You know your buyer is busy, and you know they have objections. You don’t know which ones. The customer interviews that would tell you take weeks to book, and the survey you could send tomorrow only answers the questions you already thought to ask.
Luan Wise, a chartered marketer, Fellow of the Chartered Institute of Marketing and author of eight books on marketing and social media, runs those interviews in Claude instead. In this session, she builds Sarah, the owner of a 15-person UK business, from research she already had. She interviews Sarah about cybersecurity, then switches Claude into the seat of the provider’s marketing lead to work out what Sarah’s answers mean for the campaign.
You come away knowing things before you write a single headline. Sarah rates her own protection a four out of ten. She has never asked her IT provider a direct question about it, and she’s nervous about asking. Luan is clear that none of this replaces your real customers. It gives you sharper questions to take to them.
What you will learn
- Claude builds a more useful buyer persona when you feed it the customer research you already have than when you ask it to invent a buyer from nothing.
- A seed prompt that says “stay in character, don’t give me generic marketing advice, and if something is uncertain, say so” keeps the persona answering as a buyer instead of as a consultant.
- Switching Claude from the buyer to “the marketing lead for the cybersecurity provider” inside the same chat turns the interview into a debrief without opening a new conversation.
- Asking the persona to rate her confidence from one to ten, and asking Claude to give a percentage for its own confidence, exposes uncertainty that a polite answer would hide.
- In Luan’s demo, the buyer’s biggest barrier was her relationship with her existing IT provider: she didn’t know which questions to ask, and she didn’t want the provider to feel accused.
- Luan runs the whole exercise on the free version of Claude.
Timestamp index
00:00 The buyer who assumed her IT provider had it covered
00:54 Why Luan swapped surveys and focus groups for AI role-play
02:12 Plan your scenarios before you type a prompt
03:16 The seed prompt: build Sarah from research you already have
04:32 Sarah’s first answer puts cybersecurity on the list, not at the top
05:19 The follow-up question a survey could never ask
07:17 Sarah rates her protection a four out of ten
08:30 Switch seats and let Claude debrief as the marketing lead
10:12 Why the switch works like analysing a call transcript
11:00 Do you need to run the role-play more than once?
11:34 Would Sarah actually ask her IT provider?
12:19 “I’ve been asking the wrong question all along”
13:25 What Sarah’s answers mean for your content and your jargon
15:23 Turn the avatar into a one-pager, a landing page or an email test
16:47 How to stop the persona agreeing with you
18:17 Put the persona in a skeptical buyer’s seat, then the Finance Director’s
19:05 The biggest mistake: starting from a blank page
20:02 Geoff Woods and the AI board of directors
21:14 What role-play changes in Luan’s client work
22:02 Where to find Luan
Start with what your customers already told you
Luan’s first rule comes before the first prompt. “You need to think about what it is I’m trying to do today. What are my scenarios? What are my prompts?” She walked in with Sarah already defined: the owner of a 15-person UK business, responsible for cybersecurity but not an expert in it, with a Finance Director who will want a say in any significant spend.
That definition came from research, not from Claude. “What I’m not doing here, Mike, is trying to create a target audience and persona from scratch. I already have some ideas, some insights from within my organization, and I’m using it to role-play and prompt and dive deeper.” Later in the episode, she named the blank page as the most common mistake people make: “Don’t try and start from blank sheets of paper ’cause it could lead you down the wrong path.”
The seed prompt then sets the rules of the interview. “I want to interview Sarah. Stay in character, answer from her perspective. Don’t give me generic marketing advice, and if something is uncertain, say so.” Luan’s full prompt is in the Episode 7 Blueprint below.
Ask broad, then chase the surprise
Luan opened wide on purpose: “Tell me about your business and your biggest priorities at the moment.” Sarah answered the way a busy owner would. Keeping the team utilized and billing properly came first. Cybersecurity was “on the list. It’s not top of the list.” Then came the line Luan was hoping for: “We have an IT provider. They handle our laptops, email, that sort of thing. I think they do something around security, but I couldn’t tell you with confidence.”
