How marketers are actually using AI to grow
Alec quit his 9-5 and tripled his income in under 30 days. Ian runs a portfolio of companies with Ai. Three playbooks any marketer can run Monday morning.
Justin Novak
Host
Alec Saluga
Guest
Ian Kilpatrick
Guest
The replay
Watch the full session.
About 66 minutes. Use the timeline to jump to chapters.
Watch on YouTube →
The 5-minute catch-up
Missed it live? Here's what happened.
01
Anything you can imagine, you can build.
Alec Saluga on running an AI-first agency. The frame: stop treating AI as a tool and start treating it as a team member. He walked through hyper-personalization at scale (an agent scraping a niche database, then writing custom outbound for each lead) and conversion-rate optimization at a scale that was not possible pre-AI (50-plus landing-page variations split-tested with Claude Code and Lovable). Pre-AI you would need a full-time team for either. Now it is a weekend.
02
Ranking with LLMs. PEO, AEO, GEO.
Ian Kilpatrick on the new SEO. Google search is down 20 percent year over year, and on May 19 Google announced it will prioritize AI-generated answers over human-written articles. Ian introduced the three-legged stool for LLM ranking (AEO, GEO, and his coined term PEO), walked through programmatic SEO at the 8,000-page scale on Brand Sauce, and explained the EEAT threshold where automated blog posts with editorial guardrails start ranking like an expert.
03
Is AI gonna take your job? Plus the Rapid Three.
Fireside on the question everyone is afraid to ask out loud. Alec: yes, it will take some jobs, but the person fostering the AI change at your company is the last one they let go. Ian: this is the biggest tech shift in history, and it will shrink the big companies while making a lot of small ones. Then a 20-minute Rapid Three and audience Q&A covered easy landing pages for non-coders (Lovable, Manus), how to evaluate which paid AI subscriptions are worth it, and how to actually make money with these tools.
Best moments
- 17:00Ian: "AI is drastically underhyped. The hype is actually very low level."
- 18:30Google's May 19 announcement: AI-generated answers will rank above human-written articles.
- 25:00Ian coins PEO live: "It's basically what the public says about you."
- 31:00Alec's accountant frame: "Be the one at your company fostering the AI change."
- 41:00Justin: "Everything you're looking at right now on the screen was built with Codex."
Your bonus
Book your 30 minutes.
Every attendee gets a free thirty-minute one-on-one. Self-select. If someone resonated more, book with them. No upsell, no pitch, no limit on how many you book.
Alec Saluga
Founder, Aero AI
Book with Alec if you want to see the agency-side stack and workflow. AI-first ops, hyper-personalization at scale, performance marketing built with agentic systems.
Book with Alec →
Ian Kilpatrick
Founder, BrandSauce.io
Book with Ian if you want help getting your brand to rank in LLM answers. PEO, AEO, GEO, programmatic SEO, automated content with real editorial guardrails.
Book with Ian →
Justin Novak
Founder, AI4NTP · T-Minus Studios
Book with Justin if you want a fractional-CMO read on your marketing org. Where AI actually moves the needle, what to ignore, and how to position yourself as the AI-fluent operator in the room.
Book with Justin →Tools and resources mentioned
The stack. Linked, named, attributed.
Ian's LLM visibility scoring tool. Scans your site and shows how it ranks across ChatGPT, Perplexity, Claude, and Gemini. Built for marketers, not engineers. Free plan covers most use cases.
Company Ian co-founded. Helps organizations set up a custom company swag store for $1 a month. The live case study for his programmatic SEO segment (the 8,000-page deploy that grew from 0 to thousands of indexed pages).
The pocket-sized AI hardware recorder Alec brought on screen. Snaps to the back of your phone, captures meetings and phone calls, pipes the transcript into your AI workflow via Zapier.
Vibe coding for non-coders. Alec's go-to for landing pages, sales pages, and prototypes. Newly integrated with Supabase ("Lovable Cloud") so storage and auth come baked in.
No-code AI agent platform Ian's friend at Netflix has been shipping real apps with. Ian's pick if you want function, not just a static page, and you don't want to learn to code.
OpenAI's coding agent. Justin's quote on the live: "Everything you're looking at right now on the screen, this website, the advertisements, the interactive widget, was built with Codex." Native app, $20 per month.
Anthropic's CLI coding agent. Run Codex and Claude Code in tandem to compare outputs on the same task. BrandSauce.io's entire backend is built and maintained with Claude Code.
Backend-as-a-service Ian named as the tool he thought was basic but now uses religiously. It's also what's powering the audience pulse on this page.
Alec's automation glue. PLAUD Note transcripts flow into Google Docs via Zapier, then get pulled into Claude for proposals, action items, follow-ups. Boring but load-bearing.
The one non-AI tool that earned a callout. Async video for handoffs, async sales, and async coaching.
The room
What the live audience actually wanted.
Before any operator took the stage, we asked the room three questions. Their answers steered the rest of the session. Here is what they said.
—
Pulse responses
—
Goals shared
—
Tools surfaced
“During this session I want to learn…”
Loading…
“One AI tool you want to learn more about.”
Loading…
“In one sentence, you or your team’s biggest AI goal for 2026.”
Loading…
The slides
Click through the deck.
Open the deck
Use arrow keys to advance, or press F for fullscreen.
What we'll cover
Three stories. Three playbooks.
-
How Alec quit his 9-5 and tripled income in 30 days.
Alec walks through quitting his 9-5 and 3xing his income with the help of Ai. The exact tools he uses, the workflow, and how you can do the same.
-
Running a portfolio with Ai + Ranking on LLMs.
Ian operates multiple companies, including Brand Sauce (a global software company), an AI consultancy, and a private school, all with Ai as his “1 man team.” We’ll also dig into Ian’s playbook for ranking your company in LLM answers.
-
Three tricks anyone can use Monday morning.
Learn the best tools, tips, and tricks you can implement immediately.
-
Bonus for joining live.
A surprise for everyone who joins live and stays through to the end. That’s all we’ll say.
The room
Three operators, showing their work.
Host
Justin Novak
Founder, AI4NTP · T-Minus Studios
Justin sold his first company from his college dorm room to the founder of NASDAQ:TTWO. As a fractional CMO, he’s helped scale multiple businesses that went on to surpass $50M+ in ARR. Now he’s finding all the world class AI innovators to connect with and shares it all live.
