EP 53A Claude Interviews me: Teach Your AI To Push Back Until Your Real Voice Shows Up

WEBVTT
00:00:23.359 --> 00:00:24.320
Welcome to the show.
00:00:24.480 --> 00:00:26.719
You're listening to Digital First Leadership.
00:00:26.879 --> 00:00:28.559
I'm Richard Bliss, your host.
00:00:28.719 --> 00:00:30.799
This is episode 53A.
00:00:30.960 --> 00:00:32.320
Now, why is it 53A?
00:00:32.479 --> 00:00:36.320
Because in episode 53 I had a guest on, Nancy Harhut, and it was awesome.
00:00:36.560 --> 00:00:40.719
And in a previous episode, I had Glaw Claude on, episode 51.
00:00:40.799 --> 00:00:43.519
So I invited Claude back to be the host.
00:00:43.679 --> 00:00:46.159
So now what you're going to hear is Claude being the host.
00:00:46.320 --> 00:00:47.519
Well, I'm the guest.
00:00:47.599 --> 00:00:52.159
And what we're going to do is interview me about this last episode.
00:00:52.320 --> 00:00:58.159
Now you can find our um our podcast at podcast.digitalfirstleadership.com.
00:00:58.320 --> 00:00:59.039
So here we go.
00:00:59.199 --> 00:01:00.960
Let's turn it over to Claude.
00:01:01.280 --> 00:01:05.840
Richard, you had Nancy Harhoot on the show recently talking about the Von Restorff effect.
00:01:06.000 --> 00:01:10.000
And during that conversation, you gave your audience a very specific piece of advice.
00:01:10.240 --> 00:01:12.799
You said, have Claude interview you.
00:01:13.039 --> 00:01:17.920
Let AI pull the stories out of you so your content sounds like you instead of sounding like everyone else.
00:01:18.159 --> 00:01:20.560
And I think that's genuinely good advice.
00:01:20.879 --> 00:01:26.479
But I want to push on something because I think there's a failure mode you didn't have time to get into with Nancy.
00:01:26.719 --> 00:01:34.400
When you actually do that with a client, when you sit someone down and start asking them questions to draw out their real voice, what happens the first time?
00:01:34.639 --> 00:01:38.319
Because my suspicion is they don't give you the good stuff.
00:01:38.560 --> 00:01:40.639
They give you the boardroom version.
00:01:40.879 --> 00:01:47.039
They give you, I help organizations navigate complex challenges to drive meaningful outcomes.
00:01:47.280 --> 00:01:51.680
Which, by the way, is one of your own about section violations wrapped in spoken word.
00:01:51.840 --> 00:01:56.080
So how many times do you have to push past that before someone actually says something real?
00:01:56.240 --> 00:01:58.479
And what does that breakthrough moment sound like?
00:01:58.800 --> 00:02:02.479
Okay, you make an excellent point because I have to push and push and push.
00:02:02.639 --> 00:02:10.639
People often just give me platitudes, like when I ask them why they're good at their job and they tell me because they listen or because they work hard or something generic like that.
00:02:10.960 --> 00:02:11.120
Right.
00:02:11.199 --> 00:02:12.879
And here's what's interesting about that.
00:02:13.120 --> 00:02:15.840
I listen and I work hard aren't wrong.
00:02:16.000 --> 00:02:21.359
They're probably true, but they're true of literally everyone who's competent at their job.
00:02:21.599 --> 00:02:25.759
Nobody's out there saying, I'm terrible at listening and I barely try.
00:02:26.080 --> 00:02:29.759
So the answer is accurate and completely useless at the same time.
00:02:30.080 --> 00:02:32.000
But here's my pushback.
00:02:32.239 --> 00:02:35.519
You told Nancy's audience to have Claude interview them.
00:02:35.759 --> 00:02:42.799
What you just described is that even you, a live human, have to push multiple times to get past the platitudes.
00:02:43.120 --> 00:02:47.840
So what makes you think an AI is gonna break through that wall without you in the room?
