decks, press releases — and they're programmed to deflect any question that touches on forward-looking guidance or non-public data. Luna: But who's liable if the AI slips up? The company? The software vendor? That's the part that keeps general counsels up at night. Lucas: It's a valid concern. Most implementations I've seen include a human-in-the-loop — every response gets reviewed by an IR officer before it goes out, at least for now. And the AI logs every interaction, so there's an audit trail. The key is that the tool augments the IR team, it doesn't replace them. Luna: That makes sense. And I imagine for a post ipo company with a lean IR team — maybe two or three people — an AI agent could be a force multiplier. Lucas: Absolutely. Consider a company that just went public and gets hundreds of inbound questions from retail investors on platforms like Reddit or StockTwits. An AI agent can triage those, spot trending concerns, and draft a FAQ that the team can approve and publish. It frees up the IR team to focus on the big institutional relationships. Luna: So it's not just about efficiency — it's about better capital allocation within the IR function itself. Lucas: If today was actually useful to you, the way these episodes stay ad-free is listener support — buy me a coffee dot com slash fexingo. It's a small way to keep the conversation independent. Luna: Yeah, totally. If this show helps you think about your own company or investments, that's exactly why we do it. And listener support is what keeps the lights on without ads. Lucas: So back to the AI agent trend — I want to talk about a concrete example. A friend of mine works in IR at a post ipo SaaS company, and they recently built an internal LLM trained on their last four years of earnings call transcripts. Luna: Four years of transcripts — that's a lot of data. What do they use it for? Lucas: Primarily to prep for earnings calls. Before the call, the IR team uploads the sell-side analyst models and the AI generates a summary of likely questions — based on past calls, current macro context, and the company's recent guidance. It's like having a dedicated research assistant. Luna: So it's not just reactive — it's predictive. That could help the CEO and CFO be more prepared, especially if they're new to the public stage. Lucas: Exactly. And it's not just earnings prep. Some companies use AI agents for sentiment analysis — they scrap sell-side reports, buy-side notes, and social media chatter to gauge how the market is reacting to news. That can help the IR team prioritize which analysts to engage with. Luna: I wonder how much of this activity is disclosed. Are investors aware that their questions might be answered by an AI? Lucas: It's a gray area. Some companies are transparent — they'll note in their IR contact page that they use AI tools. Others don't mention it. I think over time, best practice will be to disclose it, especially since the AI is only using public information anyway. The bigger risk is if investors feel misled. Luna: And there's also the question of AI bias. If the model is trained on past earnings calls, it might amplify certain framing — maybe overemphasizing growth metrics and underplaying risks. Lucas: That's a smart point. Bias is a real issue, especially with off-the-shelf large language models. That's why some companies are building their own fine-tuned models on their specific datasets, and they run regular audits to check for drift or skewed outputs. Luna: Let's talk about cost. For a post ipo company that's still not profitable, adding an AI IR tool — even a modest one — can be a meaningful expense. Is the ROI there? Lucas: It depends on the scale. For a company with a high retail investor base — like some of the consumer-facing tech IPOs — the volume of inbound queries can be huge. An AI agent can handle that for a fraction of the cost of hiring additional IR staff. And there's also the intangible benefit of consistency: the AI gives the same answer every time, reducing the risk of accidental selective disclosure. Luna: So in a way, the AI becomes a compliance tool as much as a convenience. Lucas: Exactly. But it's not a set-it-and-forget-it solution. The companies I've seen doing this well have a dedicated person — often the IR director — overseeing the AI's outputs and updating its knowledge base every quarter. Luna: I'm curious — have any of the big tech companies like Microsoft or Meta started using AI agents for their own IR? They certainly have the resources. Lucas: I haven't seen public confirmation from Microsoft or Meta, but it's likely they're at least experimenting. After all, Microsoft just reported a $3.2 billion gain from its Anthropic investment — that's the kind of complex story an AI agent could help communicate consistently across different investor touchpoints. Luna: Right — that investment gain is a nuanced narrative. You don't want one investor hearing a different version than another. Lucas: Precisely. And that's the promise of AI in IR: uniform, accurate, timely communication. But it requires careful calibration. I think in three years, we'll look back and wonder how IR teams functioned without AI assistance. Luna: As long as they keep the human in the loop. I don't think investors want to feel like they're talking to a machine about their money. Lucas: No, they don't. And that's the final piece — the emotional aspect. A good IR officer builds trust over time. An AI can handle the facts, but it can't replicate that relationship. So the teams that succeed will be the ones that use AI as a tool, not a crutch. Luna: So the IR job evolves: less time answering basic questions, more time on strategy and relationships. That sounds like an upgrade. Lucas: I think it is. And for post ipo companies especially, where resources are tight and the spotlight is bright, any edge in investor communication is worth exploring. The question is how fast the rest of the market catches up. Luna: Thanks, Lucas. That's a great look at a trend that's still under the radar but won't be for long. Lucas: Thanks for listening. If this made you think about your own IR strategy, I'd love to hear how your company is approaching AI — you can always reach us through the show notes.