Latest / Tech IPO Conversations with Fexingo: Public Software Companies and Capital Markets / How Post-IPO Companies Use AI for Earnings Call Prep
Transcript
- Lucas: Alright, so let's talk about something that a lot of newly public companies are wrestling with right now — earnings call preparation, and specifically how artificial intelligence is starting to change that process. Luna: I imagine the stakes are pretty high. You've got analysts, institutional investors, the media all listening for any slip-up. Lucas: Exactly. And the traditional approach is a sort of war room — the CFO, the IR team, the CEO, sometimes outside advisors — spending days crafting answers, doing mock Q&A sessions. It's intense. Luna: So where does AI come in? I've heard of some companies using natural language processing to simulate analyst questions. Lucas: Yeah, that's exactly it. There are now tools that ingest past transcripts, analyst reports, even recent news — and then generate a simulated Q&A session. The IR team can practice against a model that's trained to mimic the toughest analysts on the call. Luna: Right, so it's like a sparring partner. And I assume the quality of the simulation depends on how much data you feed it. Lucas: That's the key. The more historical data — your own past calls, your competitors' calls, the buy-side questions you've fielded — the better the AI gets at predicting the actual curveballs. Some companies are now feeding the AI real-time sentiment signals from social media or news flow leading up to the call. Luna: Let's bring this to a specific example. I'm looking at Apple — stock's up 5 percent this week, trading around 333. That's a pretty strong run. If I'm the IR team at Apple, what's on my mind going into the next call? Lucas: Great question. Apple's got a services revenue narrative that's been driving margins, but hardware — especially iPhone — still makes up the bulk of revenue. Analysts will want to know about AI integration, especially with the recent headlines about OpenAI and hardware. The AI tool would flag that as a high-probability topic and generate multiple angles. Luna: So you can prep a more thorough answer. And the AI can also catch inconsistencies — like if your prepared remarks on capex don't align with what you said last quarter. Lucas: Exactly. Some tools actually do a consistency check across all public statements. If you said you'd keep R&D spend at 8 percent of revenue, and your script says 9 percent, the AI flags it. That's huge for avoiding embarrassing corrections. Luna: What about the smaller companies? The ones that just IPO'd and maybe don't have a seasoned IR team? Lucas: That's actually where I see the biggest impact. A company that went public two years ago, with a lean finance team — they can get ai generated talking points, sentiment analysis, even a draft of the CEO's opening remarks. It levels the playing field. Luna: But there's a risk, right? If you rely too much on the AI, you might sound robotic. Or worse, the AI might hallucinate a fact. Lucas: That's the number one concern. Hallucination in a live earnings call would be catastrophic. So the best practice is to use AI for preparation, not for script generation that goes straight to the teleprompter. The human still has to own every word. Luna: And the SEC is watching. They've been pretty clear that you can't mislead investors, even if the AI wrote the script. Lucas: Right. The legal liability still sits with the company. So the AI is really a tool for efficiency — it helps you get to the right answer faster, but it doesn't replace judgment. Luna: Let's talk about Nvidia. Stock around 203 this week, down slightly. Their earnings calls are hugely watched because they're the bellwether for AI capex. Lucas: Nvidia's a fascinating case. Their investor relations team has been known to field very technical questions — about GPU architecture, data center buildout, energy efficiency. A good AI tool would need to be trained on very specific technical documents, not just financial data. Luna: So we're moving into domain-specific AI models for earnings prep. Lucas: Exactly. I know of at least two startups that are building fine-tuned models for semiconductor companies and biotech firms. The language is so specialized that a general LLM won't cut it. Luna: And the cost? For a post ipo company with limited budget, is this something that's accessible? Lucas: It's actually not that expensive. You can get a subscription to a good IR AI tool for a few thousand dollars a quarter — which is a rounding error compared to the cost of a bad earnings call that tanks your stock. Luna: If today has been useful for understanding how AI is reshaping earnings prep — and I think it has — then it's worth mentioning that this show stays ad-free because of listener support. If you find value in what Lucas and I dig into, you can help keep it going at buy me a coffee dot com slash fexingo. Lucas: Yeah, we keep the production lean and the content focused. That kind of direct listener support is what allows us to do deep dives like this without any sponsor noise. Luna: So back to the tools — I'm curious about the real-time aspect. Some of these platforms can actually listen to the call and suggest answers live, right? Lucas: They can, but most companies shy away from that. The risk of the AI recommending something off-script mid-call is too high. Instead, they use it in prep mode to build a robust FAQ, and then go live with a human team that's fully briefed. Luna: That makes sense. What about after the call? AI for post-call analysis? Lucas: Big area. Instant transcript analysis, sentiment scoring of questions versus answers, identifying which topics drove the most volatility in after-hours trading. Some hedge funds are even using AI to parse the tone of the CEO's voice. Luna: So it's a two-edged sword. Companies use AI to prepare, and investors use AI to parse. Lucas: Exactly. It's an arms race. And the companies that are most effective at using AI for IR are the ones that understand it's not just a technology play — it's a communication strategy. Luna: Alright, so what's the one takeaway for a founder who just took their company public and is thinking about their first few earnings calls? Lucas: I'd say: invest in the prep tool early, but don't let it write your narrative. Use it to stress-test your messaging, to anticipate the tough questions, and to make sure your team is aligned. The human voice still matters more than ever. Luna: Good advice. And maybe a reminder that analysts can tell when you're reading a script. Lucas: They absolutely can. Authenticity — even a little bit of vulnerability — actually builds trust with the Street. The AI can't fake that. Luna: Thanks for listening to Tech IPO Conversations. We'll be back next week with another angle on how companies navigate life after the IPO.