LEARNING HUB

Why AI Still Needs Humans (This Changes Everything in 2026)

Artificial intelligence continues to transform the way businesses market, engage, and operate—but as this week’s AI Marketing Experts Podcast reveals, rapid automation doesn’t mean humans are obsolete. Episode “Why AI Still Needs Humans (This Changes Everything in 2026)” dives deep into current AI releases, practical tools, and game-changing workflows, highlighting the vital role of human oversight in brand strategy, creative direction, and business growth.

Why does this matter right now? The pace of adoption is staggering. In just twelve months, AI marketing tech shifted from niche curiosity to mainstream essential—30% of business owners are actively integrating AI, with projections reaching 80% within a year. Nobody’s waiting for the Jetsons; flying cars and advanced automation are already piloted globally. Yet, even as AI takes over grunt work and routine tasks, industry leaders—including billion-dollar innovators—stress that “human in the loop” remains non-negotiable for creative strategy and brand voice.

Ready to catch up on this week’s most actionable AI ideas? Whether you run a small business, lead a marketing team, or advise high-value clients, here’s what you need to know—and how to act on it fast.

 

This Week in AI: Key Developments

Several headline tools, trends, and workflows were front and center:

  • CES Automation & AI Releases: The Consumer Electronics Show isn’t just about gadgets anymore. This year, AI-powered automation, including autonomous vehicles (yes, flying cars!), is moving from science fiction into stock portfolios. As John Clendenning puts it, “AI can now fly better than a human ever could.”
  • Human-in-the-Loop Emphasis: Even industry leaders building autonomous robots are warning: creative and strategic functions must stay in human hands. AI excels at executing repetitive tasks—but without oversight, brand damage is real.
  • PrettyPrompt for LLM Prompting:
    Brad Kilgore introduces PrettyPrompt, a Chrome extension that empowers anyone to instantly refine prompts for tools like ChatGPT, Gemini, and Claude. It breaks down technical barriers, letting teams achieve high-quality outputs without deep prompt engineering know-how.
  • Perplexity Integration: Smart use of Perplexity AI to read YouTube transcripts, summarize learning, and feed those insights directly into custom GPT projects enables faster ideation on the go.

These updates aren’t theoretical—they offer hands-on tools to streamline marketing, automate busywork, and boost creativity without sacrificing control.

 

Tactical Takeaways & Use Cases

Episode highlights brought practical examples and use cases for real-world implementation:

  • Automated Website Chat with Human Escalation:
    Brad Kilgore discussed advanced website chatbots that serve instant answers for common queries, yet offer seamless escalation to live video calls with humans, respecting the customer’s journey and providing high-touch support when needed.
  • Master Prompting for Brand Consistency:
    Chris detailed how even with robust, multi-page prompts (brand voice, customer avatar, services), AI shouldn’t publish content unsupervised. Humans review, correct, and guide, ensuring consistent, authentic messaging.
  • AI-Assisted Meeting Summaries and Action Items:
    Transcripts from meetings are auto-analyzed via tools like Perplexity and Comet, extracting actionable tasks and assigning them to team members in project management systems—a massive time saver that still needs human validation.
  • PrettyPrompt for Prompt Refinement:
    • Simplifies prompt creation for image generation or blog posts.
    • Is accessible via $5–$10/month, no prompt engineering required.
    • Supports “refine,” “improve,” and “library” workflows for repetitive team tasks.

Why it matters: These workflows allow small businesses to operate like enterprise teams—automating busywork, catching bloopers before they hit print, and positioning themselves a step ahead of competitors.

How to implement:

  • Add chatbots with human escalation options.
  • Use AI for initial drafts and error-checking, but always review before publishing.
  • Integrate tools like PrettyPrompt for prompt refinement, enabling team-wide consistency.

 

AI Workflow or Strategy Spotlight

Spotlight Strategy:
10-80-10 Framework for Effective AI Automation

John Clendenning showcased the “10-80-10 Framework” (originally from Dan Martell’s “Buy Back Your Time”), adapted for practical AI workflows:

  1. First 10%: Identify the task and craft clear instructions (or a prompt).
  2. 80% Automation: Let AI execute the heavy lifting—summarize meetings, generate content, analyze data.
  3. Final 10%: Human review. Check outputs, adjust messaging, approve publication.

Who’s it best for?
Business owners, marketing teams, and consultants looking to maximize AI’s speed without risking brand reputation.

How to use it:

  • Set up recurring workflows for content creation, client follow-ups, or meeting summaries.
  • Use AI to execute and draft—pause before publishing.
  • Finalize with expert human judgment.