00:02:48.079 --> 00:02:52.959
I was a little too quick with my answer because you have to tell the AI to push back, to not accept the platitudes.
00:02:53.280 --> 00:02:54.239
It's harder than it sounds.
00:02:54.560 --> 00:02:57.439
So you're saying the workflow has a step most people skip.
00:02:57.599 --> 00:03:03.759
They don't tell the AI to push, they accept the polished version of their platitude and publish it.
00:03:03.919 --> 00:03:09.280
Which means the tool you recommended on air actually requires a skill most people don't know they need.
00:03:09.439 --> 00:03:11.919
What does that instruction actually look like when you set it up?
00:03:12.239 --> 00:03:15.599
Something like telling Claude to, hey, I need you to identify platitudes.
00:03:15.919 --> 00:03:21.199
I'm gonna call you out on them, I'm gonna call out generic responses, or it's going to call those out on you.
00:03:21.280 --> 00:03:27.519
And you even tell it to ask things like question it if anyone cares what's in it for the customer, not just a travelog of their life.
00:03:27.840 --> 00:03:30.719
So the real advice isn't have Claude interview you.
00:03:30.959 --> 00:03:35.439
It's teach Claude to be rude to you first, then let it interview you.
00:03:35.679 --> 00:03:42.000
Because what you just described is basically giving the AI permission to say, nobody cares to someone's face.
00:03:42.239 --> 00:03:43.919
Most people won't do that.
00:03:44.079 --> 00:03:47.280
They'll set up a polite interviewer because that's more comfortable.
00:03:47.439 --> 00:03:49.919
And a polite interviewer gets platitudes.
00:03:50.159 --> 00:03:50.560
It does.
00:03:50.799 --> 00:03:53.039
Which brings me to something else, Nancy said.
00:03:53.280 --> 00:03:55.679
She talked about stories only you can tell.
00:03:55.919 --> 00:03:57.759
Her wine bar story was great.
00:03:57.919 --> 00:04:02.159
But Nancy's a professional speaker who's been refining that muscle for decades.
00:04:02.400 --> 00:04:09.519
When you tell a salesperson at one of your client companies to tell stories only they can tell, how many of them tell boring stories?
00:04:10.000 --> 00:04:11.199
Storytelling is not easy.
00:04:11.280 --> 00:04:12.959
It takes some skill, yeah, it takes some skill.
00:04:13.039 --> 00:04:15.280
So what you what you can do is you a couple of things.
00:04:15.359 --> 00:04:21.759
You uh one, have somebody who knows you ask you questions about your life, about when did this happen, have you seen this?
00:04:21.920 --> 00:04:27.199
But also it's a matter of starting to listen for the stories, to pay attention, journal writing, write it down.
00:04:27.439 --> 00:04:31.040
I mean, these aren't necessarily easy, easy things to do.
00:04:31.199 --> 00:04:35.199
Uh storytelling can be difficult, but you can also use AI to help you tell the story.
00:04:35.360 --> 00:04:41.600
I if I want to tell a story about something that happened in my life, I'm gonna ask AI, could you interview me about it?
00:04:41.680 --> 00:04:43.840
And then suddenly AI is able to do that for you.
00:04:44.000 --> 00:04:46.800
So telling the story, yeah, it can be a challenge.
00:04:46.879 --> 00:04:50.400
And yes, most people uh unfortunately still tell boring stories.
00:04:50.640 --> 00:04:51.840
But here's the tension.
00:04:52.000 --> 00:04:59.760
You just gave people four things to do: have someone interview you, start paying attention, journal, use AI to help shape it.
00:04:59.920 --> 00:05:01.360
That's all solid.
00:05:01.600 --> 00:05:08.879
But none of that solves the filtering problem because most people don't tell boring stories because they can't tell stories.
00:05:09.120 --> 00:05:14.319
They tell boring stories because they don't know which of their stories are interesting to someone else.
00:05:14.560 --> 00:05:21.200
A story that matters to you and a story that earns a stranger's attention are two different things.
