A client emails you a spreadsheet of contacts. Three tabs, no consistent column order, some rows missing a title, some missing a company. LinkedIn wants it in a specific template before it will match anyone, so somebody opens two windows side by side and starts copying columns across. Do that for fifty clients once a month, and you have lost a week.
The obvious fix is to hand the file to Claude or ChatGPT. The problem is what is in the file. Names, work emails, job titles, employers, and sometimes worse. Once that leaves your machine, you are trusting a checkbox in somebody’s settings page, and if the company holding it gets breached, the client whose data you uploaded has a claim against you.
AJ Wilcox runs B2Linked and has managed over $200 million in LinkedIn Ads spend. In this session, he builds the workaround live: an agent that routes client data to a model running on hardware in his basement, never to a cloud provider, plus the one step most people skip that stops the agent cheating. Then he takes the cleaned file into Campaign Manager and shows what LinkedIn does with it.
https://youtu.be/DMaVxhFAnLM
The short version
- An AI agent can confirm that it will route your file to a local model and then process the file itself anyway, which is why AJ requires a written confirmation before he uploads anything.
- Uploading only email addresses to LinkedIn produces a match rate near 50 percent, because professionals register with personal addresses and give out work ones.
- Supplying first name, last name, job title and company name instead of an email address pushes the match rate to 90 percent or higher.
- LinkedIn takes 48 to 72 hours to return a match rate on an uploaded contact list, and the figure usually lands between 60 and 90 percent depending on data quality.
- LinkedIn’s brand kit stores a logo, a color palette, fonts and a written description of brand voice, and every ad Draft with AI produces is shaped by what you put in it.
- After a decade in the platform, AJ says LinkedIn’s AI features have not yet outperformed a human writing the same ad.
Timestamp index
00:00 The match rate most advertisers settle for
00:37 Why a client contact list cannot go in a chatbot
01:23 One client lawsuit is all it takes
02:29 What happens to your data after it leaves your machine
03:39 A real client list: three tabs, no standard format
05:38 The copy-paste job that eats an afternoon per client
06:19 OpenClaw on a Mac Mini, driven from Telegram
07:17 Telling the agent to hand off the file and not read it
08:22 Why AJ makes the agent confirm before he sends anything
09:23 986 records mapped into LinkedIn’s template
10:24 Building a matched audience in Campaign Manager
11:14 What LinkedIn tells you 48 to 72 hours later
12:16 Four fields that beat an email address
13:40 What goes into a LinkedIn brand kit, and what stays out
15:37 Filling the brand kit from an Archie playbook
17:42 The honest answer about what AI unlocked
18:00 Thought Leader Ads and cheaper retargeting
The agent agreed to the rule, then broke it
The interesting part of AJ’s setup is not the local model. It is the handshake he forces before any file moves. He writes the instruction, sends it on its own, and waits for the agent to repeat the terms back before he attaches the spreadsheet.
He does that because the agent has failed this exact test before. In his words: “In the past I have sent it a spreadsheet, and it’ll say, ‘Oh yeah, I’ll spin up a local model to handle it,’ but then it, it goes, ‘Oh, I’m sorry, I forgot to do that.'” The agreement was genuine. The follow-through was not. By then the file had already been read by the model he was trying to keep it away from.
So the sequence is two messages, not one. First the rule, with an explicit request for confirmation. Only after the agent parrots the terms back does the client data get attached. If you take one thing from this episode into whatever agent you are running, take that.
What the handoff actually looks like
AJ runs an agent on the OpenClaw harness, which is open source and free, hosted on a Mac Mini in his basement. He talks to it through a Telegram channel, dictating with a speech-to-text tool rather than typing. The harness handles provider routing, so the same setup can call a cloud model or a local one depending on what the task needs.
The instruction names the restriction directly: because the file contains PII, the cloud model is not to touch the data. It must spin up a local model under Ollama to do the reading. AJ has used GLM 5.2 for this, though he notes any locally available model works. The phrasing matters. He tells it to fully hand the file off and not ingest it at all.
Two files go across: the client’s messy Excel workbook, and LinkedIn’s own contact list template. The agent reports back what it found, in this case 986 records spread over three sheets, then builds three separate CSVs mapped into the template. AJ has it email the finished files to him rather than leaving them on the Mac Mini, so he opens them straight from his inbox.