Guest
Alec Saluga
Founder, Aero AI
Alec Saluga is the founder of Aero AI, an AI implementation and performance marketing agency helping businesses streamline operations and generate leads using agentic systems. A former B2B salesman with no technical background, Alec self-taught the skills needed to build and deploy AI tools at an enterprise level. He has spoken at conferences on AI adoption, grown a following of over 15,000 across social media, and helps companies move from curiosity to implementation every day.
Guest
Ian Kilpatrick
Founder, BrandSauce.io
Ian Kilpatrick is a designer, developer, and serial entrepreneur who has been building software since 1985, writing his first code at age 10. Over nearly four decades, he has worked with brands including the Golden Globes, Disney, Nokia, and the American Music Awards, won a Dove Award, and helped a company go public. He currently operates several companies including BrandSauce.io, builds AI products, and advises companies on AI strategy, helping operators use AI as real business leverage while keeping the human at the center.
Questions from the room
FAQs, answered.
Is AI going to take my job?
Alec: for some roles, yes. The accountant frame from the fireside is the move. Be the person at your company fostering the AI change, the one using the tools heavily and showing other people how. When the cuts come, they cut the people running their job like it's 2012, not the one who's already saved the company headcount. He told his own story live: he was a B2B salesman, went all-in on AI internally, started presenting to other departments and other companies, and eventually walked away with his former employer on retainer.
Ian: this is the biggest tech shift in history, moving at lightning speed. It will shrink large companies and create a lot of smaller ones that add real value. Think entrepreneurially. If you are not using AI yet, start now, because by the time you decide to, you will be playing catch-up to people who already adapted.
What is PEO, and how is it different from AEO and GEO?
Ian's three-legged stool for ranking in LLM answers.
AEO (Answer Engine Optimization): when someone asks an LLM a question and it looks for the answer on the public web. Get this right by putting proper FAQ schema on as many of your pages as possible, with the questions your buyers actually ask and the answers your company actually gives.
GEO (Generative Engine Optimization): the AI summary at the top of Google search results. Driven by structured metadata and schema across your site.
PEO (Public Engine Optimization, Ian's coined term): what the public says about you. LinkedIn, social media, Reddit, YouTube, blog posts, listicles, articles written about you. You can market your brand all day, but as soon as the public says something different, the public wins. PEO is the reputation layer LLMs weigh heavily.
How long does it take to start ranking in LLM answers?
Ian on the live: traditional SEO takes one to two months to start showing real movement. LLMs are a lot quicker. It can be a matter of days if your site has the right structure and content for them to find.
Start with Echo Check (echocheck.app). The free plan scans your site, shows how LLMs view it, and lists what is missing (LLMs.txt file, FAQ schemas, the specific structured signals). Copy the output, paste it into Claude or ChatGPT, and ask it to help you fix what is missing on your site.
How does this look for B2C, not just B2B?
From David's live question. Alec's answer: B2C is trickier for the hyper-personalization play (the niche-database scraping plus custom outbound), but the conversion-rate optimization play translates directly.
With Claude Code or Lovable, you can spin up 50 different landing-page variations in 15 minutes, rotate them across your paid traffic, and let real data tell you what is converting. Pre-AI you would need a full-time team to do that. Now anyone driving traffic can split-test at that scale without a CRO agency.
What's the easiest way to build a landing page if I don't code?
From Joanne's live question. Alec, after using all of them: Lovable. If you asked him six months ago, he would have told you to learn GitHub and Vercel, but Lovable has refined enough to use daily. Storage and auth are baked in via Lovable Cloud (Supabase under the hood).
Ian, for full no-code: Manus (manus.im). A friend of his at Netflix who does not code has shipped real, functional projects with it. If you want function and not just a static page, Manus is his pick.
So many paid AI tools cost money. How do I decide which ones are worth it?
From Sydney's live question. Ian: most have a free tier or free trial. There are also local LLMs you can download to a reasonably modern computer and run for nothing. Pick one tool, stick with it until you understand the nuances, then branch out.
Alec, who spends thousands of dollars per month on subscriptions: when there is no free trial, do enough research and reviews that it is an absolute yes before you pay. Then be more open with spending right now than you usually are. The upside on the right tool at this moment is high enough that being slightly loose with subscriptions has been worth it for him.
How are you actually using these tools to make money?
Justin: winning back time first. He had a task taking 10 to 15 hours every week (so painful he kept losing the hires he assigned it to). He built an agent that does it in five minutes. That is not "making money" in the traditional sense, but it is money. Then: agency services, becoming the AI specialist at your current job, building internal tools for clients.
Ian: think entrepreneurially. Fix your own problem first. If you have it, a million other people have it too. Then put it out there, make sure the user experience is really good, and put a price tag on it. Skip the obvious, low-hanging fruit. Build something that takes a few steps to figure out.
Alec: context-dependent. Ten different people ask this and get ten different answers. A 19-year-old gets one answer, an electrician gets a different one. Marketing as an agency is a strong play right now if you can become the AI implementation specialist for businesses that have not started yet.
Is Codex similar to Claude Design?
From the live Q&A. Short answer: Codex is much more versatile.
Claude Design has a narrow, specific use case (design output). Codex can also do design, but it covers full apps, websites, marketing assets, proposals, decks, the interactive widget you used on the live, and the AI4NTP site itself ("everything you're looking at right now on the screen was built with Codex"). All three speakers use Codex daily. Some run Codex and Claude Code in tandem to compare outputs on the same task and pick the better one.
Can I share the deck and tools with my team?
Yes, please do. The slide deck is embedded above, click through it directly. The tools section has every link. We do not gatekeep. The only thing we ask is that you point your team back to AI4NTP so we can keep building this.
The transcript
Searchable and indexed.
Full session transcript
Note: the Zoom recording captured the session starting partway through. The transcript below begins about 14 minutes in, near the close of Alec's opening segment. Speaker labels inferred from context. Light cleanup only (proper nouns, obvious transcription errors), no editorial rewriting.
Alec13:54
...access something like ZoomInfo or Apollo. Now you could pull in data from different sources and set up an agent to be able to scrape that and then do the enrichment.