 

What This Means for Marketers in 2026

AI’s accelerating adoption isn’t a prediction—it’s happening now, reshaping everything from customer acquisition to creative production. For agencies, consultants, and internal teams, here’s what to expect:

  • Competency Becomes Table Stakes: Routine tasks are automated. Differentiation will come from creative strategy and empathetic brand messaging—functions AI can’t fully replicate.
  • Hybrid Teams Win: The most effective teams blend automation with human expertise. They’re fast, but never reckless. They use AI as a co-pilot, not the sole operator.
  • Brand Risk Is Real: As noted in the podcast, letting AI run “rogue” can damage reputation, produce off-brand content, or escalate PR crises. Human oversight—at least at final approval—remains essential.
  • Customer Experience Evolves: Chatbots, instant answers, and smart routing are the new norm, but customers still crave human reassurance, complex assistance, and nuanced responses.

Bottom line: Marketers must leverage AI to scale, but keep people in the loop to protect their brand and grow loyalty.

 

Implementation Checklist

Take action this week:

  1. Review and map your current marketing workflows for “human-in-the-loop” gaps.
  2. Deploy a website chatbot that offers escalation to live human support.
  3. Integrate PrettyPrompt (or similar) to refine AI prompts—empowering your whole team.
  4. Record and transcribe internal meetings; use AI to extract action items.
  5. Set up a “Draft Mode” for all AI-created content—never publish without human review.
  6. Create a brand master prompt: include voice, values, avatars, and services.
  7. Educate your team about the importance of oversight in AI workflows.
  8. Monitor ongoing AI releases and adjust strategy monthly.

Frequently Asked Questions

Why is “human in the loop” critical for AI-powered marketing?
AI can automate content and responses—but humans ensure brand voice and strategy remain consistent and authentic.
PrettyPrompt is a Chrome extension that refines prompts for major LLMs, enabling everyone to create high-quality inputs without prompt engineering skills.
No. AI automates grunt work and repetitive tasks, but creative direction, strategic thinking, and nuanced communication still require human expertise.
Adopt a “review-first” policy: automate drafts and busywork, but always review and finalize outputs before publishing.

Conclusion

AI is changing everything—but not without you. As this episode shows, smart marketers blend automation with human insight to win in 2026 and beyond. Want more tactical guidance? Check out the [Show Guide & Resources] for deeper implementation steps, tool links, and insider tips. Your competitive edge starts here.
http://
Transcript

JOHN CLENDENNING
0:18

Welcome everybody back to another exciting episode of the AI Marketing Experts. Jon Clendenning hosting for this week, um, Carpet Cleaner Marketing Masters, and I’ve got my distinguished panel of friends that are going to be sharing all kinds of cool ideas in the world of AI to help you better run your business, understand the massive trend that’s changing in a— right in front of our faces, um, and how things are rolling out in the industry now. A little bit of a bookmark. So, um, you’ll remember, and I’ll— I, I bug Jen all the time about how Jen and I were swearing at our, our, our ChatGPTs and our 5.1s back in August And if you’ve ever watched something called the 2027 Project, that was right on the timeline. It gets advanced to the point we get pissed off and then it turns the corner and there’s an inflection point. We feel it in the industry, like just being, living this every single day. There is an inflection point where human beings are using their AI, the, as Chris calls it, their $1,000 computer on their hip, to answer questions that they, they would have asked their friends and family. Word of mouth is now the device.

JOHN CLENDENNING
1:33

1% the beginning of January 2025, arguably 30% or higher at the end of 2025. Fastest adoption in human history. Are we 80% this time next year? Are you in the game? We keep talking about it. That’s why you’re here. That’s why you’re listening. And we’re going to go talk about what’s hitting the industry right now, trends, things changing things you should pay attention to and how you can be a leader in your industry and let your competitors lag behind so you’re pushing forward and getting the biggest benefit from that. So with that said, hopefully everybody agrees with the, the premise here. Um, CES is really cool.

JOHN CLENDENNING
2:12

So CES is the, um, Consumer— I don’t know, does anybody know even say the acronym stands for? But it’s Consumer Electronics Show. Thank you. And it’s like it’s, it’s now it’s automation and AI. It’s like flying cars with nobody in it. It’s all that kind— they exist now. They’re, they’re rolling out in, in Dubai. They’re rolling out in place. This is not just hypothetical.

JOHN CLENDENNING
2:31

You can buy stocks in these companies. They’re already starting, right? So this, this is the Jetsons 2026, right? So, and it’s moving that fast, and it’s going away because AI can now fly better than a human ever could. So you don’t even have to touch the controls, and you can go from rooftop to rooftop, um, and that kind of thing. So that’s what CES is, and they’re talking all about automations. Now, the AI and automation experts, industry leaders, are still stressing— and this is an important topic that we like to talk about— is that while AI is accelerating execution, especially on repetitive tasks and things like that— I’m just reading the headline over here— um, creative strategy and brand voice, um, still need a human in the loop. So the tools are great for the grunt work, but you’re not— it’s not a set it and forget it, hope and pray, let it go out. Oh crap, it just damaged my entire company and my brand. That’s not what we’re talking about.