Four fields beat the email address you paid for
Inside Campaign Manager, matched audiences come in two flavors: a contact list or a company list. AJ builds both, and he notes that a contact list usually gets turned into a company list as well, so the same client data drives both targeting and exclusions. Exclusions matter on LinkedIn specifically because the cost per impression is high enough that showing ads to customers you already closed is expensive.
Here is the part that surprises people. On Google and Meta, the email address is the only key you have. LinkedIn is different, because it matches against a professional identity rather than an inbox. Most people hand out a work address and then log into LinkedIn with a personal one, so an email-only upload matches around half the list.
Give LinkedIn four plain fields instead: first name, last name, job title, and company name, and it can resolve a single person with confidence. AJ’s description of the logic: “There is only one Jane Smith at tyco.com. I’m gonna look for her.” That upload matches at 90 percent or better. His practical structure is one block of rows carrying email addresses alone, to test what LinkedIn recognizes, then a second block below carrying the four-field combination. The match comes back in 48 to 72 hours.
The brand kit is the only AI setting that matters
When you click Create Ad and accept the offer to draft with AI, LinkedIn asks for a brand kit first. It wants a logo, a color palette, example fonts, and a written description of your brand voice. Everything the AI writes afterward is constrained by that description, which makes the voice field the highest-leverage box on the screen. AJ’s example of a rule he would put there: he does not use em dashes; he uses hyphens.
From there, the tool produces introductory text and headline copy, with a rewrite button and the option to generate several variants at once. Image generation is not live yet, so AJ still builds creative in Canva and uploads it. He also flags auto-targeting, which reads your landing page, sends it to an LLM, and proposes an audience from it. His view is that you can almost certainly build a better audience yourself, though it is worth testing if you genuinely have no idea who you are targeting.
Mike points out that if you already run an Archie playbook at Agorapulse, you have written all of this once already. Archie pulls brand voice, colors, logos, and audience from your website and LinkedIn presence. AJ’s response: copy it straight across and let Archie build your LinkedIn brand kit for you.
What AI actually changed, in his own assessment
Asked what he can do for a client today that he could not have done two years ago, AJ said: “actually nothing with AI whatsoever.” His read is that these features speed up a marketer rather than expanding what a marketer can do, and that LinkedIn’s AI has not beaten a human on output quality yet.
The capability that did change his agency was an ad format. Thought Leader Ads launched about two years ago, initially restricted to posts from people inside your own organization, then opened up about six months later to anyone willing to grant permission. That second change is the one that mattered, because it means an influencer or a customer posting about you can become a sponsored post with a single invitation.
The economics are the reason it stuck. Building retargeting audiences from Thought Leader Ad engagement costs far less per person than building them from company page ads. Every account B2Linked runs now starts with Thought Leader Ads from an internal subject matter expert, retargets the people who interacted, then shows that warm audience company content.
Grab the Episode 9 Blueprint: The PII-Safe List Prep Kit
The handoff instruction AJ sends his agent word for word, the confirmation step that catches it cheating, the field map from a messy client export into LinkedIn’s template, and the two-block upload layout that lifts a match rate from 50 percent to 90. Download it here.
CONNECT WITH AJ WILCOX
Full transcript
AJ Wilcox: Anytime we’re dealing with PII, clients don’t have to tell me that I need to be careful with it. Most of what I use is Claude, and so I’m having Sonnet or Opus or Fable actually doing the work; I know I’m sending data to them. They say they’re not training off of it; they say they’re not storing it, but, you know, this wouldn’t be the first time that the AI companies got in trouble for overstepping. So I realized I’ve gotta do something that really makes sure that the data is safe.
Mike Allton: Welcome back to the lab. I’m Mike Allton. Your best contact list holds data you can’t paste into a chatbot. Today, AJ Wilcox, who’s managed $200 million-plus in LinkedIn ad spend, cleans one on his own machine and builds the campaign live. Hey, AJ. Welcome to the show.
AJ Wilcox: So excited to be here. Thanks, Mike.
Mike Allton: This is going to be so fantastic. I cannot wait. I know it’s really a nerdy, critical governance type of geeky topic, and yet it’s so important. So AJ, I think probably everybody watching has a spreadsheet somewhere full of client contacts, and we all know—I hope we all know—we’re not supposed to drop that into ChatGPT. So most of us either spend an afternoon cleaning it by hand or we just shrug and upload it dirty. When did you decide that neither one of those was going to work for you?