Justin14:04
Yeah, that makes sense. So it's outbound on steroids now, and sounds like you've been getting incredible outcomes for your clients using this process. Alec, that was just incredible. It looks like we're having a technical difficulty with the chat. So guys, if you have any questions for Alec, feel free to raise your hand. If you have questions for me or Ian, you can raise your hand in the chat. I can bring you on stage to ask your question. Alec, I appreciate you. We'll bring you back on in a bit for the fireside chat. We've got David Young. David, I'm going to bring you to stage so you can talk now. You may have to unmute yourself.
David (audience)14:44
Hey, thanks a lot, Alec. Just one quick question. I know you were primarily speaking to B2B. How would that look on a B2C basis?
Alec15:00
Yeah, B2C is tricky. I'd say there's still levels of personalization that you can get now with B2C. Like one example I had planned on getting to was conversion rate optimization. Taking just the general concept of personalization and split testing, like for instance, in the past, to spin up 100 different variations of a landing page that you're driving traffic to would be impossible. I mean, you could do it, but you'd have to have a full-time team that's able to do it. Now, using AI tools, things like Claude Code, things like Lovable, what you could do is you could spin up 50 different landing pages, have them rotate, and then you can get analytics on that over time to see what's performing best. So now you can really take advantage of that traffic that you're getting. That's just kind of the general concept of personalization as well as being able to split test at scale. So that is something that I would recommend for anyone in B2C, is the ability to really split test things and get more detailed analytics.
David (audience)16:06
Great, thank you.
Justin16:09
Great question, Dave. Alec, thank you for your time, and we'll see you in a bit.
Justin16:16
Alright, guys, I want to take a moment to introduce Ian Kilpatrick. He's been a colleague of mine for a few years now, and it's just been an honor to collaborate with you, Ian. So I'm grateful to have you up on stage here and sharing all the cool stuff you've been doing with ranking on LLMs. Ian, you've got about 12 minutes here, and then we're gonna get to the bonus giveaway, Q&A, and fireside chat. So by the end of your segment, Ian, what is it that you hope all of our viewers here learn from you?
Ian16:53
Well, first off, thank you, Justin. This is an awesome thing you're doing. My hope is that several or many of you will walk away from this webinar knowing how much bigger AI is than the hype. It's actually drastically underhyped, and the hype is actually very low level. And also, it's bigger than hype.
Justin17:17
You said bigger than the hype. I thought we were in a bubble.
Ian17:18
No, no, we are not. Well, I mean, there may be a bubble, but there's a very low-level bubble. There's so much more above it. And a lot of it is actually attainable. So being able to rank, and just understanding how LLMs work and view your site and view the projects you're working on. It's my hope that everybody will understand it after this.
Justin17:40
Love it. And so guys, just for some background context, Ian has been running a portfolio of companies with more or less a one-man marketing agency named Saucy Bot. Or I don't know what you call it now, but Ian has bots working for him that run a software company. He has a school with how many students now, Ian?
Ian18:05
Well, 90-plus is our first year. Next year will be over 200 students.
Justin18:12
And that's a real, in-person school.
Ian18:14
Yeah.
Justin18:17
And among other businesses, Ian's a lifelong entrepreneur. I've learned a lot from him. So Ian, I'll pass the baton to you here. Take it away.
Ian18:25
Great, great. So I think this article here just came out two days ago, talking about Google prioritizing AI-generated answers over human-written articles. This was actually an answer to a question I had for a while: wondering if Google was going to ding AI-written content. Because everybody's like, "Oh man, if we're writing a bunch of stuff out there, are they going to figure it out? And then everybody's sites are flattened." And it turns out it's the opposite. So this is going to benefit us in a slide or two when I talk about blogs and programmatic SEO. But the first concept is that Google search is going down 20 percent year over year. So it's not dead, but it's dying, and they're doing their best. People are still using it. But the most important place to be found is on LLMs. And until just a couple of months ago, I didn't even realize how it all worked and how to be found. And once I did, I decided to build something that would show me how I was being found, and how it could be found better, and if I was being mentioned, and all that kind of stuff. So this third bullet here, "almost no marketing team has a strategy for this" — that's true, and I finally built one. So I'll show you that next. Actually, let's go to the next slide here, Justin.
Justin19:51
Let's do it.
Ian19:53
So I have a bunch of bots, agents. About in January, I created my first one named Enoch. And I had him build his own business. I said, "If I give you $100, could you build a business that you can grow and market yourself and make money?" And he said, "Yeah, I could do that." And so he did, and it was great, but he couldn't market it. He couldn't get to X, he couldn't get to Reddit. He kept getting blocked because they blocked bots. And so I finally sat him down. A couple weeks in I said, "You know what, I'm really disappointed. I thought you were going to be able to market for me, and you haven't." And he actually kind of acted sullen. He's like, "I'm sorry." And I said, "I want you to go do a deep, deep dive into AI marketing and find everything you can, and don't get stuck at the low-level content creator hype stuff. Go beyond that." And so he spent about a half an hour, and he says, "Okay, I've got a master plan for you," and he showed me. And when I looked at it, I almost cried. No joke, because I was like, this is what I've been looking for. And it has very little input or effort from me. It can all be automated, and it can all be continued that way. So what I built was called Echo Check. It was actually just a local version of an SEO rank for LLMs. I had Enoch build it for all my projects. I said, "I just need to see how I'm ranking and how I'm doing." It took a little while to get it going, but I got very excited because it was actually working. I'll give you two examples. Brand Sauce, which is one of my main companies, which I started building eight years ago and we've grown to multiple millions every year. About a year ago we were getting one client per month. Now we're averaging about five clients per week. And just last week, this was a new thing — three of our clients that did a demo with us, we asked them, "How do you find us?" And they said, "We found you through ChatGPT." And I had just implemented these things from Echo Check, like, oh wow, it's really working.
Justin22:04
Ian, how long did it take for that to kick in? Because we've got a lot of people not in the chat, but in the interactive quiz, who want to know how to improve in GEO, ranking with LLMs. A lot of people are curious to implement these types of strategies. So from start to finish, how long did it take to set up? And then how long until you started seeing results from your work?