JOHN CLENDENNING
3:23

So with that headline in mind, I just thought that was a really intriguing headline to me is like, hey, that kind of dovetails into what we’re talking about. And these guys are making like autonomous robots and stuff like that. So if they’re saying that, how does that fit back into the local marketer? So who wants to go first and say, agree or disagree the human in the loop with these using all the AI tools that we talk about? Laura?

LAURA SOUTHERLY
3:46

1,000% agree.

JOHN CLENDENNING
3:48

OK, we’ll go with Jen. Yeah, Laura was smiling most, but we’ll go with Jen first.

JENNIFER CREGO
3:53

I think it’s so important to have a human in the loop. Just if you think about your AI, it’s, it’s going to want to please you and agree with you and do all of everything correctly. And which is great for many things. But when you’re talking about your brand and your messaging and like all of that, if all you’re doing, if AI is always just agree, agree, agree, agree, without pushing back in the areas that, that humans would and should, your brand messaging and voice, it just might be wildly different compared to what you want it to be, um, just in that process. And that’s just like one example of why you would want a human in the loop, uh, with brand. So Super cool.

JOHN CLENDENNING
4:58

Chris, what’s your input on that? Because I know you build things that where you put a lot of checks and balances where it does something and stops and waits for somebody to look for this exact reason. And that was one of the reasons why the headline caught my attention. You know, even just to let it loose on your team, hey, write this email back to this client. Is there, you know what I mean? And things like that. You wouldn’t want it potentially, you tell me, that it just sends it out with nobody looking at all.

LAURA SOUTHERLY
5:21

No.

CHRIS HUNTER
5:21

So, AI is smart. We all know that, right? It’s, uh, but it can get things wrong. And I’ve got a, I’ve got a good story about this that was actually this morning. So this is a timely topic. A friend of mine shared, uh, on Facebook, uh, you know, in one of his groups that I’m part of. He’s an SEO guy, right? So he shared that, um, he had, he had built— because he’s built some automation type, you know, things to do his process for, SEO, competition and, and, you know, all that kind of stuff, right? And really figuring out, what the holes and the gaps are in the SEO strategy currently, right? Well, it came back with this report. He showed it to the client. The client said, those competitors are not my competitors.

CHRIS HUNTER
6:10

Those, those people are like chucking at trucks. That they’re not my competitors, right? And so, you know, and he went on in his post about how it’s important to have that human in the middle, right? To ensure that we’re not just trusting AI to do absolutely everything just yet, right? It might get that way to where we can, but it’s just like a brand new employee, in my opinion, when they don’t really know what to do, right? They know, they know stuff, but it might not be right. Okay, so you’ve got to kind of keep some checks and balances in place there until such time as you actually trust them to do the entire job, you know, start to finish. At that point, you know, and so it’s the same thing with AI. I think at that point, eventually we’re going to be there, but there’s a lot of things that aren’t right. So you mentioned I’ve got a few things with humans in the middle. One of them is content., right? Because of that simple fact that it might not be right, and we’ve got to do some error checking and fact checking and ensure that, that the content itself is right and it’s on brand. We’re not saying something that the brand doesn’t agree with, kind of a thing, you know.

CHRIS HUNTER
7:29

So anyways, yes, 100%, 1,000%, human in the middle at this point, uh, when it comes to anything really with AI.

JOHN CLENDENNING
7:37

And I’ll add a little detail into that because, Chris, you’re the one— you’re the, um, we talked about the master prompt, you put it out first. So you’re saying you’re even pausing, and you understand, and it’s something we’ll train on, um, coming up here in, in, in the, the near future, how to build a master prompt, like 5, 10 pages of your brand voice, your services, everything about your company, your ideal customer avatar, all that kind of stuff. So it already knows all that in your world, and I know it does, and you’re still going, oh no, you’re not posting that out without me looking at it, or whatever, right? So Think of the average person who goes, I don’t even know what a master prompt is. Definitely have the human in the loop for that kind of stuff.

LAURA SOUTHERLY
8:14

What are you thinking, Laura? Well, it comes back to what Chris was saying. And again, this is a topic we bring into the conversation basically every week because of the importance. And when it comes to hiring and having those key people to our team, our employees, anybody that’s helping us, it is true. We may say, here’s our brand guide, here’s X, Y, and Z, and provide them. We may do trainings with them, but it it is stuff that gets forgotten or looked over. And even this week when I was looking over our business e-newsletter that we send out, it was referring to an article that hadn’t been posted for, you know, a few issues back. So yes, even though it was having a conversation, it kept referring to something. I’m like, that’s not even in this issue.