A survey can’t follow that up, because you wrote its questions before you knew Sarah would say it. Luan could. “Okay, you’ve said you use an IT provider. Have you asked them about cybersecurity?” Sarah admitted she had only ever raised it as “a side comment.” Asked to rate her confidence that the business was protected, she said four out of ten, and added without prompting that a year ago she’d have said six or seven.
Mike’s explanation for why the answers ring true: the models have read the books, and also “all of Reddit and YouTube videos and LinkedIn profiles,” so they know how people in that role talk and what worries them. Luan had her own check. She ran the same scenario a couple of days before recording. The wording changed between the two runs, and the overall picture stayed the same.
Switch seats to debrief
At 08:43, Luan changed the role: “Okay, switch position. You’re the marketing lead for the cybersecurity provider. What has this role-play told us so far about our persona?” Claude, now speaking as the marketing lead, pulled out the findings. Sarah doesn’t know what she’s already paying for. Her low score reflects honest uncertainty, not fear. She reasons by analogy rather than technical detail, and cybersecurity has to compete with recruitment and other systems for her attention.
Mike compared the move to taking the transcript of a real call, dropping it back into the chat and asking what you learned. Luan compared it to pausing a focus group to ask, “What does this mean? How am I gonna respond? What else do I need to ask?” She kept switching back and forth for the rest of the demo, and she restated the role each time, because the tool “tries to draw on memory sometimes, and sometimes the wrong memory.”
One thing to watch. The debrief also pointed out that nobody had asked about the Finance Director yet. Luan’s view is that Claude always tries to suggest your next step, which helps right up until it pulls you off the question you came to answer.
The real objection is the existing relationship
Back in Sarah’s seat, Luan asked how likely she now was to raise cybersecurity with her IT provider. “Quite likely,” Sarah said, but “not a fire drill.” She’d bring it up next time they spoke. Then she said something Luan called a really useful insight for any marketer: “I’ve realized I’ve been asking the wrong question all along.” She also admitted she was nervous about the conversation, because she didn’t want the provider to think they’d done anything wrong.
If your marketing asks a buyer to question or switch a provider, that’s what you’re up against. As the marketing lead, Claude turned it into a content brief: guides, comparison content, and “what good looks like” material that Sarah can read before that call, in place of book-a-demo offers. Keep the tone gentle, because “she doesn’t want to feel accused.” Cut the jargon too. “B2B, we all have jargon,” Luan said, and for a buyer who isn’t a specialist, it gets in the way.
Keep the persona from agreeing with you
“AI tools will always want to please you,” Luan said. The seed prompt’s instruction to skip generic advice is the first guard. The second is a question she asks throughout: “How confident are you about that answer? Give me a percentage.” If the number comes back low, she asks what would raise it. She compared it to questioning a junior member of the marketing team and asking why until you get what you need.
Mike added a tip he’d picked up elsewhere: tell the AI to act as a skeptical buyer, then move it through different stakeholder seats, the Finance Director included. Luan agreed, with a caution. Mapping a full decision-making unit and its objections still takes time. “It just means that you can do it from your desk.”
From avatar to campaign
Once the role-play has given her enough, Luan turns it into something the rest of the team can use. You can upload an avatar template you already like and ask Claude to complete it, or ask for a one-pager or a crib sheet for the sales team, written in the persona’s own language.
From there, she drafts a landing page or an email, puts it back into the chat, and asks how Sarah would respond, which objections she’d raise, and how many emails a reluctant buyer with other priorities might need. With clients, the benefit is running that scenario together in the room, instead of asking whether they’ve ever done any customer research. None of it needs a paid plan.
Grab the Episode 7 Blueprint
Luan’s seed prompt word for word, the interview sequence, both switch-position debrief prompts, the confidence check, and a persona briefing template to fill in before you open Claude. Download it here.