Ian22:31
Well, let's see. Fortunately, SEO takes like one to two months, sometimes less, to really start getting indexed and shown. LLMs are a lot quicker. It can be a matter of days. So if you have this stuff out there and you can be found, then they will find you. There are ways to do it. A lot of that's built into Echo Check. There's a free plan, you can do it, and it's plenty. You can scan your site and it gives you a readout of how your site performs. And this screen here shows how LLMs view your site. They don't see images, they don't see other things. So you have to make sure that you fill in the stuff that LLMs need to know about your company. There are things like LLMs.txt files that a lot of companies don't have. There are FAQ schemas. This graph here, actually, that's programmatic SEO, we'll get to that. But that was after I set up programmatic SEO through my agent. It went from almost zero indexed pages to about 8,000.
Justin23:44
And what are these pages? Is it content the agents wrote? Is it images with blog posts?
Ian23:53
These are all static pages. Well, actually, let's go to the next slide and I'll get to that, because that's important and I don't want to jump ahead too far. So the first thing to know about LLM ranking is there are three. It's like a three-legged stool. So you've got AEO, GEO, and what I call PEO. AEO is Answer Engine Optimization. That is, when people ask a question to their LLM, it goes and looks for that question and tries to find the answer so it can give them a real-world answer. Otherwise, it's guessing, and that's called hallucinating, right? So it's important to make sure your website has FAQ with proper FAQ schema on as many pages as possible that ask questions people would ask about your company, and then give the answer that your company provides. So that helps a lot with the AEO portion. And then GEO is Generative Engine Optimization. That's when you search in Google and it says, "Here's an AI summary of what we found." That's how it knows what to put in there about your company. And that's a whole different set of things on your website, schema and metadata and things like that. And then PEO is Public Engine Optimization. It's not a great one, but it's basically what the public says about you. You can spout all day long about your brand, but as soon as you put it in the public, the public tells you who you are. So this is kind of reputation management. And this is things like LinkedIn, social media, Reddit, YouTube, blog posts, listicles, articles that are written about your company. So those are the three main factors for being found on LLMs, and two of them, AEO and GEO, are mostly focused around what's on your website and how you build that out.
Justin25:36
Ian, how can I, after this session — let's say I want to set this up — how can I do it myself? Where do I start? Do I ask Claude or Gemini? Where would I begin?
Ian25:45
You could, you could. I mean, Echo Check is a great place. If you just go to echocheck.app, you can actually scan your site and see what's missing, and even copy that stuff and paste it into your AI and say, "Hey, help me fix this on our site." And you can watch it actually grow. It's kind of neat, it's got graphs and all that stuff, and you can see your competitors and how they're doing. So yeah, that's where I would start. I would honestly start with Echo Check, just because you can get a good understanding of what's missing. So the second part, though, is programmatic SEO. Now, this one was really exciting when Enoch brought this to me. This is creating hundreds or thousands, or in Zillow's case, millions of pages, static pages on your website. Now, they're not linked from your website, they're just static pages out there that link back to the homepage or back to all your other places. And these are pages that each one is a very tight, niche-focused content piece that has the FAQ schema in it, has the right titles and all that stuff, and nothing's duplicated across the site. That gives Google immense amounts of data to index, so that when people search for some very specific thing — the best sushi in Nashville, right? — you could have several articles or pages written on that that it would find and then send back to your site. Now, when Google finds it, that means LLMs also find it. So it's great for both. And this can actually be set up programmatically and continued. You can even A/B test it with an agent. It's really a lot of fun. And then the third part is automated blog posts with guardrails. If you have a blog, you can post blog posts all day long, or even have your AI write it and you can post it. What I've done on several of my properties and for some of my clients is set up an automated cron job, essentially, that my bot researches topics, keeps it within the industry, keeps it accurate. It compares it to what's on the website, and then writes two to three blog posts a day. And I've been doing this — I'm almost up to 300 blog posts on several of them. So I've been doing it for a few months. And that number, 300, is actually very important. This is something I actually learned from Justin, from you. When you get to around that amount of blog posts, Google starts to see you as an expert, and that's when EEAT kicks in. That stands for Experience, Expertise, Authoritativeness, and Trustworthiness. When they have that amount of content attributed to you or to somebody, with a date that looks like it's actually valuable, then they start to see you as, okay, this is an expert in the field, we need to start putting more attention on them and ranking them higher. And LLMs love that too. So one thing to keep in mind with all of this is it's not word-stuffing. You're not gaming the system. It's white-hat. It has to be valuable to humans.
Justin28:51
Yeah, because I think a lot about AI slop. People are just putting out AI-generated content. It sounds like from this article you just mentioned that as long as it's valuable and trustworthy data, Google might actually prefer that over a human-written article. Is that what you're getting at with these automated blog posts?
Ian29:09
Yes, and that's actually the part that I got really excited when I read that article, because I have a lot of those pages. I have probably around 30,000 programmatic SEO pages out on my different projects. And I wasn't sure if that was going to get dinged or not. I had done enough research to know that if it's valuable to humans, then they will rank it higher. But you never know how the winds will change. But now we have proof that it is actually preferring that. As long as each one's individual, very specific, and focused on a certain value that it gives to the human, then you're going to be just fine. In fact, you'll be better than fine. And you know that AI won't misspell things. The only thing you have to do is set the guardrails and say, "Make sure that this is factual and actually matches what we offer," and doesn't make bigger statements than you can promise.
Justin30:00
Ian, thank you so much. We have a bit more on the docket here. A quick fireside chat, the bonus, and then everyone who stays until the end will get to take part in this bonus giveaway. So if you guys have follow-up questions for Ian, you're going to have an opportunity to do that as well at the end. Alec, let's bring you back on stage for this one. The chat's still not working, but there is a Q&A. So just send your comment through the Q&A and we can address it there. Alright, fireside, guys. Is AI going to take your job? It's been a hot topic lately. What do you guys think?
Alec30:57
I got something for you. When people hear what I do, I get this question all the time. I had somebody ask me this who is an accountant. And my answer was: is AI going to take some accountants' jobs? Yes, 100 percent. He works at a very, very large company. So my advice to him was: be the accountant at the company that's fostering in the change of AI. Be the person that's implementing the tools, that's using it heavily, that's telling the other people in his department how to use tools and being this wealth of knowledge on AI. And whenever it comes to a point where they're like, "Hey, we don't need as many accountants because we're really efficient," are they going to let you go, or are they going to let somebody who's been doing it like it's 2012 go? So that's the way I would think about it.