LAURA SOUTHERLY
9:00

So just as simple as that, it made me pause thinking, wait a minute, we’re looking at a current thing, it’s remembering something old. Just like humans do as well. But having me look over it, having us as humans look over any of the AI work does help tremendously. And there, you know, obviously you can find those oopses online a lot where people will call out when someone just printed something and actually went to physical print, like into a magazine without looking at it or you know, checking anything, and it was easily seen, you know, ask me any questions, or, you know, can we finish this article? And they just copy-paste it, done, print it, and that went to full, you know, out there that everybody’s now seeing as a blooper. But human in the loop, very important, no matter what we put in place.

JOHN CLENDENNING
9:55

Yeah, and don’t, don’t go on X and follow politicians or anything like that, because you’ll hear in the comments, people, this was an AI post, I can tell, you got the long ellipses, you got this wrong, you got that wrong, and they’re getting, they’re getting busted for not even being authentic as like, yeah, like a world leader and stuff like that. So imagine, yeah, like it’s a real thing. So, and I love those stories. What’s your thought, Brad, on human in the loop and how clients and teams and just the general?

BRAD KILGORE
10:25

So yeah, so some good things were shared already from our other cohorts here. Is that what you call you? Is that what I call you guys? Are you guys my cohorts? Sure. My partners in crime. So we’ve developed a new system that I think dovetails right off of what you just said, where, you know, we know how important and how popular website chats are now and how AI integrates into those. And, you know, that instant gratification that most customers nowadays are needing, you know, they’re like, Okay, well, I just want to chat with somebody. I need a quick answer. What’s your hours? What’s, what’s, what’s the price for this? Where are you located? Whatever it is. So, so those things are becoming super commonplace now where you have a chatbot or a chat of some sort on your website that can answer basic stuff.

BRAD KILGORE
11:18

It’s trained to do that, right? But what we’ve developed now is something where the human-in-the-loop portion where literally they can chat they can actually have a voice conversation with an AI bot as well. But there’s also a little button that says start a video call with a human. And because that part of it is a lot of times what people are looking for. Think about it this way. Let’s say you’re, you’re, you’re a female. Laura and Jen are out shopping on Rodeo Drive and they’re, they’re out looking at fancy clothes, right? And they may not want to talk to anybody just yet because, as Rodney Dangerfield used to say, I’m just browsing, right? But there comes a point where they need help, right? They need help of, okay, what about this? What is the, what is the thread count on this? What is this, right? So now they need to talk to you. But before that, they really didn’t want to be bothered, right? Well, it’s the same thing with your website. You know, people are going to be browsing, they’re going to be looking at things.

BRAD KILGORE
12:21

And maybe they don’t want any help. Maybe they don’t want to be interrupted or talk to a chatbot or talk to a video per— a person on a video. But that’s really where it’s going. And that’s what we’ve developed is where they literally, whenever they’re ready, we can kind of nudge them and say, hey, I saw you were, you know, I saw you had an interest in this. Did you want to talk? And so literally, they just hit the button. And now, as long as it’s not midnight, you know, they could talk to a human., to, to go face to face. Because in all of our industries, there is that portion, uh, you know, our clients’ industries, whatever it is, there is that, that point where show and tell, so to speak, is important to say, oh well, uh, yeah, I can show you this, or yeah, let’s look at this, let’s hear this, let’s see this. You know, you can’t just do that with AI, but with a human you can.

JOHN CLENDENNING
13:11

Yeah. And I’ll, I’ll put a pin in at the end of this. I fully agree with you. In my world, I like, I like the average carpet cleaner to have a 5 to $800 residential job. If you’re listening to this, carpet cleaners, and you just— your jaw just dropped, I can show you how. But you can’t do that if the bot is trying to self-book, because how do you influence a person who thinks carpet cleaning is 3 rooms in a hall for $99 that it isn’t? And how can you run influencing? Well, that’s scripting. That’s all the stuff that we know how to do to sell a better product. There’s a reason why people are willing to pay $10, $12, $15 for a Fufu Lapa Lapa coffee at Starbucks, but they they pay $1.50 at Dunkin’ Donuts and think it’s a ripoff, right?

LAURA SOUTHERLY
13:51

What kind of coffee was that?

JOHN CLENDENNING
13:55

What kind of coffee was that? I am not that person that goes to Starbucks. So yeah, you guys can fill in the blanks. But another one is like, so my story was I talked to a client recently who took over his, decided to take over his own local service ads. I said, well, you’re— you go out in the field, you have a role to play that can generate a $10,000, $20,000, $30,000 a month carpet cleaning company worth of efforts if you follow a proven path. Local service ads, there are experts that can do that. Well, no, I just chat GPT. I got it all figured out. I’m stuck on a few things here and my, you know, this happened.

JOHN CLENDENNING
14:32

Okay, cool. So what did you do about image optimization? Oh, I put images in. No, no, no. How did you optimize them? To actually sell at a higher price than your competitors? What have you done? And I asked, I went down a list of 3 or 4 or 5 advanced strategies that we do because we’ve been doing it for years now and we know how to optimize for the highest-end clientele to not get tire kicker calls. He goes, you know the answer to that, John. I’m going, yeah. So why don’t you just hand that back over to us or hand that over to us? He hadn’t. And do these things I’m telling you that will land you $10,000 worth of work in the next 30 days.