Connect
CONNECT WITH LUAN WISE
Listen and subscribe
Full episode transcript
Luan Wise: You know, as we all do with shiny new tools, you wanna dive straight in. But no, plan this. Think about this. Who is this avatar? We’re not using it to start off from scratch and say, “I sell cybersecurity software; who’s my customer?” Maybe it’s like, “I know what customers I already have in the business, and I want more people like that.” Or you could say, “I’m looking to grow my business and reach new audiences. This is what our existing customers already do.”
Mike Allton: Welcome to The Lab. I’m Mike Allton. Stop guessing how your target audience will react and start asking them directly. Today, Luan Wise, a veteran marketing consultant and author, shows us how to build custom B2B decision-maker avatars in Claude to pressure test social campaigns live. Hey, Luan. Welcome to the show.
Luan Wise: Hi, Mike. Thanks for having me.
Mike Allton: So glad to have you on. What made you start building these kinds of AI avatars instead of just running standard customer surveys?
Luan Wise: So I’ve always been absolutely passionate about marketing fundamentals. My career started before social media and, of course, before AI as well, and I’ve always found that if you’re reading any books, texts, or advice, it skips over the target audience bit really quickly. It’ll just say, “Understand your target audience, know something about them.” In practice, it’s not that easy. It’s also not that easy because people don’t have the time to do it, and they very often don’t have the resources either. If you’re going to do this properly, you want to be speaking to your customers or prospective customers through research, questionnaires, and focus groups. That takes time, energy, and can have costs—it’s a really long process to do properly.
So when we started playing with AI—and it’s amazing to think that it’s nearly four years since ChatGPT landed and we started experimenting with all the new shiny things—we did the quick wins, but then it was, “How can we really use this? How can we do more with this?” I love the opportunity to use AI tools to go deep, dive into things, and experiment. Role-playing is something that I heard someone talk about, and I was like, “I’m going to give that a go.”
Plan Before You Prompt
Mike Allton: Awesome. So let’s pull up your screen and dive in. Show us where you start when you’re building a brand new B2B decision-maker avatar in Claude.
Luan Wise: I think the important thing when you’re doing something deep in an AI tool is that you don’t just have your screen open like this and start typing straight away. You need to think about this. You need to think, “What is it I’m trying to do today? What are my scenarios? What are my prompts? What is it you’re trying to achieve?” In the same way that you would prepare if you were doing a real-life focus group, you would plan your outline and your interview.
So I’ve planned this. I’ve thought about who my decision-maker is and what kind of organization she works for. I’m role-playing as if I’m a cybersecurity provider, and Sarah, my persona, my decision-maker, is running a business. We’re going to ask her where she sits on this right now, whether she’s got provisions in place, and we’ll see how that pans out.
So let me copy and paste my first prompt in, which I’ve got right here.
The Seed Prompt
Luan Wise: I’m giving Claude the context that you are Sarah, the owner of a 15-person business. I’m also saying based on the research that I’ve already done—so what I’m not doing here, Mike, is trying to create a target audience and persona from scratch. I already have some ideas and insights from within my organization, and I’m using it to role-play, prompt, and dive deeper.
So I’m telling Claude this and then saying, “I want to interview Sarah. Stay in character and answer from her perspective. Don’t give me generic marketing advice, and if something is uncertain, say so.” So I’ve given the context and then the scenario: I’m a cybersecurity provider, and Sarah is the decision-maker. She is not a cybersecurity expert, but she is responsible for it. And I’m going to ask Sarah, “Tell me about your business and your biggest priorities at the moment.”
Mike Allton: Now, is there anything in there that you feel helps that avatar act like a real buyer?
Luan Wise: I think stating the size of the business, what her role is, that she’s not an expert already, and asking her about the business and her priorities. It’s a broad question, and I will be narrowing it down, but I’m still giving those scenarios as well, and we’ve got the answer right here.
So Claude, as she is repeating back, “We’re a 15-person business based in the UK. This is what we do. Our biggest priorities at the moment, honestly, are keeping the team utilized and billing properly. That’s the day-to-day.” And then this is where it’s already getting interesting: “Cybersecurity is on the list, but it’s not top of the list. It’s one of those things on my mind.” And then another insight already: “We have an IT provider. They handle our laptops, email, that sort of thing. I think they do something around security, but I couldn’t tell you with confidence.” So she is aware she hasn’t done something, but she’s assumed something in her business, which is kind of scary that people would do that around something like this.
