Latest / Sales Leadership with Fexingo: Quota Carriers, Sales Managers, and Revenue Teams / How Top Reps Use AI Without Sounding Like a Robot
Transcript
- Lucas: There's this moment in almost every sales organization right now—someone on the team feeds a prospect's LinkedIn profile into ChatGPT, gets back a perfectly formatted email, and sends it without changing a word. And the prospect writes back saying, 'Nice form letter.' That's the problem we're drilling into today. Luna: I've seen that exact thing. A rep at a friend's company got a reply that just said, 'Did you actually read my profile?' Brutal. Lucas: Right. So here's the uncomfortable truth: AI can make you faster, but it can also make you sound like everyone else. The best reps we've studied have figured out how to use AI without losing their own voice. And there's a specific case that illustrates this perfectly. Luna: Let me guess—the one from last year's Sales Hacker conference? A company called Enablix? Lucas: Close. It's a B2B SaaS firm, not public, but they shared their numbers. They rolled out an AI coaching platform—think Gong but with a real-time feedback layer—across a team of forty reps. After six months, close rates went up eighteen percent. But here's the catch: the improvement only showed up in the reps who used the AI to prepare, not to execute. Luna: I want to unpack that. How did they measure 'preparation versus execution'? Lucas: They tracked two behaviors. First, reps who used AI to generate call scripts and then read them verbatim. Second, reps who used AI to summarize the prospect's company news, product updates, and competitive landscape—and then went into the call without a script. The second group closed at a rate that was twenty-two percent higher than the first. Luna: So the AI was better as a research assistant than a co-pilot. That makes intuitive sense—prospects can smell a script from a mile away. Lucas: Exactly. And this is where we get to the first of three rules that top reps are using. Rule one: use AI for preparation, not conversation. The AI should help you know what the prospect cares about, not tell you what to say. One rep I interviewed—she's at Salesforce, top ten percent of her pod—she feeds the AI the prospect's last two earnings calls and asks for three strategic priorities. That's it. She writes the outreach herself. Luna: That's smart because the AI can synthesize a ten-page transcript in seconds. A human would take forty minutes. But the actual phrasing—the empathy, the timing, the humor—that's still on you. Lucas: Right. Rule two: let AI handle the data, but keep intuition for the read. A lot of tools now offer sentiment analysis during calls. They'll flag if a prospect sounds frustrated or confused. But the machine doesn't know why. A top rep I watched recently—he got a flag that the prospect's tone turned negative when he mentioned integration timeline. The AI just said 'negative sentiment detected.' The rep knew, because the prospect had mentioned a failed implementation two years ago, that the fear was about disruption. So he pivoted to talk about their onboarding success rate. The AI gave him the signal, but his memory and empathy decoded it. Luna: That's a great example. The AI is like a dashboard light—it tells you something's wrong, but you still have to pop the hood. And I love that he used a personal detail, not a generic reassurance. Lucas: Rule three is the one that gets the most pushback from sales managers: always edit ai generated content to sound like you. I know this sounds obvious, but the data from that same company showed that reps who sent ai generated emails with zero edits had a reply rate of four percent. Reps who spent two minutes personalizing the first paragraph—changing a reference, adding a specific observation—had a reply rate of eleven percent. Luna: Nearly triple. And it's just two minutes. So why don't more reps do it? Lucas: Speed pressure. Quota pressure. They think the AI output is good enough. And sometimes it is—for a follow-up email that's purely logistical. But for the first outreach or the discovery call follow-up, generic is a liability. The rep I mentioned from Salesforce—she has a rule she calls 'three swaps.' She takes any AI draft and changes three things: the greeting, the reason for reaching out, and the call to action. She keeps the structure, but the language is hers. Luna: I want to complicate this a little. What about the argument that AI can actually help you find your voice—by giving you a starting point that you then refine? Some reps say they write better after they see what the AI produces, because it sparks an idea they wouldn't have had otherwise. Lucas: That's a fair point. I've heard that too. And I think it depends on the rep's experience level. A junior rep who's still learning the product might genuinely benefit from seeing a well-structured email. But the danger is that they never move past imitation. The best reps I've talked to use AI as a brainstorming partner, not a ghostwriter. They'll prompt it with 'give me three angles on how cloud migration could reduce IT overhead for a manufacturing company' and then pick the one that resonates. They don't copy the language. Luna: So it's about intentionality. Using the tool, not being used by it. There's a great line from a sales trainer I follow—he says, 'AI can help you scale your best self, but only if you know what your best self sounds like.' Lucas: That's exactly it. And that's why this episode is really about self-awareness more than technology. The AI tools are changing fast. By this time next year, the models will be better at mimicking human tone. But the core skill—knowing what's authentic for you and your prospect—that's not going to change. Luna: Before we wrap, I want to touch on one more thing: the risk of over-relying on AI for call coaching. We've seen platforms that score every word a rep says and suggest alternative phrases in real time. Does that help or hurt? Lucas: I think it can hurt if the rep becomes dependent on it. There's research from MIT—actually, from a study on call center agents using AI suggestions—that found agents who followed the AI's recommendations too closely had lower customer satisfaction scores. The AI optimized for efficiency, but customers wanted to feel heard. The best agents used the AI as a safety net, not a script. Luna: That's the same pattern. Efficiency without empathy is just automation. And in sales, automation doesn't close deals—people do. Lucas: Yeah. And that's a good note to end on. But you know, this is exactly the kind of conversation that we have on this show every week—real, practical, no filler. We deliberately don't run ads on these episodes because we want the focus to stay on what works, not on selling you something. Luna: And we've heard from a lot of you that that matters. One listener said it's the only sales podcast she can listen to without feeling like she's being pitched to. Lucas: If these conversations have moved your work forward in some small way, and you want to support that, the link is buy me a coffee dot com slash fexingo. It's not required—we'll keep doing this either way—but it helps us keep the ad-free model going. Luna: Totally. And now back to the practical stuff. Lucas, you mentioned a framework earlier—the 'three swaps' rule. Can you walk us through exactly how that works in a real email? Lucas: Sure. So let's say the AI drafts something like: 'Hi, I noticed your company recently launched a new product line. At, we help businesses like yours streamline their supply chain. Would you have 15 minutes to chat next week?' That's not terrible, but it's forgettable. The rep would swap the greeting to something specific—like 'Hi Sarah, I saw your interview about the product launch on LinkedIn.' Then swap the reason: 'Your focus on reducing lead times aligns with what we do for clients like Acme Corp, who cut their lead time by 30%.' Then swap the CTA: 'Would you be open to a 10-minute call Thursday at 2pm to see if there's a fit?' Now it's personalized, credible, and direct. Luna: And it takes two minutes. I can see why that works. The structure stays efficient, but the words become human. Lucas: Exactly. And that's the whole point. AI can handle the heavy lifting of research and structure. But the human part—the empathy, the specific observation, the genuine curiosity—that's what earns a reply. Luna: So if you're a sales manager listening, the takeaway might be: don't ban AI, but do coach your reps on how to edit it. Run a workshop where you compare an ai only email with one that's been through 'three swaps.' Lucas: Great idea. And for the reps: next time you're tempted to hit send on an AI draft, pause. Read it out loud. If it sounds like it could be from anyone, it's not ready. Make it sound like you. Luna: Alright, that's a solid place to land. Thanks, Lucas. Lucas: Thanks, Luna. And thanks to everyone listening. We'll be back next week with another episode.