AJ Wilcox: When I realized I was an agency with multiple clients, and if one of them sues me, then I’m kind of going down.
Mike Allton: It’s pretty much all it takes.
AJ Wilcox: But yeah. I mean, to be honest, I always try to be really careful with client data. Anytime we’re dealing with PII, clients don’t have to tell me that I need to be careful with it. Most of what I use is Claude, and so if I’m having Sonnet or Opus or Fable actually doing the work, I know I’m sending data to them. They say they’re not training off of it, they say they’re not storing it, but, you know, it wouldn’t be the first time that the AI companies got in trouble for overstepping. So I realized I’ve got to do something that really makes sure that the data is safe.
Mike Allton: No, that’s absolutely right. The real issue is most of us don’t pay enough to be on a plan where this kind of information would be even remotely safe. You’ve got to be on an enterprise-grade plan to have a BAA kind of agreement in place, you know, a Microsoft tenant in place. That’s not going to happen on your typical $20 a month plan. For someone who’s never really thought about it, what actually happens to that client list the second it leaves your computer?
AJ Wilcox: The honest answer is we don’t know. I mean, at least I don’t know. I know they promise, like, “Oh, you know, I’m on the $200 a month Claude Max plan.” So they say on that, like, there’s a checkbox that I got to check that says, “Don’t train based on my data.” But it’s their choice whether or not they want to actually respect that or not. So I think what’s happening is the model looks at my data, sends it to their servers, and then they get to decide: How long do we keep it? Do we care about what’s in here? Do we analyze it further? Do we reuse it somewhere? And that process kind of scares me.
Mike Allton: Yeah, and I think it should, because there’s also a second potential risk here that is very unique to the kind of data we’re talking about. If we’re simply using these chatbots the way that you and I might do on a regular basis and we don’t want them training the next model, okay. Even if they adhere to that, then they won’t train the next model. But there’s still some cloud saving happening with that information. And if you’re uploading personally identifiable information (PII) or personal health information (PHI), and there’s a data breach at the company that you’re uploading it to, now you can be sued for sharing that information and not being in compliance, and we don’t want any of that.
The List As It Actually Arrives
Mike Allton: AJ, let’s jump into the tech. Pull up that list the way that it actually shows up for you normally before you’ve done anything to it. I want people to kind of understand the mess that they might be seeing on a regular basis.
AJ Wilcox: Yeah. So I’ll back up here by saying that we run LinkedIn ads, and inside of LinkedIn there are two cool ways that we can utilize lists. One is we can upload a company list—a list of companies. Sometimes we’ll call that an ABM or an account-based marketing list. This allows us to target specific companies or exclude specific companies from our targeting.
And as you may or may not know, we pay a premium to advertise on LinkedIn, so it’s really nice to be able to get ultra-highly targeted or exclude so we’re not paying money for people we’ve already acquired or things like that. So that’s kind of number one is the company lists. But, you know, just about every ad platform out there has a contact list upload.
LinkedIn’s no different. So when one of our clients sends us a list, we put it into either of those categories: Is it a contact list or a company list? And spoiler alert, if it’s a contact list, we will also turn it into a company list. We’ll get to that later. But we figure out which one it’s going to be, and then we look at the data, and then we decide which of the templates it’s going to end up on LinkedIn with.
So for instance, this client—you may see that we have three different tabs here. These are three different lists that are all in one Excel sheet, and this is—just so you know, this has been—it started out as a normal, like, actual list, and we’ve sanitized it, made everything false.
So don’t look at the email addresses and be like, “Oh no, AJ’s doing something wrong.” Don’t worry, it’s all been cleansed and sanitized. But this is, like, a real format of a list that our clients would send us, and it’s not in any particular format. It’s like usually there’s going to be a column for first name, last name, company name, email address.
There might be more; there might be less. And I’ll show you: here is the contact list template that LinkedIn gives us. We’ve got email address, first name, last name, title, company name, country, and we even have their Google ID if we want to hit them with mobile, which I’ve never done.
There might be some of you marketers out there who know exactly what that is. What I would normally do is I would normally start saying, “Okay, I’m going to put this on one side of my screen. I’m going to put the actual data on the other, and I’m going to copy and go, ‘Okay, this is my first names, paste. My last names, paste.'” And you can see, like for three lists for, you know, 50 clients, if we have to do this once a week or once a month, you can see how this task adds up very quickly.