Chasing the Answer a Survey Would Miss
Luan Wise: So this is where it already gets interesting because if you were doing a regular survey, you would have predefined your questions, and you wouldn’t know how your persona was going to respond. But I have been doing this for a while, so I can immediately look at this and go, “Okay, there’s something there I want to dive into.”
So now I’m going to say, “Okay, you’ve said you use an IT provider. Have you asked them about cybersecurity?” So I’m going to prompt into that.
Mike Allton: And what a lot of folks don’t maybe realize is these AI tools have read books, yes, and they’ve read articles, but they’ve also read all of Reddit, YouTube videos, LinkedIn profiles, and that sort of thing. So they know a lot of demographic and psychographic information about any role or buyer you could potentially put in there, and they’ve seen the kinds of comments from those kinds of people on forums and threads like Reddit. So they know what they talk about, they know what they’re concerned about, and you can trust that this is a fairly realistic conversation you’re about to have.
Luan Wise: Yeah, it feels realistic as well, just in the way the response is. So here we have, “Yeah, I have in a roundabout way, but I don’t think I’ve asked them directly, ‘What exactly are you doing to protect us from a cyber attack?’ It’s more a side comment.” And then she’s kind of going on saying, “At the time, that reassured me, but thinking about it now, I’m not sure that’s right. Never had a proper conversation, maybe, but it’s crossed my mind.”
So I’m understanding already that this owner of a 15-person business knows it’s something on the list, but not top of the list. She has assumed something, never really asked the question, and that’s kind of dangerous, I think.
And now, my point in role-playing as a cybersecurity provider is providing me with some really clear insights. I’m kind of relieved by the responses because I did a practice run-through, as I shared with you. The responses aren’t identical to what I did a couple of days ago, which is interesting in itself, but the overall concept is the same. So it’s not identical, but it’s giving me the same kind of feelings and prompts.
So just to keep going with this, I’m now going to ask, because we’re still in the role-play scenario, “Do you feel confident that your business is protected right now on a scale of one to 10?”
Mike Allton: And so far I would say, yeah, those responses make perfect sense to me as well. Having worked with people who have businesses at that size, they’re very much focused on the day-to-day of their business, and thinking about risk scenarios usually isn’t top of mind.
Luan Wise: No, not until something happens or you see something in the media and it alerts you to it. So, the response now is, “Maybe a four. I haven’t got a specific reason to think we’re at risk right now, but I can’t confidently say we’re an eight or nine.” Interesting.
And I think one of the good things about this is if you were doing this in real life with real humans, you might need to prompt a little bit more, whereas Claude is kind of going into a bit more detail here by saying, “I’d probably say a year ago it was a six or seven if you’d asked me. Four feels honest. I don’t think we’re about to be attacked,” but she’s now realizing that this is an issue and might need to do something about it. And this is not the prompt now.
Switching to the Marketing Lead
Luan Wise: This is, again, something different, but now I’ve got so far with this and I’ve got some insights, so I’m now thinking, “Okay, I want to switch roles and start to summarize this now.” Because if I want to prompt with more questions, I need to have done some thinking as a marketer on how this is going to impact my activity, my campaigns, and my messaging.
So my next prompt that I had ready was, “Okay, switch position. You’re the marketing lead for the cybersecurity provider. What has this role-play told us so far about our persona?” So what have we got? She doesn’t know what she’s already paying for. Her low confidence score isn’t fear; it’s honesty about uncertainty with some reasoning. She has a live, specific worry about her IT provider’s incentives. She reasons by analogy, not by technical detail, so that’s the language that she uses to talk about cybersecurity, not techy language. That’s important. She had a trigger event that was a near miss, but not big enough to really do something about at the time. It’s competing with—and not losing to—other priorities such as recruitment and other systems.