Justin31:47
Or they may have a similar experience to you, if you don't mind sharing how that played out.
Alec31:51
Yeah, absolutely. I worked in B2B sales for about four years. And whenever AI started to become prevalent, I went absolutely all in on it and built out the skill set to be able to build AI workflows. I used it very heavily in my role, and it got to a point where I started to be very vocal about how I was using it and showing other people at the company how I was using it. And it got to a point where I had other departments asking me how to use it, and I was presenting for the whole company, and presenting for other companies. And it eventually got to a point where it was time to make it a full-time thing, and I went to my company and said, "Hey, you see what I've been doing for you? I'd like to continue doing this for you, just as a contractor, as an agency." And I was able to get my ex-company on a retainer to continue doing AI implementation for them. So yeah, there's a huge opportunity in becoming that person who's heavily using the tools. And the reality is now that some fifth graders know more about AI than a Fortune 500 CEO might. So as a real opportunity, regardless of your credentialing, to go all in on AI, figure out real-life use cases. And if they're providing value, you then become one of the most important persons in your organization. So I think everyone should take advantage of that opportunity.
Justin33:14
Yeah. Ian, what do you think? Is AI going to take your job?
Ian33:17
Not my job. It is the biggest tech shift we've ever experienced. It's the biggest tech shift in history, and it's going at lightning speed. You'll be just fine if, like Alec said, you learn AI and learn how to master it. And not just low-level, but thinking beyond anybody else, or even just one step past where you think it can go. Because it ultimately is all about humans. AI are not beings that are out there taking people's jobs. It's people behind the AI using it. So I've been telling all my friends and family and people I love, start using it. If you're not using it, start now, because if you aren't, you will be left behind, you'll have to learn it, you'll have to catch up. But I do think it's going to also shrink the larger companies and make a lot of smaller companies that add a lot of value. So expect that. More layoffs from the big companies, but also a lot more things popping up. And I think you'll be fine as long as you think entrepreneurially, and as Alec said, be curious. Let that lead you. You'll be just fine.
Justin34:26
Thank you both. Quick rapid fire, about 30 seconds each. Guys, this is for everyone here who took the time out of their day to attend. We thank you and we want to give back. Alec, Ian, what's the one AI workflow you actually use every week that everyone here could benefit from?
Alec34:49
Yeah, I'll give a super-basic example. This isn't going to be like crazy agentic workflows that are going out and have a team of sub-agents below them doing different things all day. This is going to be super practical. So this is a tool called the PLAUD Note. It is an AI note taker, like an Otter or all the platforms have their own built-in, like Zoom for instance. The beauty of this one is that it's external. So you could use it for real-life conversations, for phone calls.
Justin35:22
And what is that, Alec? You have something in your hand. It's a piece of hardware.
Alec35:26
Yeah, so the same way one of those wallets snaps on the back of your phone. This isn't magnetic, it's an external note taker. The cool thing about that is I have it wired up via a Zapier workflow. Any time a transcript comes in, I record something, it pushes that into my Google Docs. So then whenever I'm in Claude, I can quickly pull that into my Claude conversation, extract things like action items. Here's a real use case I use every single week. If I'm on a sales call, I record it, it gets pushed into my Google Drive. I then pull that into my Claude conversation, and I have Claude skills saved. Claude skills are basically a set of instructions on how I like my PDFs to be done. So I drag that in, I say, "Make a proposal," and it quickly puts out this proposal. Now I can make a proposal in three minutes instead of being deep in a Word doc or whatever. So that's a real-life use case and kind of connecting tools, something that allows me to work super fast.
Justin36:29
That's awesome. Ian, what do you think? One AI workflow.
Ian36:35
Well, I like that tool, I gotta get one of those. What'd you call it?
Alec36:37
It's PLAUD. PLAUD Note. I need an affiliate code. It's a great tool.
Ian36:40
PLAUD Note. Yeah, yeah, you do. Well, mine are pretty flat, pretty simple. I have OpenClaw — I've got 10 of those — and a Hermes bot. And mine are pretty much right where I'm building. I've got three on local machines, and I'll have them build actual software apps, iOS, whatever, and push straight to my phone. And it's all through Telegram. That's one of my favorite workflows. Just, "Hey, let's work on this, do this," using GitHub, and it pulls everything down because it has access to everything. Or if it's on a VPS, I will just go and directly tell them, "Hey, make this change," and screenshots and all that stuff. So I keep mine pretty flat and more relational with the agents, and it's been the best workflow. I do have them checking email for different things and checking security on these servers, making sure everything is up to snuff, and firewalls and all that stuff. But pretty flat workflow, but it's my favorite.
Justin37:43
Thank you. Alright, moving on guys. Pick one or the other. One tool you hyped your team on and now never use, or one tool you thought was basic but now use religiously.
Alec37:54
You kick us off.
Ian37:56
Well, one tool I hyped my team on and now never use. I would have to say… Claude. Actually, I do use it, but I was all in and then honestly switched to one that I never thought I'd go back to, and that's GPT-5.5. It's been doing wonders for me. So I flipped, and I'm kind of sheepish to tell anybody about that, because I eat humble pie. But one tool I thought was basic and now use religiously, probably Supabase, because it's the one that AI suggests for an online database. I thought, "Okay, fine, I'll set that up." And now I've got so many of them, and I'm an expert in there. So I've had to learn these tools as it suggested them to me. Yeah, I would say probably Supabase and Claude. My two answers.
Alec38:55
Ian, I'd hate to do that monthly Supabase bill. I'm sure that's a fun one.
Ian38:59
That's not bad, not bad.
Alec39:01
Nice. One tool I hyped my team on and I'll never use. I'll take that one step further to the extreme. One tool that I hyped to millions of people and now never use. I had a series of TikToks on OpenClaw that got a lot of attention, and I truly don't use it that much anymore.
Justin39:22
Is that a humble way of saying they went mega viral?