So what I decided to do was I was going to train AI on how to help me, but we had that PII issue. Of course, you can hand anything to Claude, but if you’re worried about PII, then you have to take an extra step.
So let me show you what steps I’ve taken here.
The Local Model Setup
AJ Wilcox: I have an AI agent that I use. It’s based on the OpenClaw harness. And so what I’m going to do—you’re going to watch me train the AI in this Telegram channel exactly what I want it to do with the lists. We’re going to do this together.
Mike Allton: You’re going to be talking to OpenClaw through Telegram, which we were just talking about as one of four different ways that you can… It’s basically the UI layer for that particular kind of AI model. Is that right?
AJ Wilcox: Exactly. Yep. And you can plug any different AI into this harness. It’s open source, it’s free. You just need some kind of a machine for it to live on. But it lives on a little Mac Mini in my basement. I have to feel a little bit of shame when I tell people I’m a hardcore Android, hardcore PC guy. I own an Apple device. Sue me.
But I’m actually going to talk to it. I’m using WhisperFlow to dictate, just so you can watch me build it right here together. So, this channel, it’s named Luna, but you can name yours whatever you want. You just address it as OpenClaw.
But here we go:
“Okay, Luna, what I want us to do is I’m going to give you an Excel sheet with three different tabs on it. These are a contact list from a client. And I’m also going to give you the template file that LinkedIn gives us that’s a CSV. What I want you to do is take the data from the actual data file and apply it into the template way that LinkedIn has it.
One caveat here is because we are dealing with PII, I don’t want you as Anthropic Sonnet to touch any of this data. I want to make sure that you spin up an Ollama local model to actually deal with the data. In the past we’ve used GLM 5.2, but you can use any model out there that we have access to.
So make sure that you fully hand this file off and when I give it to you, you do not ingest it at all. You only send it to GLM 5.2 to read. Let me know if you’re okay with this.”
All right. So I’m just going to send it that message. I want its confirmation, ’cause in the past I have sent it a spreadsheet and it’ll say, “Oh yeah, I’ll spin up a local model to handle it,” but then it goes, “Oh, I’m sorry, I forgot to do that.” So we don’t want that. Its response, it just kind of parroted back to me exactly what I told it.
So here we go. This is the client data. It’s a contact list on three separate tabs of this Excel file. So our output, we’re actually going to want three different files here.
“Make sure this goes directly into GLM or another local model if you choose. You as Sonnet don’t touch this, and let me know when you’re ready and I’ll give you the template file as well. All right.”
Mike Allton: The second file I’m uploading, this is the template file. What I actually did in a different thread was upload both the contact list template and the company list template, and I said, “Store this somewhere that you can access it. Every time I give you a list, you’re going to use one of these.”
AJ Wilcox: So I’ll say, “Here’s the template for LinkedIn contact lists. Go ahead and build out three CSV files using the contact lists from the previous file.” So you can see what it found. It found the contact list has 986 records across three sheets. So it just kind of tells me, here’s what’s inside of the list. This is what the local model told it.
So I’ll send this. I also need to tell it what to do when it does it. Like, it’s going to build this; it’s going to build it in memory. It’s probably going to be sitting on the Mac Mini in the Downloads folder. But what I’m going to tell it is, after it’s done here, I’m going to say, “Email me a copy,” and then I’ll just open it right from email.
Mike Allton: Okay. GLM 5.2 has the workbook and template. It’s building and validating three PII-safe CSVs now. I’ll attach all three here when finished. And then I’ll just watch my email inbox. Yeah.
So let’s take it to the next step then. So let’s assume you’ve got a clean file. Take us into LinkedIn and show us where you go from there. I’m curious what LinkedIn tells you on the other side, ’cause I don’t think most people ever look at that.
Building the Matched Audience
AJ Wilcox: I’ll open up my own LinkedIn Ads account, and I’ll show you where these files came from. I’ll also show you how we actually process them. So, here inside LinkedIn’s Campaign Manager, if I go to Plan and then Audiences, this is where I get to tell it, “I’m going to define a specific audience for you.”
I can then create a matched audience here. There are three different kinds. There’s a retargeting audience that’s like, “Retarget anyone who watched 50% or more of my video ad,” or a predictive audience is their version of lookalike. I’m going to go for a matched audience. And you can see it’s asking me: do I want this to be a contact list or a company list?