Claude has picked up now from the original scenario—I didn’t read it out, but I did say Sarah isn’t the only decision-maker. That is one of the things about B2B: very rarely will you have a single decision-maker. You could have eight, nine, 10, or possibly more. So if you read back through the prompts, I did say her finance director will want to get involved in this as well, and at this stage, Claude has picked up and said, “We haven’t yet asked about the finance director’s involvement. We might do that.” This is one of the nuances around AI—it’s always trying to be helpful and prompt us as to what to do next. That can, of course, be useful, but at the same time, it’s important to stay on track with what you’re trying to do through this exercise.
Mike Allton: Just to summarize and make sure I understand—that way the audience understands—the first part, the first prompt, and then a couple of the successive prompts were all about setting a scenario as though you were interviewing a real, live person to gain some insights that the AI truthfully and transparently generated. And now you’re taking those, and with a final prompt, analyzing that conversation as though you had taken a transcript of a call, dropped it back into that same chat, and said, “Okay, this is the person I talked to. Tell me what we’ve learned from this conversation.” Does that sound about right?
Luan Wise: Yeah, and it’s almost, again, if you were doing this in real life, you’d probably want to stop a focus group and go, “What does this mean? How am I going to respond? What else do I need to ask?” So being able to do it so quickly, but also pause when you need to, is really valuable about being able to role-play with an AI tool.
Mike Allton: Now, you mentioned how when you had run this before, you had gotten a somewhat different set of responses. Would you imagine you might want to do this several times in a row before you actually finalize the avatar, or do you think one run-through is probably sufficient?
Luan Wise: Having seen it run a couple of days ago and running it again, I don’t think there’s anything significantly different that would change my avatar. But it could be that you do run it with different criteria, perhaps around the persona…
Mike Allton: Got it.
Luan Wise: …maybe a slightly bigger business or something else.
Would Sarah Ask Her IT Provider?
Luan Wise: So my next one is kind of now switching back. So if I just copy and paste this in here, I’m now saying, “Back to your role as Sarah: given this conversation, how likely are you now to ask your IT provider about cybersecurity?” So I’m thinking, you know, I’m having a conversation with a prospect potentially and flagging some items in her mind. If we were to do that in our marketing, would she go away and ask her IT provider about this?
“Quite likely. I think I’ll ask them. Not a fire drill—we need to talk urgently—but just next time I’m speaking to them.” So there’s still not that real sense of urgency to her as a business owner that, “Oh, this is something that we need to do tomorrow,” but it’s more like, “Yeah, it’ll stay on the list, and the next time, I’ll ask, ‘Have we got this covered?'”
Interestingly here, I think what’s changed is I’ve realized I’ve been asking the wrong question all along. That’s also really useful insight for a marketer—you know, you’ve got your prospective customer who doesn’t really know what they don’t know and doesn’t know how to ask those questions to an existing provider. So she is now saying, to be honest, she’s nervous about having that conversation because she doesn’t want them to think they’ve done anything wrong. This is really highlighting existing provider-supplier relationships and how people feel about it. If your role as a marketer is to get people to switch providers, that’s really useful to know as well—what’s going to trigger that or what’s worrying them.
So to finish this response: “I won’t treat this as urgent enough to drop everything, but next time on a call, I’ll see how that goes.” She’s being realistic. She’s kind of saying this isn’t a top priority for her.
I’m going to switch back. I’m going to keep switching back now in my role-play because I want to build on this persona and the insights, but also, at the same time, how does this impact my messaging?
What the Persona Means for Your Content
Luan Wise: So we’re switching: “You are the marketing lead for a cybersecurity provider.” I sometimes like to remind Claude or the AI tool I’m using, because it tries to draw on memory sometimes—and sometimes the wrong memory—so I keep it in there. “Again, what can we learn from this to guide marketing activities?”
So she’ll ask, but only when prompted. It’s not going to be a cold start.
Mike Allton: Yeah.
Luan Wise: The timing will be the next call, not today. There’s no urgency on this. She wants to prepare before the conversation, so she’ll write down some questions and do some reading around it first. And again, as we know in B2B, people will go and do their research before they ask a real salesperson. So even with an existing provider with a long-term relationship, she still wants to go away and do a bit of research and prep before having that conversation.