JOHN CLENDENNING
15:09

Because again, there’s the human in the loop and there’s also the expert that knows how to maybe do a service where ChatGPT— like, I would never go to ChatGPT and say, you know, whatever, I’ve got this ailment, this affliction, you know, what do I do? Oh, drink bleach. Okay, you know, whatever. Like, there might be a point at which you, you get some advice, go to your doctor. I’ve got this, what do you think, right? So again, just, just understand that there are still experts that need to finalize the decisions, like Chris said, like Brad said. Um, but yeah, I just thought that’s a great headline that, you know, industry leaders are stressing this as well. Guys that are— guys and gals that are building billion-dollar companies using AI and automation going, whoa, but don’t let it just run rogue, right?

LAURA SOUTHERLY
15:53

Not at all. Have those checks and balances in there and stop. And even if it’s just having a quarterly check, start again small. But if, as you feel like, okay, things are running with less checks between your teams checking it. Maybe you as the business owner gets to just check in once in a while now, but definitely let those on your teams know who has what responsibilities to do that final check before things go live, before things continue posting. There’s a lot of benefit and it keeps you in with the conversation too. So especially we get some questions like, how can I just have, you know, AI automatically respond to posts, etc.? Well, that takes you out of that conversation that’s going on. If you know what questions they’re asking online, guess what? You’re going to be the expert they look at to say, please answer this for me.

LAURA SOUTHERLY
16:43

So I encourage them to stay especially involved in that. They may have somebody else on their team that does it daily. But hey, wrap up the report. Let me know what were the top questions being asked, because again, that gives you a way to get into those conversations.

JOHN CLENDENNING
16:58

Yeah. And you said something great. Wrap up that report and let me stay in. I’m going to go back to Chris for a second because one of the things he shared with me that was amazing that we still do manually that we don’t do manually starting this week, um, is at the end of an EOS, Entrepreneur Operating System, running your L10 meetings. And Chris is now— I was like, of course, once you get into the— once you get, you know, into the beast, all of a sudden you’re like, oh my God, this and this and this. And it’s like, yeah, so wrap up the report. Maybe don’t spend 3 hours writing the report. Maybe take the transcript of the meeting and have a really well-trained prompt, right?

CHRIS HUNTER
17:33

And so mention that a little bit, Chris, because like you do this better than most. So, um, I shared with y’all, right? Uh, some of my prompts that I’ve, I’ve got for my shortcuts for, um, Perplexity Comment, right? And so one of those, like what you talked about, was after, after every meeting, uh, and, and I shared this with, with my account management staff as well to help speed them up. Okay, is essentially open up Comet. I’ve got it set up as a shortcut. You could just copy and paste this prompt if you want to, but basically you’re telling it, go into OneNote because OneNote is what we use for all of our notes and pull out all of the, the action items out of our notes that we took during the meeting. Okay, now put those into ClickUp, which is the tool that we use for our project management. And assign those out, right, with a date due of next week with, uh, whoever it’s assigned to and, and the urgency set to high. Okay.

CHRIS HUNTER
18:41

Um, so it, it speeds you up. You can still look at those and go, not quite, and maybe it should be this, but that’s like the 20% of the, of the stuff that you have to do versus having to, uh, because it was just a, it was a a horrible task. Honestly, I hated it. And I’m, I’m, when it comes to this kind of stuff, I’m like, what is the stuff that I hate doing? That’s what I’m going to automate as much as I possibly can, right? I’m going to spend 2 hours trying to figure out how to automate a 5-minute task, but I’m still going to do it because I hate doing it. And so therefore it gets put off, right? And so that’s how I approach that kind of stuff. But coming full circle back to that human in the middle, you still have to have that in all sorts of things at this point, like I’ve already said.

JOHN CLENDENNING
19:28

So I think the best way to think of that as well— I’m just gonna put a point in this and then we’ll move on— but is, um, so Dan Martell, if we think in frameworks, frameworks really help us identify, do the same things over and over again in a great way to teach our team, like your employees, how— here’s the framework of how we’re going to arrive on a job and do a roof, here’s the framework, you know, frameworks, everything’s a framework. So now you can visualize it, see it, train on it, stuff like that. Dan Martell’s got a bunch of really cool frameworks in Buy Back Your Time. Um, 10-80-10 is a great one. 10%, identify the, the, the task at hand, what it is. So in the AI world, write a great instruction. That’s what Chris just said. 80%, let, let comment browser turn blue and go, go do stuff, other things on your computer.