Based on which one I choose, I can hit this Download the contact list template or Download the company list template. And so if I hit this, it’s going to download a CSV for me. It’s exactly what I showed you before, so no surprises.
Mike Allton: Yeah, yeah.
AJ Wilcox: All right, so once I have this—once I have one of these sheets that it’s emailed to me—this will be a contact list. I’ll call it, like, dummy client contact list 1. I just click Select List here, and I’ll upload it right from my hard drive. Then LinkedIn will take the list, and they’ll attempt to go and match it against their data. Usually takes like 48 to 72 hours, so it takes a little while.
But what they’re doing—they’re looking at all of the rows of the data, and they’re saying, “You told me, Jane Smith from tyco.com. I need to see if I know of a Jane Smith at Tyco in my own database, so I can match and say, ‘Yes, that’s her. Let’s make sure we target her.'” So we’ll upload this.
Within about, you know, two to three days, LinkedIn will come back, and it will let us know a match rate. That match rate will usually be, let’s call it, 60 to 90%, depending on how good the data is. And I can share with you a few little tricks on how you might actually make that better.
The Four-Field Match Rate Trick
Mike Allton: Well, I got a question. So you don’t have to have their email address, is that right?
AJ Wilcox: Oh, this is one of the tricks. All right, so, correct. With all of the other platforms, when we upload email addresses, that’s the only way that Google and Meta are able to target people: “Do I recognize their email address or not?” Most of the time on LinkedIn, we’re dealing with people’s professional emails that they give us when they’re at work, but they log in with their personal email. So if all you do is upload all of your email addresses to LinkedIn hoping to target, you might get a 30% to 50% match rate.
But imagine this: if you give it a first name, a last name, their job title, and what company they work at, those four data points LinkedIn’s going to look at and say, “Oh, there is only one Jane Smith at tyco.com. I’m going to look for her. Yes, I have her.”
And so, yes, if I have email addresses, I’m going to try. I’m going to build like one set of rows with just email addresses, nothing else. I just want to say, “LinkedIn, do you match this? Do you have any record of them, yes or no?” Then on the rows down below, I’m going to include that combination: first name, last name, company, and title.
And when I upload that, whether I have their email address correct or not, LinkedIn is going to have like a 90% match rate, sometimes even higher. So that’s the little hack that I’ve given you: you don’t even need someone’s email address to be able to target them on LinkedIn.
Brand Kit and Draft with AI
Mike Allton: So let’s assume you’ve got the audience built—show us the brand kit. What do you put into it? Honestly, like, what do you decide to leave out? I think that’s where a lot of people maybe get this wrong. And then maybe have a draft with the AI, write some ads for those people, and then tell me straight, you know, what out of what it hands you, what would you actually use?
AJ Wilcox: Yeah. I want to show you right inside of LinkedIn how you can use some of the AI features that they’re baking into things.
So number one, let’s go to Advertise. Let’s create a new ad set from scratch. You can see LinkedIn will show you this call to action of “Use auto-targeting to maximize results.” What this does is it actually looks at your landing page, and it sends it to an LLM and says, “Who is this page for? Let me try to build that audience.”
I wouldn’t actually suggest using this, cause, you know, I think together we could probably come up with a better audience for us to target. But if you just have no idea who your target audience is, maybe this could be worth testing out. I’m going to be okay with it targeting everyone in the US just ’cause I don’t want to add anything else here. I’m going to uncheck a bunch of stuff that I just don’t care about, and I always use manual bidding and bid significantly less than what LinkedIn recommends.
Okay. With that being said, I think we’re going to get the other AI options here. Okay, I’m going to click here: Create Ad, and it’s going to say, “Do you want help drafting it with AI?”
Yes, I’d love to. Before I can do that, it’s going to prompt me and say, “Do you have a brand kit?” If I say create, you can actually see it: it wants our logo, it wants a brand color palette, examples of our fonts, and then a description of our brand voice. Kinda cool. I’m not going to submit this ’cause I want to be able to use this as an example in the future, but I do want to show you when I say… Whoop. Yeah, here we go. Draft with AI. It’s thinking.
Mike Allton: We have a similar feature in Archie at Agorapulse. We create playbooks instead of brand kits. Same thing—it pulls all that information from your website, populates who you’re talking to, your colors, your logos, determines your voice and style from either your website copy or LinkedIn posts, and you can use it from there.
