High Signal Podcasts Evidence ledger
Method
Browse
How the evidence works
← All source episodes

Lenny's Podcast / episode intelligence

The ultimate guide to AEO: How to get ChatGPT to recommend your product | Ethan Smith (Graphite)

14 Sept 2025 21 published claims 2 attributable people
Open original episode ↗ Read the evidence

Speakers in the public record

Ethan SmithLenny Rachitsky

Claim mix

belief 8evaluation 7recommendation 3prediction 2uncertainty 1

Evidence policy

Every row below preserves an exact excerpt. Identified speakers are linked; unresolved voices are labeled and excluded from people counts.

Claim ledger

The useful parts, with receipts.

21 published records

01 / recommendation

Recommends Emotional Intelligence. Recommends Cialdini's Persuasion book. Recommends Thinking, Fast and Slow. Recommends Persuasion. Recommends How to Measure Anything.

“Number one is Emotional Intelligence, and people talk about the concept of emotional intelligence, but there's actual research and psychology around that. I believe it was published in the '80s, but there's a really good book that summarizes the foundational research around emotional intelligence. And it's very useful when building relationships and communicating with people to understand their emotions. So that's the first one. And doing growth because growth is getting people to use your stuff. And so if you have frameworks to inform how people will use your things, then you can be a more effective growth person. Which brings me to my second book, which is Cialdini's Persuasion book. Robert Cialdini does a bunch of books around persuasion, but again, there's frameworks for how to persuade somebody to sign up, buy something. And so he breaks down his framework for that, and again, it's based on psychology. And I think especially in growth, there's all kinds of psychology research and behavioral economics research to inform tests. And if you just read Thinking, Fast and Slow, Persuasion, Emotional Intelligence, you can basically take those frameworks and apply it to growth in all kinds of different ways. And then the last is How to Measure Anything. So How to Measure Anything is about measuring things that are not immediately obvious to measure. They give this example of they wanted to measure how good an orchestra conductor was and they could survey or they could see the number of standing ovations for each orchestra conductor. And the more standing ovations probably means it's this better one and that you don't need to survey people. But much of growth and business is things that are not immediately obvious for how to measure, but anything could be measured, and so that's my third record.”
Speaker
Ethan Smith
Publisher
Lenny's Podcast

03 / uncertainty

Such a good point, because I don't know if I have competitors in this exact space, but basically they would show up instead and then I lose all that traffic.

“Such a good point, because I don't know if I have competitors in this exact space, but basically they would show up instead and then I lose all that traffic.”
Speaker
Lenny Rachitsky
Publisher
Lenny's Podcast

10 / belief

One interesting use case that I think will eventually come would be, I say, "Plan a trip to San Francisco," and decisions are made for you without any intervention.

“One interesting use case that I think will eventually come would be, I say, "Plan a trip to San Francisco," and decisions are made for you without any intervention.”
Speaker
Ethan Smith
Publisher
Lenny's Podcast

12 / evaluation

Especially for things like commerce and local and hotels, because they have these rich modules where you can click on stuff and go somewhere, which was not true before.

“Especially for things like commerce and local and hotels, because they have these rich modules where you can click on stuff and go somewhere, which was not true before.”
Speaker
Ethan Smith
Publisher
Lenny's Podcast

13 / evaluation

Probably it's because you're so primed because you're having a conversation with multiple follow-ups, and so there's so much intent that you've built.

“Probably it's because you're so primed because you're having a conversation with multiple follow-ups, and so there's so much intent that you've built.”
Speaker
Ethan Smith
Publisher
Lenny's Podcast

15 / evaluation

I didn't even think about this side of it when we started talking about this, but I think that's an important thing to note, is just this has nothing to do with the training data.

“I didn't even think about this side of it when we started talking about this, but I think that's an important thing to note, is just this has nothing to do with the training data.”
Speaker
Lenny Rachitsky
Publisher
Lenny's Podcast

16 / evaluation

Meaning the amount of questions that are very specific is larger, the share and the volume. And there's probably questions that have never been asked before and questions that have never been searched before, because search can't support lots of really specific, super-specific stuff.

“Meaning the amount of questions that are very specific is larger, the share and the volume. And there's probably questions that have never been asked before and questions that have never been searched before, because search can't support lots of really specific, super-specific stuff.”
Speaker
Ethan Smith
Publisher
Lenny's Podcast

18 / evaluation

Companies that we work with started in January and it started, one, because of more adoption, but two is because the answers became a bit more clickable.

“Companies that we work with started in January and it started, one, because of more adoption, but two is because the answers became a bit more clickable.”
Speaker
Ethan Smith
Publisher
Lenny's Podcast

19 / prediction

What's different about the tail is that the tail is larger in chat than in search. So the average number of words, I think, Perplexity said this to somebody else, said that it was around 25 words, where versus Google words it's around six words.

“What's different about the tail is that the tail is larger in chat than in search. So the average number of words, I think, Perplexity said this to somebody else, said that it was around 25 words, where versus Google words it's around six words.”
Speaker
Ethan Smith
Publisher
Lenny's Podcast
Search evidence