JOHN CLENDENNING
20:10

It’s working in the background. Let it do one step at a time as an agent, right? It’s, it’s running agentically. And then have it pause and not do or not post or not finalize, leave it in draft mode, and you come back and do the, the other 10. So whether that’s 10 employee, I want this done, come back to me when you’re ready with this entire new brand kit, this thing, whatever, um, this new marketing, whatever it is, think of your— and then at the, the 10 on the end, I will see what you did, reiterate, readjust, and then we’ll deploy. Well, That’s the same framework, just put an agent in the middle where it can. Same thought process. So cool. So yeah, so I think again, great headline, great thought process.

JOHN CLENDENNING
20:52

Hopefully we help people start thinking through a whole bunch of really cool ideas. Um, speaking of cool ideas, we’re at the segment of the show where we share some cool ideas. Um, and I know Brad’s got something he wants to share. I don’t even think he told us what it was, but he said it’d take 2 or 3, 4 minutes, so What do you got for us, Brad?

BRAD KILGORE
21:11

Well, it’s top secret, but I’m about to reveal it because you guys don’t even know how cool this is, right? So we talked a little bit about obviously human in the loop, but also time-saving and systematizing things, right? And so whether you’re using ChatGPT or Gemini or any of the other well-known ones, this tool, this is very new to me, is pretty powerful. So I’m going to share screen. Okay. So what this, what this does is it’s an extension and it’s a paid service. It’s not very expensive. It’s called PrettyPrompt. All right. So when you’re in ChatGPT, I know myself and several people on my team, my wife, whoever it is, struggle with how to put a good prompt in there, right? So I made it as super simple as possible so that I could emphasize how this works.

BRAD KILGORE
22:08

So, you know, whether you’re creating images, whether you’re creating blog posts, whether you’re doing any number of things in your LLM, now with this little extension here, I can say, okay, make me a nice photo of birds in Colorado. This is super generic, right? I mean, ChatGPT could probably do something with it normally, but it’s kind of lost, right? But with this, I could say improve the prompt, refine the prompt, or go to my prompt library, right? So for my team who struggle, I can have a prompt library they can pick from for repetitive tasks, right? But if I go to improve the prompt, it’s going to quickly take what I have there, right? And it’s going to make me a really nice prompt., right? And so it’s saying, you know, you’re an expert nature wildlife photographer specializing in blah, blah, blah. The image must be a nice photo. It’s going to talk about this, right? So it can create it for you. And then of course, you would just apply it and then create what you’re going to create. Or you could hit the refine button here where now it’s, you know, giving you, okay, I’m in Colorado. And so what specific type of bird are we even going to talk about, right? So I’m going to say birds of prey. And then I’m going to say what kind of setting.

BRAD KILGORE
23:32

I live in the forest. So that sounds cool. But this one sounds cool. But I like near the river. You know, what artistic? I want it to be hyper-realistic sharp focus, right? And then submit my answers., and then away it goes, right? So this is just, to me, a time saver where now, again, it’s got my refined prompt in there, which if I had to figure that out, I might as well have went to a stock photo place or something to try to just buy a photo by the time it would have taken me to write a prompt this detailed, right? I just hit apply, boom, done.

LAURA SOUTHERLY
24:09

I’ve got what I need. Brad, can you go ahead and hit apply? Because I think some of our Viewers and even the listeners might be unfamiliar with one of the new tools as well is when it’s creating an image. Now you’re able to— once you see the image, you’re actually able to use the Acrobat, Adobe Acrobat to refine your image. Now it’s limited and it’s very much limited. Like you can remove the background, it’s rough and everything, but that is an upgrade that I’ve been seeing. That surprised me. So as you’re working with something like this and you have an image created by ChatGPT, you’d be able to refine it now.

BRAD KILGORE
24:50

Yep, you sure could. Yep, there’s a lot of things you can do with screenshots or other images. You can mark up screenshots these days of what you’re trying to explain to AI and then put it in there.

JOHN CLENDENNING
25:02

But yeah, this— Yeah, if you go back a year ago and you went into Flux, so I went down this rabbit hole to try and get better pictures, more lifelike to me before Nanobanana came out and all that stuff.. And I, so I went down some rabbit holes on Reddit and other places like that. And the guys that were getting guys and gals that were getting really great images, they were talking about the camera aperture and their, what kind of camera. And like, it was this whole, and I’m going, I just typed in, I want a picture of me looking like Santa Claus. And they’re going, oh, heck no. You, and you had to learn all of what LORA determined. Well, LORA, what the heck’s a LORA? It’s like, I thought that was a friend of mine down the street, right? So that we had coffee last week or so. Anyways.

JOHN CLENDENNING
25:44

I realized that I had no clue. Brad has just realized that I am now brilliant because I don’t need to have a clue. And it will help for— and then if it does, it turns out wrong, you can, like, again, any AI, I assume you could probably take it back in and go, well, that wasn’t quite right. This is what I got. How do we fix it?