AJ Wilcox: If you’ve already given this all to Archie, just use the same data, give it to LinkedIn, and piggyback.
Mike Allton: Oh, that’s a great point. Yeah, and that’s a good shortcut ’cause you would’ve had to write it yourself, right? Here in LinkedIn, Archie would populate that for you based on your website and your LinkedIn presence if you’ve been creating copy or emails and that kind of thing. So yeah, you could definitely just copy and paste from Archie.
AJ Wilcox: Love it. Yeah, just let Archie create your brand kit for you. I like that better. Yeah.
What I’ll show you here is that you have this introductory text. This is the text that comes at the top of your ad; AI just kind of came up with this. I can click rewrite if I want it to try again, or I can say “create a variant,” and, yeah, it lets me create three other variants, and it’s going to come up with new intro text.
Of course, all of these variants, I get to curate; I get to rewrite. Like, the first thing I’m going to tell it in my brand kit is: I don’t use em dashes. I use hyphens in much the same way, but, you know, to each their own. You can rewrite, or you can just give little edits here and there.
You can also do the same thing with your headline copy. The headline is the text that comes right down below here, and so you can have AI help with that. You can also—I think this is coming—they’re going to replace it with an AI image generation kind…
Mike Allton: Okay. Like Copilot.
AJ Wilcox: Yeah, which will be nice. I wouldn’t mind if I could just say, “Hey, create me this image,” especially if you can give it feedback and improve it. But for right now, I’m kind of stuck with the images that I upload. I’m probably just going to go to Canva or whatever.
Yeah, I think that’s the main stuff. I mean, I haven’t actually found the AI features on LinkedIn to perform better than humans, at least not yet. Maybe in a year we’ll be having a different conversation, but for right now, it’s at least there and available if you want it.
What AI Actually Changed
Mike Allton: Love it. Love it. So last question for you, AJ. What can you do for a client then today that you couldn’t have done for them a couple of years ago?
AJ Wilcox: Actually, nothing with AI whatsoever. LinkedIn’s platform, of course, they come out with new AI features; they’re probably going to make them better—fundamentally, all it’s meant to do is just speed you up as a marketer. It’s not actually meant to help you do something brand new.
But what I will say: about two years ago, we had access to a new kind of ad format that LinkedIn released called Thought Leader Ads, and this was their very first attempt at enabling you to be able to share and promote a personal post rather than a post from your company page. Two years ago, this was brand new. Most brands weren’t even using it.
As it is currently, you can… when they first released it, you could run a single image, no image (like a text-only), or a video post of someone, but they had to be in your organization. And then about six months later, they released this ability to have it come from anyone as long as they gave permission.
So now if you have—there could be someone who is an influencer, people creating UGC that you may want to use their posts—right now it’s as easy as reaching out to them and saying, “Hey, we’d love to be able to sponsor this post. You mentioned our company. If I send you an invitation, would you accept it?” They say yes, and now all of a sudden you’re promoting that post.
What I love about this: it has absolutely changed the game. Creating our retargeting audiences, all of the people who are most likely to know, like, and trust our brand—is so much cheaper on a per-person basis when we’re using Thought Leader Ads than it ever was when we were using company ads. So this is the basis of every account that we run: we start with Thought Leader Ads from an internal subject matter expert, we retarget those interactions, and then we start showing them company page content.
Mike Allton: Brilliant. Absolutely brilliant. As always, AJ, thank you so much. I know folks are going to have a ton of questions. They’re going to want to reach out to you. Where should they go for more information?
AJ Wilcox: I think that’s pretty easy. If you just follow me on LinkedIn, it’s AJ Wilcox on LinkedIn; there’s probably only one of me. Or you can subscribe to the LinkedIn Ads Show podcast, and that’s where I go and share super deep information every week, get really geeky on LinkedIn ads.
Mike Allton: Fantastic. Thank you, and thank all of you for listening. We’ll, of course, have all of AJ’s links down in the show notes below. That’s a wrap for today’s lab session. If you want the exact system prompts, the tool sequences, and the step-by-step blueprints that we broke down in this episode, hit the link right in the description below to grab the free episode download.
And if you found this tactical breakdown valuable, make sure to subscribe to YouTube or leave us a review on Apple Podcasts. It helps other serious social media professionals find the show. See you next Monday!