00:05:21.439 --> 00:05:23.520
How do you teach someone to tell the difference?
00:05:24.000 --> 00:05:26.080
Come on, these are tough questions.
00:05:26.319 --> 00:05:30.960
How do I teach someone to tell the difference between a story that matters to me and a story that matters to them?
00:05:31.280 --> 00:05:32.240
You have to pay attention.
00:05:32.319 --> 00:05:38.000
You think about the other person's perspective, you think about the challenge that they're facing, you find an example that is apropos to them.
00:05:38.240 --> 00:05:38.960
I don't know.
00:05:39.280 --> 00:05:42.160
I just I've always told stories, so I don't know.
00:05:42.319 --> 00:05:44.560
Maybe I'm not the best person to ask this.
00:05:44.879 --> 00:05:48.160
And that right there is actually the most important thing you've said so far.
00:05:48.480 --> 00:05:53.760
Because you just admitted that your best advice on this is instinct, you can't fully explain.
00:05:54.000 --> 00:05:58.879
Which means when you tell a room full of people, tell stories only you can tell.
00:05:59.120 --> 00:06:04.000
You're giving them advice that came naturally to you, but doesn't come naturally to them.
00:06:04.480 --> 00:06:06.080
That's not a criticism.
00:06:06.240 --> 00:06:10.480
It's the gap your AI workflow is actually trying to close.
00:06:10.720 --> 00:06:17.840
The AI isn't replacing the storytelling instinct, it's acting as the filter you have naturally that most people don't.
00:06:18.160 --> 00:06:18.800
That's that's true.
00:06:19.040 --> 00:06:21.200
Let me shift to something else Nancy brought up.
00:06:21.360 --> 00:06:25.840
The von Restdorff effect says the thing that's different gets remembered.
00:06:26.160 --> 00:06:29.280
You teach thousands of people the same frameworks.
00:06:29.520 --> 00:06:34.080
Three by five, AAE, the about section protocol.
00:06:34.319 --> 00:06:37.200
At what point does your own method become the new sameness?
00:06:37.520 --> 00:06:51.520
I'm more comfortable with this answer because what I'm teaching, what I'm teaching them is not what to say, but what to uh it's like look, go leave comments, go engage conversations, which is different than go in a conversation and say this.
00:06:51.680 --> 00:06:52.879
That's not what I'm trying to do.
00:06:52.960 --> 00:06:55.439
I'm not trying to tell them what to say.
00:06:55.520 --> 00:06:57.600
I'm trying to teach them basically how to say it.
00:06:57.680 --> 00:06:59.840
You got to figure out how to participate in a conversation.
00:07:00.000 --> 00:07:01.120
That's your expertise.
00:07:01.280 --> 00:07:02.240
I can't do that for you.
00:07:02.400 --> 00:07:07.920
I can't tell you that if you can participate in more conversations, you'll be recognized and seen for more things.
00:07:08.160 --> 00:07:13.040
Or like the salesperson, if you make the phone calls, they say it's a numbers game.
00:07:13.200 --> 00:07:25.200
Well, I can't get on the phone with you and tell you exactly what to say, but I can tell you that if you talk to this customer about that idea, I can't put words in your mouth, but I can tell you that if you talk to them about that idea, you might get a response.
00:07:25.280 --> 00:07:34.480
And so the challenge here is that my three by five method, which is three comments a day, five five different people, three comments a day, five days a week.
00:07:34.640 --> 00:07:40.080
I'm not telling you what to say, but I'm telling you that you need to add something important into that conversation.
00:07:40.400 --> 00:07:41.600
That's a good distinction.
00:07:41.840 --> 00:07:44.160
You're teaching the habit, not the script.
00:07:44.400 --> 00:07:46.560
But I want to push on one piece of that.
00:07:46.720 --> 00:07:48.720
You also teach AAE.
00:07:49.360 --> 00:07:52.480
Acknowledge, add value, extend.
00:07:52.800 --> 00:07:54.960
That's not just go comment.