LAURA SOUTHERLY
26:02

Right? Refine it.

JOHN CLENDENNING
26:02

Right?

LAURA SOUTHERLY
26:03

Exactly.

JOHN CLENDENNING
26:03

So cool.

CHRIS HUNTER
26:04

That was a great share.

LAURA SOUTHERLY
26:06

Yeah.

CHRIS HUNTER
26:06

Great share. Thank you.

BRAD KILGORE
26:08

Does that tool work with multiple LLMs or is it only ChatGPT? ChatGPT, Gemini, Claude. From what I understand, it can now integrate into things like Lovable when you’re doing vibe coding. It’s insane. And what I’m seeing as well is this particular software is low cost, but there’s a lot of investment into it and it’s growing fast.

JOHN CLENDENNING
26:34

What, John?

BRAD KILGORE
26:35

What is the cost?

JOHN CLENDENNING
26:36

I got a lifetime deal on it because you know me. Of course you did. So anybody wants to just ping Brad privately about how to get lifetime deals on things. He is my go-to on lifetime deals as well, just so we all know. So I got one share. I got one share I want to share as well because I do a lot of two things. One, do a lot of thinking when I’m talking. Turns out it’s my style.

JOHN CLENDENNING
27:00

Learned that over doing a personality thing over the Christmas break. But I do talk and brainstorm and I do a lot of thinking and ideation when I’m driving. I’ll listen, I’ll put a pod, like a YouTube podcast. I’m not watching it or whatever, learn something. And I’m like, oh, that’s great. ’cause I was used to like trying to write it down on a notepad in front of me and driving into traffic and all that kind of stuff. So yeah, I usually drive alone as my family says. So, but now I’ve got this down to a science where on my phone, I’ll show it to you, I’ll share my screen, but on my phone I’ve got Perplexity and then I’ve got my comment browser or Perplexity.

JOHN CLENDENNING
27:37

Perplexity is cool because it can read YouTube transcripts. So the cool thing about that is you don’t have to like, where ChatGPT is not quite, can’t do that yet. Perplexity can.

LAURA SOUTHERLY
27:51

So if I go to entire screen and it will be— While John’s doing that, it’s $5 to $10 a month for a pretty prompt. Oh, that’s nothing. Oh, nice.

JOHN CLENDENNING
28:00

Is it a Chrome extension, Brad?

BRAD KILGORE
28:03

Yes.

JOHN CLENDENNING
28:03

Oh, that’s brilliant. But you can use this— We’ll put a link in the show notes. I can’t guarantee it’s not an affiliate link if Brad’s got an affiliate link, but that means you pay the same price and Brad makes some money. But we’ll talk about that on another episode. So here, for example, today, just before we jumped on. Um, I watched this YouTube video, so I’ll just click on this. So again, it’s— so it was on my phone, just my iPhone. Um, I watched the YouTube video, copied the link, moved it into Perplexity, right? Um, and so now it’s sitting in my library.

JOHN CLENDENNING
28:33

I said here, here’s the YouTube link. Uh, I’m creating a, a custom, um, GPT on project to train. It’s for some communication coaching and things like that, better webinars, things like that. Um, all this kind of stuff, right? And the guy’s name apparently is Joseph Zarr. I’d never heard this guy before, and I just told it I would like a summary of this and a document that I can add because again, I, I use multiple LLMs, but over in ChatGPT, I have my communication coaching project, right? So I started a project. It’s trained on, um, a book called The Coaching Habit by Michael Mungay, um, is, uh, Surrounded by Idiots. Lovely book, actually. Um, so these are the PDFs of the books that I bought.

JOHN CLENDENNING
29:21

And then, um, Crucial, um, Confrontations. And then my own personality assessment and stuff like that. So inside this project, I can have any conversations I want about different things. Create me a training program. Let’s have a conversation back and forth. Um, give me a hyper, you know, a scenario, a confrontation. Let’s banter it out because, you know, literally just talk to my my gal Sam is what I call her. But now I, because of that call, I can now take this document right here.

JOHN CLENDENNING
29:49

This is sort of it summarizing it and giving me the frameworks that it talked about, the power arc, the give-getting scale, and all of that kind of stuff, and other links to this communication coach. The only thing I’m gonna do right now is the next prompt I will give it will be, and I won’t type it, I’m just gonna speak it and just say, can you put that into markdown language in a code frame so I can easily copy it because then you have a little copy button up there and paste it into my communication log, or paste it in as another file in for the whole entire, um, um, the whole entire project. So now this project is all about just this thing. I’ve got a project like, you know, with my finances and my life and my all this stuff. I’ve got one about our messaging. And so this project is now just about communication. So I can, when I go to do a presentation, I can say, hey, I’m going to do this presentation. How would I make this point more clearly? Right? Give me some ideas, stuff like that.