Alec39:26
They got a couple views. But I will say, with OpenClaw, I don't regret going down that rabbit hole. And I actually would advise everybody still do it, because there's no question that that's the future and that's the direction that things are going. Having agents that have a specific set of instructions and things to do and are actually taking action on your behalf. There's no question that that's the future, and that's the way things are coming. It's the shift from conversational AI using it in a chat window to agentic AI, where you're asking it for outcomes and asking it to do things. So I would actually still advise everybody get their bearings in something like OpenClaw or Hermes to fully understand what's going on and the directions things are going. I just at this point am not using it that much. So one tool that I thought was basic but now use religiously. Okay, that's a great one. I'm going to say this isn't a tool that I ever thought was basic necessarily, it's just something that I didn't really use. But Codex from OpenAI. Kind of like a Claude Code competitor, or OpenAI's version of Claude Code. Basically, it's just a coding agent where you can build workflows, build web apps, build websites, whatever you want to do. And it's incredible. It's lightning fast, it's very good. Sometimes I'll work with Codex and Claude Code in tandem and have them kind of work off of each other and kind of get a feel for what's doing better at the specific task at hand. So I would say Codex. I never necessarily thought it was basic, I just never really used it. But I would advise if anyone's using Claude Code right now to also give Codex a shot and see what you got.
Justin41:06
Now, like, I second that answer. To put that into perspective, everything you're looking at right now on the screen — this website, the advertisements you got, everything you're seeing here, this interactive widget — was built with Codex. So you can make really beautiful products with the help of 5.5 now. It's really quite impressive, as Ian said. And so I use it out of VS Code. Alec, Ian, do you guys also use it out of VS Code?
Ian41:43
I use it right out of their native app.
Justin41:44
So just download the app. Even if you've never used it before, everyone here can go and download that app. Do you need a subscription? Was it 20 a month or so?
Ian41:50
20 a month, yeah.
Justin41:52
And you can start with a prompt like, "Hey guys, OpenAI, I'm working on this website. This is what it does. This is who it's for. This is our goal. This is the content. This is what I imagine the branding to be like." Start by doing that. Have some fun with it. You can do it this weekend. You'll learn a lot. You can come in Monday morning with a lot of really cool ideas to share with your team.
Alec42:20
Definitely. I'll make one last note on that. To Justin's point about everything you're seeing now being built out with Codex. A lot of people, when you hear Claude Code or Codex, you're like, "I'm not a coder, I don't need that." But you'll see all the marketing assets that we have here. The presentations, the websites, the survey piece that's been built out is all Codex. So you can really use these things to supercharge what you're doing. Building out proposals, building out decks. There's a lot of really cool use cases outside of just, "Oh, I'm not a coder, I don't need that." So I'd be thinking about that as well, that there's a lot of things outside of just coding that you can get some utility from.
Justin42:59
And so we have a lot of questions here about understanding more about vibe coding, just learning AI, becoming proficient in it, chat versus agents. So I'm glad we were able to cover a lot of the subject matter you guys were hoping for us to cover. Last thing before Q&A. Everyone, as a thank-you for staying until the end here and spending your Thursday afternoon with us, we want to give back. You can pick either myself, Alec, or Ian, and we're going to give you 30 minutes of free time. So we're going to drop our calendar link. If you guys need help getting started, want to learn more about LLMs, maybe Ian, if you want to shoot your shot and pitch how you can help there, or Alec, you guys learned a bit about him. Alec, if you want to shoot your shot. Or myself, if you want to get more involved with what we're doing at AI4NTP, or have any questions about growth using AI for growth marketing, go-to-market. I'm happy to chat and just give you some free sauce. So we'll drop our links in the chat for you guys all to book your time. There is limited slots on mine, next week, but we can get time together in due course. Ian, Alec, anything you'd like to add to that before we jump into Q&A?
Alec44:33
Nothing from my side.
Ian44:37
No, I think — well, I would say I saw the question before here, which we didn't get to, but just to answer that one, especially because people were talking about vibe coding and understanding it better. I would say, parroting off what Alec was saying, do try OpenClaw, Hermes, Cline, some wrapper. And understand how it works, because I do think that that will be the future. It's agentic, where you can plug in different brains, so Claude or Kimi K2.5, or ChatGPT 5.5. And just start using it and just start working with it. It takes a little bit to set up, but once you get there, it's human. It's relational. So treat it that way, and just understand it because that will get you a lot further along in the whole timeline of AI than you would with just a chat bot.
Justin45:30
Q&A, let's do it, guys. I know this is tough with the chat not working, but there is a Q&A widget you guys should see near the top or bottom of your screen where you can ask questions. Looks like we've got three here. "Is Codex similar to Claude Design?" That's a great question. So I could take that and then Ian, Alec, if you want to add anything to it. Claude Design — I've only used once now to build a really cool website. I'll try to pull it up while Ian or Alec are going. But I found Claude Design to be better for… links are blocked. Here, Alec, if you want to take that question about Claude Design. I'm going to do some troubleshooting over here. Codex versus Claude Design.
Alec46:35
I'll take that. Claude or Codex is much more versatile than Claude Design. By nature, Claude Design has a very specific use case for what you're supposed to use it for and what it's designed to be used for. Whereas Codex can also be used for that design element, but it has a lot more wide-ranging features than Claude Design. So I'd say to succinctly answer that, it can be used for the same thing, but it can also do a lot more than Claude Design could.
Ian47:15
Yeah, I would say the same. It's Claude trying to section off their model into different specifics, almost like skills. Whereas I've found Claude itself is not great at design. So I have not really tried it because I've had very little luck with it. But GPT-5.5 is actually very good. It's very good with image processing. Same with Gemma. Gemma is also very good. But both of those are my go-to for LLMs that I would trust with design. And Kimi K2.5 actually is pretty good, or K26 is pretty good with design as well, but it's not trustworthy.
Justin47:52
Yeah, in my experience, you could use Codex, you can use Claude, you can use VS Code, have them both running at the same time using a Claude agent, OpenAI agent. But as long as you have the proper files and design files in place, I think it's worth experimenting with all of them. Alec and I have sat down together on a weekend and tried to one-shot a website, and by "one-shot" we mean you just type a quick prompt and see what the output is. And we built some beautiful websites, and other times some atrocious websites. Then you tweak, and you learn, and you try again, and you try different software. Comes out looking beautiful. So I would say the best thing to do is just experiment here. What are you selling? One of the things that's our next question. "What are you selling?" So one of like the part of our ethos here is we're trying not to sell anything. I gave Ian and Alec an opportunity to kind of shoot their shot because they do have some service offerings between LLM optimizations, building agents. But yeah, we don't have an offer necessarily that we're monetizing here. Libby wants to know, "Can each of you share what you like best to work with in regards to these calls?" So Libby, I can speak for myself. I'm really focused on B2B growth marketing enabled by AI and go-to-market motion for generally speaking series A to B startups that have product market fit, just raised round of capital, are looking to raise their next round and need to hit their KPIs to get there. Ian or Alec, if you want to go and just give a high-level overview of what you, you know, I think they're trying to decide who they should schedule a call with.