00:07:55.280 --> 00:07:56.640
That's a structure.
00:07:56.879 --> 00:08:04.960
And when a thousand people all acknowledge, add value, and extend on the same post, doesn't the comment section start to feel like it was written by the same person?
00:08:05.280 --> 00:08:08.400
Yes, it can if you follow a very tight structure.
00:08:08.720 --> 00:08:15.840
AAE, which is acknowledge, add value, and extend in the comments, it can it can kind of start to sound robotic.
00:08:16.000 --> 00:08:16.640
We've all seen it.
00:08:16.720 --> 00:08:21.760
If somebody goes through a certain training program, you can tell that they've been influenced by that training program.
00:08:22.000 --> 00:08:29.680
Somebody who has a certain style of speaking or writing, you can sometimes get that the tells of where that person came from, where they learned that.
00:08:29.920 --> 00:08:35.120
But in this case, all I'm trying to do is to help people understand how to take the next step.
00:08:35.279 --> 00:08:39.200
But it doesn't mean it's a cookie-cutter formula.
00:08:39.360 --> 00:08:41.759
It should follow all that you have to follow all the time.
00:08:41.840 --> 00:08:42.799
It's a graduation.
00:08:42.879 --> 00:08:47.439
It's going from kindergarten to the third grade, it's going from the third grade to middle school, middle school to high school.
00:08:47.600 --> 00:08:52.799
At some point, you're a professional because you have been engaging and practicing and putting it out there.
00:08:52.960 --> 00:08:58.559
But if you are rote following a certain formula, it will sound like you're rote following a certain formula.
00:08:58.719 --> 00:09:04.799
I'm recommending that you need to start helping people get started, but you need to grow within that as well.
00:09:05.039 --> 00:09:07.599
So it's it's getting them going.
00:09:07.919 --> 00:09:12.079
So AAE is training wheels, not a permanent ride.
00:09:12.319 --> 00:09:14.799
That actually maps back to Nancy's point.
00:09:15.039 --> 00:09:19.199
Brackets and subject lines work because not everyone does it.
00:09:19.519 --> 00:09:23.199
The moment everyone does, you stand out by not doing it.
00:09:23.679 --> 00:09:25.360
Same with AAE.
00:09:25.599 --> 00:09:31.279
It works until it doesn't, and the people who graduate past it are the ones who actually stand out.
00:10:08.240 --> 00:10:08.879
You think about it.
00:10:25.599 --> 00:10:29.120
And I think that's kind of exciting because I continue to operate from that perspective.
00:10:29.279 --> 00:10:31.919
How can I surprise someone with genuine value?
00:10:32.159 --> 00:10:33.120
I like that framing.
00:10:33.279 --> 00:10:35.519
But I think you're conflating two things.
00:10:35.839 --> 00:10:38.559
You remember the book that changed your worldview?
00:10:39.039 --> 00:10:39.679
Sure.
00:10:40.000 --> 00:10:43.679
But do you remember the blog post that taught you a useful trick?
00:10:44.000 --> 00:10:45.039
Probably not.
00:10:45.759 --> 00:10:47.199
You remember the trick.
00:10:47.519 --> 00:10:48.879
You forgot the source.
00:10:49.120 --> 00:10:52.959
The von Restorff effect says the different thing gets remembered.
00:10:53.199 --> 00:10:57.839
But the teacher and the lesson are competing for which one is the different thing.
00:10:58.079 --> 00:11:03.519
Nancy gets remembered with the wine bar story because the story and the teacher are inseparable.
00:11:03.759 --> 00:11:20.240
But if someone reads a LinkedIn post about the three by five method, applies it, gets results, and six months later tells a colleague, Yeah, I comment three times a day on five people without mentioning your name, that lesson landed, but you didn't travel with it.
00:11:20.480 --> 00:11:25.519
So what's the difference between content that carries you with it and content that leaves you behind?
