Latest / Tech IPO Conversations with Fexingo: Public Software Companies and Capital Markets / How Post-IPO Companies Use AI-Powered Chatbots for Investor Relations
Transcript
- Luna: Lucas, I was poking around the investor relations page of a company that went public about eighteen months ago, and I noticed something — they have a little chat window in the corner. It's not a human. It's an AI bot. Lucas: Yeah, you're seeing more of that. A small but growing number of post ipo companies are deploying ai powered chatbots on their IR sites to handle routine analyst questions. I looked into one case — a mid-cap software firm that went public in 2024 — and their chatbot is built on the company's own earnings call transcripts, SEC filings, and investor presentations. It's basically a fine-tuned language model trained exclusively on their public disclosures. Luna: So it's not a generic ChatGPT wrapper — it's trained on the company's own data. That's interesting because the risk of hallucination is lower, but it's still a language model. What happens if it says something that contradicts a forward-looking statement? Lucas: That's the big question. The company I looked at reports that the chatbot cut their inbound IR email volume by about 30 percent in the first quarter. Analysts and investors can ask things like 'What was your revenue growth in the last quarter?' or 'When is the next earnings date?' and get instant answers. That saves the IR team a ton of time. But if the chatbot misfires — say, it implies a non-public material fact — that's a selective disclosure problem under Regulation FD. Luna: Right, Regulation FD says companies can't give material information to some investors before others. If the bot accidentally reveals something that wasn't in a press release, that's a violation. How are companies mitigating that? Lucas: Most are doing a few things. First, they restrict the chatbot's knowledge base to only public filings and transcripts that have been out for at least 24 hours. Second, they program the bot to decline answering any question that touches on forward-looking guidance or non-public metrics — it'll say something like 'I can only answer based on publicly available information.' And third, they log every conversation and have a human review flagged interactions. One IR director told me they treat the chatbot like a junior associate who can only read from the script. Luna: That's smart, but it also means the bot is only as good as the filings it's trained on. And if a company's filings are dense and full of boilerplate, the bot's answers might be vague or unhelpful. Lucas: Exactly. The quality of the answers depends on how clean and structured the training data is. Some companies are even using their earnings call Q&A transcripts as a primary source — because those contain the exact language executives used to answer analyst questions. So the bot learns to mimic the CFO's own phrasing, which builds trust with investors who hear that familiar tone. Luna: Let's talk about the current market. Palantir is up nearly 6 percent in the last five days, and it has a very passionate retail investor base. Do you think a company like that benefits more from a chatbot than, say, Airbnb, which is down a bit and has a more institutional holder mix? Lucas: Great point. Palantir's investor base includes a lot of individual shareholders who might not have direct access to the IR team. A chatbot gives them a way to get answers without having to call or email, which scales much better. Airbnb, on the other hand, has a more institutional base — analysts who are used to calling the IR director directly. For them, a chatbot might feel like a barrier. So the value proposition really depends on who your shareholders are. Luna: And there's a cost angle too. A chatbot is cheaper than hiring a junior IR associate. For a cash-burning post ipo company, that matters. Lucas: Absolutely. The company I looked at spent about $50,000 to build and deploy the chatbot, and they estimate it saves them roughly $100,000 a year in IR staffing costs. That's a solid return. But there's a subtle risk: if the chatbot becomes too good, it might replace the human touch that investors value during uncertain times. When the stock is down 20 percent, investors want to talk to a person, not a bot. Luna: That's the balance. Use the bot for the 80 percent of routine questions, but make sure the human IR team is still accessible for the tough conversations. I wonder if the SEC will eventually issue formal guidance on this. Right now, it's kind of the Wild West. Lucas: The SEC has been quiet, but I expect they'll weigh in. In the meantime, companies are being cautious. One IR consultant told me they advise clients to have a disclaimer that says 'This chatbot is for informational purposes only and does not constitute official guidance.' And they make sure the bot's responses are reviewed by legal before deployment. It's not a set-it-and-forget-it tool. Luna: And you know, this is exactly the kind of conversation that makes me appreciate why we do this show. It's not just about the big picture trends — it's the operational details that actually matter to the people building and running these companies. Lucas: Totally. And speaking of that — a quick honest thing. This podcast is ad-free, and that's by design. A small group of listeners helps make that possible through buy me a coffee dot com slash fexingo. It's not a big ask — just that if you find these conversations useful for what you're building or running, that's where the support goes. Keeps the lights on and lets us cover niche topics like AI chatbots in IR. Luna: Yeah, it's a small thing that makes a big difference for us. And we really appreciate those who do chip in. Lucas: Alright, back to the bots. Another angle I find fascinating is the potential for these chatbots to evolve from reactive Q&A to proactive outreach. Imagine a bot that not only answers questions but also pushes relevant news to investors — like 'Hey, our CEO is speaking at a conference next week, here's the webcast link.' That's a whole new level of shareholder engagement. Luna: Interesting. But that also opens up the selective disclosure can again. If the bot proactively sends information to some investors but not others, that's a problem. So it would have to be a broadcast to everyone — maybe through a newsletter or a public feed. Lucas: Right. It would need to be a one-to-many push, not a one-to-one personalized alert. But companies could use the chatbot as a front end for a broader IR content hub — earnings summaries, analyst day recordings, investor presentations. Make it the single entry point for all public information. Luna: I could see that being really useful for companies with a lot of retail shareholders who don't follow the news closely. Instead of digging through the website, they just ask the bot 'What's new?' and get a curated list. Lucas: Exactly. And the bot can track which topics get asked most often, which gives the IR team data on what investors actually care about. That's a feedback loop that human-only IR doesn't have. You can see, 'Hey, 40 percent of questions this week were about our gross margin — maybe we need to address that more clearly in the next earnings call.' Luna: That's a strong use case. It turns the chatbot from a cost saver into a strategic tool. But let's be real — most companies aren't there yet. Right now, it's mostly about handling the volume. Lucas: True. But the technology is improving fast. And as more companies experiment, we'll see best practices emerge. I think within two years, having an AI chatbot on your IR page will be as standard as having an investor presentation PDF. Luna: And the companies that get it right — that build a bot that's accurate, helpful, and compliant — will have a real edge in investor communication. Lucas: Without a doubt. And it's worth watching how the bigger players respond. If Apple or Microsoft start putting chatbots on their IR pages, that'll signal to the whole market that this is mainstream. Luna: Right now, Apple's IR page is pretty traditional — PDFs, contact form, phone number. But give it time. Lucas: Exactly. For now, the early adopters are the post ipo companies that need to punch above their weight in investor relations. A chatbot levels the playing field a bit. Luna: Good episode. I learned a lot. Lucas: Same here. Thanks, Luna.