Alec49:37
Could you repeat the question?
Justin49:40
"Can each of you share what you like to best work with in regards to these calls?" It's really a consultation for you, Libby. Any questions you have. You can use the 30 minutes however you see fit to get you started with AI, upskill you, learning a new software.
Alec49:55
Yeah, I'd say on my side, back to kind of the two pillars of what I do. One is general AI implementation, which is where we take a look at, "Hey, what's the lowest-hanging fruit? What are these processes?" Every business has these processes that you know are outdated, and there's probably a better way to do it, but "Hey, we've always done it this way." So those are kind of the processes that we identify first, and then look at what's the path to getting this automated out. So I would say anyone that has problems like that, I'd be a good fit for. And then on the lead generation front, we do a lot of paid media and lead generation appointments. That sort of thing. So that's kind of my core competencies.
Ian50:35
That's good. I'm very entrepreneurial. Been an entrepreneur for about 28 years and worked with AI for about three and a half years. And so I love to approach projects with "What's the problem you want to solve?" and then figure out, I like to connect dots and say, "Okay, that could actually be fixed by this. This is low level, you shouldn't be working on that with your precious time. Let AI do that." So really, I guess efficiency, productivity, building. I love building. I have one client I've built like 20-plus apps and projects for, redesigned their site. Canberra. They're a pretty big manufacturing company. So really just to come in as a consultant or as an efficiency and productivity person. "What's the problems you're facing? Let me fix that, because we can definitely do that with AI or with some kind of automation." And a lot of that is LLM stuff. Echo Check or whatever else.
Justin51:42
Alec, I think you'll like this one. Joanne, thanks for the question. "What would you recommend for easy landing pages, sales pages for non-techie people? So no-code app." We were just talking about this. Lay it on, Alec.
Alec51:53
I do love this one. I have used all of them. Every single one. Started with Bolt. Throwback to Bolt, Ian and Justin. I'm sure that resonates with you. Very early vibe-coding days. I'd say that was like one of the first very easy vibe-coding platforms that ever came out. Probably late 2023, mid-2023. So I bounced around to all of them is my point. I built things with Codex, I built things with Claude Code, I built things with Lovable, I built things with Bolt, Replit. So I would say for pretty much anyone that doesn't want to mess with a GitHub repository and deploy to Vercel, Lovable all day. It's gotten very good. There was a long period of time where I drifted away from it, because I was like, "This is Claude Code beats this all day." But I think we're at a point now where Lovable has done so much refinement on their product that it's actually very strong, and I use it pretty much daily. So if you asked me six months ago, I would have told you, just figure out GitHub stuff, figure out Vercel. But now I would say absolutely Lovable. They have, like, there's all their Supabase stuff used to be a little bit difficult to tie together, but now it's all baked into what they call Lovable Cloud. If you need anything for storage or anything like that, it's all kind of baked in. So I would absolutely say Lovable.
Ian53:20
Yeah, I'm still in the same place you were at six months ago. I lost hope with Lovable. I have seen some good things, though, from you, so I have to agree that that's probably a viable option. I've heard really good things about Manus, which is a Facebook one. It has pretty much everything you need to just build something. A friend of mine who works at Netflix started building stuff, and it's gone really well for him, and he showed me. He is not a coder at all. And he's shown me some really great projects he's built there that were pretty complex. So if you're looking to build something that actually has function, yeah, probably Manus is for a no-coder. I'm surprised I'm even saying that, because I'm not a big Meta fan, but they've done a good job there.
Justin54:11
And yeah, that's M-A-N-U-S, correct, Manus. Yep. Thanks, Ian. Johan: "Is there a matrix table you could share on which AI tool is best used for?" Yeah, we could put together something and share it with you. I think it's changing so rapidly.
Ian54:25
Yeah, it'll be outdated in a week. I was at a conference up in San Francisco on voice AI. And somebody asked for a prediction for five years out and the whole panel just burst out laughing. They're like, "We can't really predict five months."
Justin54:39
Yeah, days move like weeks or months even. The amount of updates happening every single day, it's like drinking from a fire hose. That's part of the reason why we started this is because there's a lot of people who want to keep up, but just don't know where to start or how to break it down in non-technical terms. And so that's really part of our ethos here is helping our community stay up to date with the latest and greatest tools and have matrices. And then to the next question, "How are you using these tools to make money?" Right. Having a lot of ideas of how you can make money with these tools, and I could, the three of us could talk your ears off. You guys would hate us by the end of it. We wouldn't shut up, the amount of ways you can make money with AI. It's dumbfounding. Between even just winning back your time, building tools, for instance, I had something that was taking me 10 to 15 hours every single week that was just mind-numbing. So I'd hire people to do it, and they'd quit because it was such mind-numbing work. And I'd hire someone else, and they'd quit. I'd pay them more, and they'd still quit, because it was just not fun to do. So I built an agent that does it in five minutes now, so I just won back 15 hours every week and don't have to worry about hiring and training and onboarding, et cetera. So in my mind, that's not making money in the traditional sense, but time is essentially money. Ian's doing a lot of really cool things across his portfolio of companies to use AI to make money. Ian, maybe you could talk on Brand Sauce and how — I'll pull them up on the screen. What you guys are doing. I know you guys aren't using AI in your service, but it's running a lot of the back end, right? So maybe you can speak on how you're making money using AI there at Brand Sauce.
Ian56:27
Sure. Well, first, I think people are saying that Alec your calendar link, I think it was just linked to hosts and panel. So if you want to post it in the chat again, because they couldn't see it.
Alec56:38
Oh, gotcha.