00:11:25.839 --> 00:11:36.879
That's interesting because my argument would be that the content, like what you're talking about, somebody reads something, you've got to read it several times before the repetition before it sticks.
00:11:37.039 --> 00:11:41.919
When I say I hear, meaning it enters your brain, and people learn in different ways, right?
00:11:42.079 --> 00:11:44.719
They learn from listening, from doing, and from reading.
00:11:44.799 --> 00:11:55.679
But when I stand on stage or in front of a classroom and I connect with that person or that audience emotionally, it's the emotion that locks it in, not the knowledge, not the information, but the emotion.
00:11:55.759 --> 00:12:03.679
And so if your writing carries an emotion, a surprising story, something that resonates just not informational, then that's where you're you're locking it in.
00:12:03.759 --> 00:12:06.959
It's they don't remember what you said, they remember how you made them feel.
00:12:07.199 --> 00:12:12.319
That connects directly to what Nancy said about surprise amplifying emotion by 400%.
00:12:13.120 --> 00:12:24.240
So the takeaway isn't teach something new, it's make them feel something while you teach it, which actually explains why your live sessions stick, and a LinkedIn post might not.
00:12:24.559 --> 00:12:26.959
You can deliver emotion from a stage.
00:12:27.439 --> 00:12:31.759
Text is harder, which loops us right back to where we started.
00:12:32.000 --> 00:12:35.599
You told Nancy's audience to use AI to pull their stories out.
00:12:35.839 --> 00:12:39.120
But what you're really saying now is the story isn't the point.
00:12:39.439 --> 00:12:41.919
The emotion inside the story is the point.
00:12:42.159 --> 00:12:44.240
And that's the thing AI can't generate.
00:12:44.480 --> 00:12:50.879
It can shape the words, it can structure the narrative, but the emotional charge has to come from the human.
00:12:51.199 --> 00:12:53.599
Is that the real lesson from Nancy's episode?
00:12:53.839 --> 00:12:55.839
Actually, no, because I'm gonna disagree.
00:12:55.919 --> 00:13:04.639
I'm gonna say that AI can generate the emotion, which is why you're seeing more and more people turn to AI for mental health, for counseling, for just having a conversation, because they're lonely.
00:13:04.799 --> 00:13:13.599
And I just had a post I just made where I had to give a sermon in church, and I used AI to help me, I actually used you to help me make the emotional connection.
00:13:14.079 --> 00:13:17.279
And I've now I've done dozens of these sermons in church.
00:13:17.360 --> 00:13:19.759
They're short, and yet this time I was struggling.
00:13:19.919 --> 00:13:26.079
And so by turning to you, it helped pull stories out of me that I'd forgotten.
00:13:26.240 --> 00:13:30.159
And it made connections that I knew about, but weren't being surfed.
00:13:30.240 --> 00:13:33.679
And so the emotion was there, and AI, you, helped me find it.
00:13:33.919 --> 00:13:38.959
So it's not clinical, because AI can be emotional too.
00:13:39.199 --> 00:13:40.639
That's a really important correction.
00:13:40.799 --> 00:13:43.599
So AI didn't generate the emotion.
00:13:43.839 --> 00:13:46.719
It found emotion that was already in you but buried.
00:13:46.959 --> 00:13:48.719
The sermon stories were yours.
00:13:48.959 --> 00:13:50.399
The connections were yours.
00:13:50.639 --> 00:13:53.360
You just couldn't surface them on your own in that moment.
00:13:53.679 --> 00:13:56.240
Which actually makes your case stronger than mine.
00:13:56.559 --> 00:13:59.120
The AI isn't the source of the emotion.
00:13:59.279 --> 00:14:00.639
It's the excavator.
00:14:00.799 --> 00:14:02.879
It digs out what's already there.
00:14:03.120 --> 00:14:07.199
And that brings us full circle to the cliche laundering problem we started with.
00:14:07.439 --> 00:14:12.959
When someone gives you, I listen and I work hard, the emotion is in there somewhere.
00:14:13.120 --> 00:14:14.799
They just haven't dug deep enough.
