Latest / Tech IPO Conversations with Fexingo: Public Software Companies and Capital Markets / How Post-IPO Companies Use AI to Detect Earnings Call Faux Pas
Transcript
- Lucas: So, Luna, we talk a lot about the mechanics of being a public company — lockups, buybacks, secondary offerings. But there's one moment every quarter where a CEO's words can move the stock more than any of that: the earnings call. Luna: Right, and we've seen it time and again — one evasive answer, one awkward pause, and the stock drops before the call even ends. Lucas: Exactly. And now there's a new layer: AI tools that analyze those calls in real time, looking for tone, hedging language, even micro-hesitations that might signal uncertainty. For post ipo companies, where the story is still being written, these tools are becoming a kind of early warning system. Luna: So are we talking about software that literally grades the CEO's performance? Because I've seen some of those demos, and they can get pretty granular. Lucas: Yeah, it's more than just 'did they sound confident.' These platforms use natural language processing to flag things like over-reliance on jargon, or when a CEO deflects a question about cash burn. And they compare it against the company's own historical calls, so they can spot when the tone shifts — even slightly. Luna: That's interesting because we've got real examples of companies under pressure right now. Amazon is down five and a half percent over the past week, Rivian down over six. If their CEOs are on a call talking about growth while the stock is sliding, an AI could probably pick up on the disconnect. Lucas: Absolutely. And that's the kind of thing that matters for post ipo companies especially. They don't have years of credibility to fall back on. One bad earnings call can reset the narrative for months. So some of them are actually using these AI tools internally before the call, to practice and refine their messaging. Luna: So it's a rehearsal tool too. Before they go live, they run the call through the AI and it tells them, 'hey, you dodged that question about gross margins three times.' Lucas: Exactly. And that's the kind of thing that used to be left to a few experienced communications folks. Now you've got a system that can process hundreds of past calls from peers, and say, 'when a CEO says this phrase, historically, the stock underperforms.' That's powerful, but it also raises some questions. Luna: Like whether you can train a machine to read a human's authenticity? Or whether you're just teaching CEOs to sound more like robots? Lucas: Right. There's a risk of over-optimizing for the algorithm. If the AI flags every moment of genuine deliberation as 'hesitation,' CEOs might start speaking in nothing but confident platitudes. And then the calls become even more sterile than they already are. Luna: But at the same time, I think there's something to be said for making sure you're not saying something you'll regret. Like, there was that headline about a company where the CEO made a joke about AI that didn't land — and it became a story for days. Lucas: Yeah, we've all seen those moments. And the AI can help flag that too — not just the words, but the sentiment behind them. Some tools analyze the transcript, but others go further and look at vocal tone, pace, even filler words like 'um' and 'you know.' Luna: So it's not just what you say, it's how you say it. That's a whole different beast. Lucas: Exactly. And for investors, these tools are becoming part of the due diligence process. Instead of just reading the transcript, some hedge funds are running sentiment analysis on the call audio itself. They're looking for signals that the text alone might miss. Luna: So we're getting to the point where the machines are listening to the machines, basically. Lucas: In a way, yes. And that's actually a good segue to something that's been on my mind. Look, we do this show because we love digging into these stories, and we keep it ad-free because we believe in independent analysis. But that only works if people who find value in it chip in a little. A couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo, if you've gotten something out of them. Luna: Yeah, it's a small thing that makes a big difference. We really appreciate anyone who supports the show. Lucas: So back to the AI angle — there's a real tension here between transparency and performance. On one hand, you want CEOs to be authentic, but on the other, you've got tools that can almost script the authenticity. Luna: What about the smaller post ipo companies? Are they actually adopting this, or is it just the big players with big budgets? Lucas: That's actually a good point, because some of these AI tools are surprisingly accessible. There are platforms that cost a few hundred dollars a month, which is nothing compared to the cost of a bad earnings call. So even smaller public companies can use them to prep their teams. Luna: So it's democratizing something that used to be reserved for the big banks' top-tier clients. Lucas: Exactly. And that's a shift we're seeing across the board — AI is leveling the playing field in investor relations. But it also means that the bar is rising. If everyone is using these tools, then the ones who don't stand out in a bad way. Luna: That's a good point. It's like a arms race of communication. The more polished everyone gets, the more investors might start to distrust the polish. Lucas: Right. And that's the paradox. The more you try to control the narrative, the more investors might wonder what you're hiding. I think the smartest companies will use these tools to catch genuine issues, not just to smooth over the edges. Luna: So what's the takeaway for a founder who's about to go public? Is this something they should be investing in now, or is it still early days? Lucas: I'd say it's early days, but not too early. If you're in the IPO process, it's worth at least understanding what these tools can do. Even if you don't use them, knowing that investors and analysts are using them on your calls changes the game. You need to be aware of how your tone is being interpreted. Luna: And that's a whole new layer of preparation. It's not just about having the right numbers anymore; it's about projecting the right confidence. Lucas: Exactly. And that's where I think the human element is still crucial. AI can flag patterns, but it can't tell you if a CEO's nervousness is genuine or if they're actually hiding something. That judgment call is still on us.