So what it’s telling me from a marketing perspective—because I’ve asked it about marketing activities—is that guides, comparison content, and “what good looks like” could be useful. Prep material, educational stuff, not the “buy now,” “book a call,” or “book a demo” conversation. We need to be gentle as we go through this. She doesn’t want to feel accused, like, “You’ve not done something,” or “You’ve missed something.”
Jargon is also a barrier. In B2B, we all have jargon. As marketers, we have jargon, and we don’t always realize it. But this is also going to be really important to her as a prospect: in anything we do as marketers, we need to simplify the language, not make it too heavy, and be seen as being helpful and educating as well.
Mike Allton: Yeah, that’s so appropriate. Love all that advice. You’re already getting really great insights. So anybody listening and watching, if you’ve ever wondered how you’re supposed to resonate with your target audience, particularly if you’re trying to sell something that maybe they don’t even know that they need, this sounds like a really smart conversation to have.
Turning the Avatar Into Campaign Assets
Mike Allton: Now, does this avatar just exist in this chat, or do you turn it into a project, a skill, or what do you do with it to actually save it as an avatar?
Luan Wise: Yeah, so if anyone’s done traditional avatars, you almost have a template that’s like a record card or a poster on the wall where you give it a name, a photo, and fill in some boxes. So I think you could keep doing this role-playing with Claude, and then maybe if you already had an existing avatar template you like to use, you could upload that and say, “Complete this,” or get to the end of this and say, “Right, I now need to share this information with the rest of my team, with my sales team, or others within the organization.” You know, we’d love some sales and marketing alignment in this ideal world. You could ask, “Create me a one-pager,” or “Create me a crib sheet of bullet notes of what’s going to be valuable to them in their language.”
The other thing is, once you’ve gotten this far and have that, we could maybe draft a website landing page. We could draft an email and then pop that back into Claude and say, “Okay, given the avatar that we’ve created and role-played, how do you think Sarah would respond to this email? What objections might she have? What’s the workflow that I could work through, knowing that this isn’t a top priority, that she might be reluctant, and that it needs to be jargon-free?”
You can then start planning your marketing campaigns in here as well.
Keeping the AI From Agreeing With You
Mike Allton: Love it. Love it. Yeah, I could imagine having it create that documentation for you, saving it to a Claude project, or if you’re using one of the other platforms, a ChatGPT notebook or something like that. That way, you can continue to have lots of chats with it. How do you keep it though from just agreeing with you or giving you generic, positive feedback—like, “Oh, this sounds like a great offer. I would buy today”—kind of a thing that we know is not realistic? How do you keep it from doing that?
Luan Wise: AI tools will always want to please you. They’re always trying to keep you happy, and they’re trying to prompt you on what you want to do next. Again, in my original prompt, which I know I copied and pasted quite quickly, I put, “Don’t give me generic advice, and if you’ve got any questions, let me know.”
Something else I quite often do as I’m diving into this is ask, “How confident are you about that answer? Give me a percentage.” And then, if it’s a low percentage, I’ll say, “Okay, what could we do to make you more confident of that?” So almost just as if you had someone sat next to you in real life—maybe a junior member of your marketing team—you would just keep interrogating almost that kind of “why, why, why” until you get to what you need to know.
We can’t take at face value what an AI tool is giving us. They don’t have the experience or the knowledge. They have so much training and input, but this is where we add the value. And it just makes it quicker, real-time, and cost-effective as well. You can do this with a free version; you don’t need a paid-for version of the tools to do this.
Mike Allton: Yeah. One of the tips that I’ve seen elsewhere says put in the instruction, “Act as a skeptical buyer.”
Luan Wise: Yeah.
Mike Allton: Right? And just that word “skeptical” tells the AI everything it needs to know about the role that it’s playing in that conversation. I think to continue on with the example and the technique that you were showing us there, at each step of the way, you could tell the AI, “Okay, now I want you to be a skeptical buyer,” or “Now I want you to be in the finance person’s role,” right? And put them in different stakeholder seats throughout the organization.