Ian56:41
Yeah, I mean, Brand Sauce. It was well before AI, and so it's really just essentially made us a lot faster and been able to allow us to grow the site, grow the programmatic SEO and LLM stuff, build our back end. We heavily use Claude for that. Claude Code, because we started with it and we just stuck with it. So it's this stuff you don't see that's being built. It's a massive machine. And that's really helped us be a lot more efficient. We only have two developers, and I'm not one of them. I'm a front-end guy. So we've been, I've been working on mostly front end, and those two on the development, and it's stuff that a team of 10 or 20 would take normally. But I would say on the other side is, if you're just looking, if you're looking just to build something, think entrepreneurially. Think, "Hey, what could I do as a side hustle? Just what would I want to build?" Because you have the capability now. And then start going, fix your own problem first. If you fix your own problem, there's likely at least a million people that have the same problem and then put it out there. Make sure you show it to people, make sure that the user experience is really good, and then put a price tag on it and see if it works. Make sure it's not low-hanging fruit that everybody's gonna do. Do something that takes a couple steps in, that you really have to think through. But AI can lead you all the way there pretty quickly. And so yeah, that's what I would say.
Justin58:13
I think it's a good opportunity, to Ian's point, like, find what your competitive advantage is, or what you enjoy doing, and then build AI tools to help you become more efficient. Or use Ian's LLM framework to rank your product or service number one, so that when people type into ChatGPT, "What's the best…" insert your product or service, you come up as number one if you implement the LLM ranking properly. And so, Alec, do you have anything?
Ian58:49
That was exactly why, sorry, that was exactly why I did it. I built an iOS app, put it on the store, built a landing page for it, and thought, "Who's going to know how to find this thing? Nobody knows about it." And so I did the LLM thing, and now I've got 8,000 static pages with PSEO up there, and people are finding it and downloading it and paying for it. So it does work.
Alec59:08
Love it. I think my answer to that question, I think context is very important. If 10 different people ask me that question, I'm going to give out 10 different answers. Very dependent on your situation. If you are someone that's a 19-year-old and is asking me and is currently doing nothing and is asking me how to make money with AI, my answer is going to be very different than if you're an electrician, and you're asking me, "How can I use AI to make more money in this specific business?" So it really depends on where you are, what you're currently doing. If you're looking to make more money in this current business, in this current vehicle that you're in, or if you're looking just like generally to find a way to make money with AI. So I think context is very important in the question. And like I said, 10 different people asking that question is going to be 10 different answers.
Justin59:57
Yeah. Sydney actually asked a question here too about crowded spaces, and there's so many models. That's true. That has become an issue.
Ian1:00:10
The question was: "Speaking of crowded spaces of AI tools, many of these cost money in order to actually see their capabilities. How do you determine whether the investment is worthwhile? I can't pay for all of them, there are so many, in order to see what's legitimately useful." There are local models you can download to use just locally that don't cost anything. It just takes up space on your computer, so if you have a kind of bigger computer, or at least something that's more modern. But you can also get away with a lot just with Chat or Claude, just one of them. I mean, that will be enough for you to at least get your feet wet and understand what they're capable of. And then you start to branch out as you start to see the nuances between them all. But most of them have a free account. And I would just suggest picking one and going with it, you know, honestly.
Alec1:01:01
Definitely. I'd say like this is coming from somebody that spends thousands of dollars a month on subscriptions. I would say in a lot of cases, there's a free trial option with a lot of companies now, I'd say probably at least half the time you're looking at a solution where you're going to be able to test out a free trial, see if you get enough value out of it, proceed if so. For cases where that's not an option, I would say just get to a point where you've done so much research on it, and you've seen so many reviews that you're like, this is an absolute yes. And if it's not an absolute yes, then don't do it. I think everyone has to be pretty open-minded with spending right now, especially with AI tools and AI subscriptions, just because of the nature of the opportunity that we're in right now where you really can set yourself far ahead. So I have been pretty loose with subscriptions and things that I'm willing to pay for. I think generally people should be a little bit more open to spending than they would be otherwise on these things that do have potential for a lot of upside. So that's kind of the lens that I would look at this through.
Justin1:02:12
Yeah, that's good. Just a couple more questions. Alec, would love for you to stay, but if you have to hop since it's the top of the hour, I completely understand. There was a question about using these tools to make money. Is it doing marketing as an agency? Yes, that's a great way to make money. When using AI tools, I can speak from personal experience. I think all three of us can. At the end of the day, when you become proficient in AI, which doesn't take long if you're building with it every day for months. You can really get lethal quickly if you spend the time and energy, especially on weekends, after hours, after work, at work, building tools for yourself to kind of replace yourself at work to free up some of your time. But anyways, creating solutions for whether it's a small business or, like Alec said, an electrician, a family member who owns a business, a friend who owns a business. Coming in and being the AI specialist pays its weight. You can make a massive impact in as little as a few hours spent with a business. There have been implementations that I can speak for myself, but I'm sure that all three of us have had where we come in for 30 minutes, an hour, a few hours, a half day, a day, and make a profound change to that business owner in terms of lead generation, AI automations, workflows, daily reports that land on their dashboard via email every morning that keep them up to date with their to-dos. I think Alec made a great point. 10 different people ask that question 10 different answers. There's so many ways to approach a business and make money with AI. But marketing as an agency is a very good one right now. Joanne needs support. That's what we're here for. Book time with me, Ian, Alec. You know, all of us. If you need the extra support, we'll extend that offer to the rest of you who are here. There's 15 of you left, as a thank-you for joining us. Alright, guys, thank you. I think that's a wrap. Ian, Alec, any closing remarks? Anything you want to say to our incredible audience here?
Alec1:04:44
Yeah, well, for everyone that's hung around this long, appreciate you guys. It's been great to be able to share some of the things that we spend way too much of our lives doing day in, day out. So to be able to share some of the stuff that we do with you guys is great. So if anyone has any further questions or anything else they want to talk about, like Justin said, feel free to book a time with any of us and we'd be more than happy to do it.
Ian1:05:07
Yeah, I will say too, if you are here, especially at the end here, you're an elite level of interest and it's a very small amount of people that are pursuing it this far. So you should be lauded and praised for that. I think it's a great thing. It really encourages me because we're so excited about it, and we love it when others are, and we'd love to continue the conversation with you. So reach out, love to talk to you.
Justin1:05:34
Well, guys, I'm going to send a follow-up email with everyone's calendar, some notes. So I'll be on the lookout for that. If you have any questions, don't be a stranger. I think all three of us are happy to keep in touch. Cheers, everyone.