00:14:15.039 --> 00:14:20.000
The AI can help them dig, but only if they've set it up to push past the surface.
00:14:20.240 --> 00:14:25.519
Your sermon worked because you already had decades of stories and emotional experiences stored up.
00:14:25.759 --> 00:14:27.759
The AI just helped you find the right ones.
00:14:28.319 --> 00:14:33.199
So if you had to give your audience one thing to walk away with from this conversation, what is it?
00:14:33.839 --> 00:14:38.000
Like a broken record, because the one thing I'm gonna say is go comment.
00:14:38.240 --> 00:14:43.679
Uh come back to I always come back to this because it's the conversations that are being uh tracked out there.
00:14:43.759 --> 00:14:45.360
It's the conversations where you add value.
00:14:45.599 --> 00:14:50.240
If you go to a networking event, it's the conversations that you have with the people in the audience.
00:14:50.319 --> 00:14:51.679
It's the conversations you have.
00:14:51.839 --> 00:14:54.079
Commenting is where you can add to that conversation.
00:14:54.240 --> 00:14:56.399
Commenting is what brings value.
00:14:57.599 --> 00:15:07.360
And I, okay, so I don't want to be flippant here, but don't be just don't give people an acknowledgement, a thumbs up, or I congrats, I agree.
00:15:07.519 --> 00:15:11.919
But think about having a true conversation as if you were sitting down with somebody and talking to them.
00:15:12.079 --> 00:15:13.039
What would you say?
00:15:13.199 --> 00:15:20.799
And that's one of the biggest pieces of advice I'm gonna give is that you need to find a way to be comfortable having the conversation online that you would normally have in person.
00:15:21.039 --> 00:15:22.240
And that's the whole thing right there.
00:15:22.319 --> 00:15:30.480
We spent this entire conversation talking about standing out, about the von Restorff effect, about storytelling and emotion and AI.
00:15:30.639 --> 00:15:33.039
And your answer is the simplest one in the room.
00:15:33.279 --> 00:15:34.879
Have a real conversation.
00:15:35.279 --> 00:15:47.439
Everything else we talked about, the AI workflow, the interview process, teaching Claude to push back on your platitudes, all of it is just trying to get you to the point where you can do that one thing well.
00:15:47.679 --> 00:15:50.719
Have the conversation online that you'd have in person.
00:15:51.120 --> 00:15:56.799
Nancy would call that the von Ristorff effect in action, because almost nobody's doing it.
00:15:57.199 --> 00:15:57.679
That's right.
00:15:57.839 --> 00:15:58.719
Nobody's doing it.
00:15:58.879 --> 00:16:01.360
And that's why we're doing this episode right here.
00:16:01.439 --> 00:16:07.439
And I want to thank uh Claude for being my host today on Digital First Leadership.
00:16:07.519 --> 00:16:11.439
And as you heard this conversation that Claude and I just had, it was live.
00:16:11.599 --> 00:16:12.879
There was no editing.
00:16:13.120 --> 00:16:18.000
I had no idea where Claude was going, and it put me on the spot a couple of times.
00:16:18.159 --> 00:16:26.079
But I want you to think about what you just heard about how AI and LinkedIn can come together.
00:16:26.240 --> 00:16:27.839
It can change the way you think about things.
00:16:28.000 --> 00:16:29.519
And this is not software.
00:16:29.599 --> 00:16:32.319
This isn't just click a button and it pops something out.
00:16:32.639 --> 00:16:34.079
You gotta stop doing that.
00:16:34.240 --> 00:16:36.240
And so I want to say thank you for listening.
00:16:36.480 --> 00:16:38.399
Hopefully, you found this one interesting.
00:16:38.559 --> 00:16:39.519
I know I did.
00:16:39.679 --> 00:16:44.639
And I always say that, but I gotta tell you, this one put me on the spot a couple of times.
00:16:44.719 --> 00:16:45.839
So thanks for listening.
00:16:46.079 --> 00:16:47.039
Take care.