JOHN CLENDENNING
30:49

I can triage after I’ve done something that, that didn’t land well. Hey, this didn’t land well. Here’s my transcript based on all of this. What should I have done? Right? And all of that is so the two, two, two hacks in one there while you’re out and about, you see something, you hear something on your phone, just have Perplexity, pop it in. Next time you go onto your computer, it’s sitting in your library. Just have the same account and now you can do it. You can use just all the way through Perplexity and use it in a shortcut, which is pretty much similar, like, and Spaces, sorry, Spaces are like projects, or you can bounce between tools. And now you’re just being way more effective and without having to touch the keyboard or type a word.

BRAD KILGORE
31:32

It’s all done by talking. Wait a minute, you’re doing all this at 80 miles an hour, John?

JOHN CLENDENNING
31:37

I think that this is illegal.

BRAD KILGORE
31:39

All this. Well, I was actually going 120 kilometers an hour.

JOHN CLENDENNING
31:42

I’m Canadian. So you’re in Canada, so it’s kilometers. Yeah. Anyways, so I grab my McDonald’s coffee, do a loop around the countryside, listen to something, give it some thought. That’s my lunch break. And I usually come back with ideas that I can’t enact right away. And I, it was notes all over my desk. Now it’s a framework for the next time I get to it.

JOHN CLENDENNING
32:05

That’s fantastic. Love that. Cool. Yeah. So I think we, I think we’ve, like we always do, drinking from a fire hydrant. I think we’ve left our audience with a lot to think about and, and stuff like that. Any famous last words from, from the panel on, uh, what’s coming up?

LAURA SOUTHERLY
32:20

Any thoughts on, or anything like that before we wrap her up? I will bring to the table next week’s Google LM that it just elevated my slides like no tomorrow. I was really impressed with, you know, that’s where I get bogged down when I’m doing these talks in the presentations that I’ve been out doing some speaking on AI and what they have put in place allows you if you just put a framework around what you want in your presentations, it now has a very uniformed look where before when I demonstrated it was one slide at a time in Google Slides. Now bringing it up through Google LM, the notebook, it’s amazing.

JOHN CLENDENNING
33:02

It’s just amazing for Notebook LM. Yeah, well, I’m excited to see that. So hopefully everybody tunes in next, next week for that one. And with that said, make sure you like, follow, subscribe, tell your friends about it, share it out. We’re here to help you understand this world as it rapidly changes. In the comments, anything you want to hear about, any advice, positive, negative, anything like that. Keep an eye on the description for, for Brad’s link. And, uh, with that said, uh, we bid you a fond adieu till next week, and, uh, happy marketing.

00:00 Welcome + why this show exists

01:46 The adoption curve & CES: automation everywhere (but brand voice needs humans)

03:19 Human-in-the-loop: protect messaging before anything goes live

10:59 Chat to voice/video: the “talk to a human” button that increases conversion

13:16 From $99 specials to $800 jobs: offers, scripting, and positioning

17:27 Prompts + end-of-meeting wrap-ups: your new ops habit

18:08 Perplexity + OneNote to ClickUp: automatic action items (you approve)

20:06 10-80-10: agent in the middle, you on quality control

21:16 Pretty Prompt demo: better prompts, faster outputs

30:30 Drive-time learning to frameworks you can reuse

AI is moving fast this AI news cycle changes weekly but AI marketing in 2026 still needs a Human in the Loop if you want consistent brand voice, fewer mistakes, and better results.

In today’s episode, we break down a practical, copy-today game plan for using ai in marketing without letting automation ship bad outputs. This is the kind of ai video you can rewatch and turn into a repeatable workflow.

You’ll learn why “set it and forget it” creates ai slop, how to build a Master Prompt that protects your messaging, and how to use AI for execution while keeping a human as the final quality gate.

What you’ll learn (fast):

The Human-in-the-Loop rule: a simple approval system that prevents AI from publishing brand-damaging mistakes.

Real-world AI fail (and fix): how AI can misidentify competitors in an SEO workflow and how to catch it every time.

Master Prompt framework: encode your brand voice, offers, ICP, and constraints so outputs stay on-message.

Website conversions: use AI chat/voice to capture leads and add a “talk to a human now” handoff (even after hours).

Local-service reality check: why high-ticket selling still needs human judgment, positioning, and trust.

Research shortcut: transcript-driven research workflows that turn insights into reusable playbooks.

Claude AI

Grok AI

Meta AI,

Kling AI

HeyGen AI

Manus AI

Higgsfield AI (and a Higgsfield AI tutorial)

Google AI Pro

SI AI

Share the Post:

Related Posts

BE IN THE KNOW ABOUT THE LATEST IN AI

Join thousands of forward-thinkers getting the best AI news, guides, and updates delivered straight to their inbox.

Quick Links

RESOURCES

Shopping cart0
There are no products in the cart!
Continue shopping