Luan Wise: Yeah, absolutely. And that’s why this isn’t a quick process. In real life, building your decision-making units and understanding who’s involved and all those objections still will take you time. It just means that you can do it from your desk.
The Biggest Mistake: A Blank Page
Mike Allton: Love that. That’s a great tip. What are some other mistakes, or the biggest mistake maybe, that you see other people making when they’re trying to role-play with AI for the first time?
Luan Wise: I think it’s one of those, “Oh, I didn’t know I could do that.” And, as we all do with shiny new tools, you want to dive straight in. But as I said earlier, it’s like, no, plan this. Think about this. Who is this avatar? What do you know about them already? That is going to be really key. We’re not using it to start off from scratch and say, “I sell cybersecurity software. Who’s my customer?”
Maybe it’s like, “I know what customers I already have in the business, and I want more people like that.” Or you could say, “I’m looking to grow my business and reach new audiences. This is what our existing customers already do.” So give it lots of context and insight. Don’t try to start from a blank sheet of paper because it could lead you down the wrong path.
Mike Allton: Love that advice.
Luan Wise: You can do this by yourself, but maybe do it as a team exercise in a boardroom and share this as well. Do it as a group and make sure that you do share it.
An AI Board of Directors
Mike Allton: I love that you mentioned boardroom. I was listening to Geoff Woods, who wrote The AI-Driven Leader, and he was sharing that he loves doing this process. The story he shared was that he walked into a business where they were struggling with their board, presenting to the board, and getting shot down all the time.
So he had them create an AI avatar for every single board member. They interviewed people who knew them, created the avatars, and then presented their next presentation to each different AI board member and got feedback, integrated the feedback, and delivered the new revised presentation. It was the best board meeting they’d ever had in their existence because they had that back-and-forth with the AI giving them feedback as if they were talking to the board members without actually doing it.
Luan Wise: Yeah. On your recommendation, I read that book, and I would highly recommend it. It was a brilliant read with really practical insights and prompts, and again, almost taking you through this kind of process as well, so you can read it and then implement it straight away. That’s the other good thing. As a practitioner, you can read textbooks and go, “Well, when am I ever going to get the chance to do a focus group or run an interview with my customers?” You can put all these ideas that you read about or hear people talking about into practice and get started on them straight away.
Role-Play With Clients
Mike Allton: Love it. My last question, Luan, I’m just wondering: are there specific outcomes or time savings that you’ve unlocked for your clients now that you can pressure-test these campaigns, like you said, immediately?
Luan Wise: I think it is about being able to pressure-test and have that conversation, and not talk to a client and say, “Have you asked your customers this? Have you ever done any research?” It’s more like being able to do this scenario with my clients and say, “Let’s pop this in. Let’s have a chat and see what they think, and what would you do?”
And the benefit I’ve found—and this is part of the consultant’s role—is that you’re not so in the weeds on a day-to-day basis, thinking about what your manager might say or what a real customer could say. I can be quite objective about this and probing. So actually doing that with a fresh pair of eyes, alongside inside knowledge, insight, and detail, is a good combination, I think.
Mike Allton: Love it. Luan, this has been so fascinating. I can’t wait to do this by myself, and I’m sure all the folks watching want to do it themselves as well. But for folks who might have more questions for you, where can they connect or reach out to learn more?
Luan Wise: Yeah, I spend most of my time on LinkedIn, so you can find me on LinkedIn.
Mike Allton: Fantastic. We’ll have all Luan’s links down in the show notes below, as well as the link to download all the prompts and workflows that Luan just walked us through. If you found this technical breakdown valuable, please make sure you subscribe on YouTube or leave us a review on Apple Podcasts. It helps other serious operators find the show. See you next Monday.
That’s how real practitioners are winning with AI on social. The AI Social Playbook is brought to you by Agorapulse and archie.app. Make sure to subscribe to the channel on YouTube so you don’t miss a single hands-on session, and grab the link in the description for the complete episode blueprint. See you in the lab.
