Episode 835 | The Right Way to Use AI in Your Startup Marketing

Rob Walling (00:00): You’re listening to Startups for the Rest of Us. I’m Rob Walling. In this episode, I talk with Taylor Hendricksen. He’s been doing performance marketing online for well over a decade, including e-commerce, B2C, B2B, and has bought tens of millions of dollars in ads over the years. He said it was both Meta and Google when I asked him in the interview, but I’ve known Taylor for about a decade and he’s just extremely knowledgeable when it comes to copywriting and making incredible offers, general marketing, getting people interested in something and actually paying for it. Taylor is one of our coaches at SaaS Institute, saasinstitute.com. If you’re doing at least one million in ARR and you’re a SaaS company looking for a coach like Taylor. Before we dive into our conversation, I’ve been writing emails again, essays that I’m sending out to my email list. I haven’t blogged regularly since 2010, 2011. 

Rob Walling (01:07): I started a few months ago writing my thoughts. These are new thoughts. Some overlap lightly with the podcast, but a lot of them are new thoughts, frameworks, et cetera, that I’m mulling over. You can head to robwalling.com/emails if you’re interested. And with that, you’ll get a free sample chapter of The SaaS Playbook. But realistically, the list on my site says, “Hey, you want some proven cheat codes for growing your SaaS in 2026? Marketing channels that actually work, building a sustainable moat, refining your pricing.” That’s the kind of stuff that I’m talking about when I send these emails. It’s one email a week and it’s an essay of my original thinking, robwalling.com/emails if you’re interested. And with that, let’s dive into my conversation with Taylor Hendricksen where we cover several topics, but one of the ones that we focused on a lot was where founders of all types should be using AI in your marketing and where you shouldn’t be using AI. 

Rob Walling (02:03): And I really appreciated the back and forth that it was kind of impromptu. I had some questions and an outline of what we were going to cover, but it was so interesting to me the way Taylor was thinking about it. And then he and I piggybacked off each other a bit thinking through using AI in marketing operations, content generation, all that kind of stuff. So that’s a highlight of this conversation. Let’s dive in. 

Rob Walling (02:34): Taylor Hendricksen, welcome to Startups for the Rest of Us. 

Taylor Hendricksen (02:37): Thanks for having me. 

Rob Walling (02:38): Yeah, it’s great to finally have you on, man. So you and I have known each other for at least 10 years, maybe 11 or 12. You and I have hung out at Rhodium. Chris Yates had a mastermind going and we hung out for a weekend at one point. You spoke at MicroConf several years ago, talked about, was it funnels and ad stuff? 

Taylor Hendricksen (03:01): Everything that was working at that time. Yeah. 

Rob Walling (03:03): So Facebook ads, funnels, quizzes, all the hot stuff at that time. That goes back to 2018, something like that, 2019. 

Taylor Hendricksen (03:10): I would guess around there. Yeah, 17, 18. Some of that works still. That’s how it always is, right? Half the stuff I talked about at my MicroConf talk back then still works and the other half doesn’t. 

Taylor Hendricksen (03:19): Oh yeah, for sure. 

Rob Walling (03:20): I mean, when I think of your career, what I know of it, you’ve run a lot of Facebook ads. I don’t know if you ran Instagram and other stuff, Google AdWords, but a lot of ads, sold a lot of, correct me as I get done with this if I’m missing anything, but I think a lot of stuff for e-comm and then info products and had some of your own, also did affiliate deals or for-hire stuff. Is that a summary? 

Taylor Hendricksen (03:46): Yeah, it’s pretty good. I would describe myself as a full-stack marketer. So everything from ideation, product design through the copywriting, some of the actual putting the product together, less operations, but then how do you take it, do all the landing page design, connect the funnels, and do the paid media on the side. So you can kind of go soup to nuts with taking a product that’s already there and taking it to market. We’ve done it for online education, some of the digital info stuff, a lot of lead generation, specifically in the financial services industry, and just tried to figure out how to convert cold traffic. We’ve spent a tremendous amount on Facebook as well as Google. So across YouTube, display, Facebook, Instagram, those are the kind of the main levers I pulled in the time. 

Rob Walling (04:22): That makes sense. And you of course had involvement in SaaS as well, otherwise we wouldn’t have you speak at MicroConf and you wouldn’t be a coach at the SaaS Institute, which folks have heard me mention many times on this show. So I guess given that you are now coaching and giving people advice on how to grow, how much do you think you’ve spent on ads across all of this stuff you’ve managed? 

Taylor Hendricksen (04:55): We’ve done easily tens of millions, up to multiple six figures a day in spend we personally controlled and overseen, and affiliates doing almost up to a million a day. 

Rob Walling (05:04): I just want to give people an idea that you’re on here for a reason because you’ve been walking the walk and talking the talk for a very long time and you know what works and what doesn’t. So as I think about you coaching SaaS Institute, B2B and B2C SaaS founders at one million and up in ARR, what’s the number one thing that you hear keeping these folks up at night? 

Taylor Hendricksen (05:24): I think the biggest thing right now is probably AI. Like the boogeyman that gets thrown around. It supposedly has the worst approval rating out of anything out there, but it is the constant thing. With SaaS, with software, we’ve seen these big public companies absolutely taking nosedives in terms of their valuations because AI is going to come and take everybody’s jobs and ruin everything. And the biggest thing now is there’s still this kind of break between AI being a cool whiz-bang toy and actually being effective. So you have this gap that these founders are going through, and a lot of it now is like, “How do I prove my relevance? How do I keep my pricing in the time of AI?” Especially now when customers are coming and saying, “Hey, you use AI to do this.” 

Taylor Hendricksen (06:02): “Shouldn’t it cost a tenth as much, or be doing 10 times more than you normally do?” And I think this is an interesting question of how do you prove your value in a time where people just say AI is a buzzword or a fix for something, even though it’s not there yet. We have one founder in particular who does high-end UI/UX design services for startups and SaaS companies. And the bigger question is, yeah, by all means, go have Claude vibe-code your whole product, and it’s going to be the generic AI output that people are used to. But what comes with the curation and taste and actually having an expert drive it? It’s like basically giving somebody a Lamborghini. If somebody is brand new, you put a toddler in there, they’re going to crash it in a second. 

Taylor Hendricksen (06:42): But somebody who knows how to drive, that thing is a perfectly tuned machine. So a lot of us are trying to figure out how to integrate AI that’s not just a cheap ChatGPT wrapper, but really utilize it as a human-first tool rather than a full replacement agentic thing. 

Rob Walling (06:59): I heard you noodling on something where it was like, use AI for ops and scale, but keep your content and maybe your front-facing stuff authentically human. Do you want to double-click into that? 

Taylor Hendricksen (07:13): For sure. Yeah. I think we’re in the very early days of seeing what AI can do for video, for text, for everything else. And it’s taken basically what was a human-level labor arbitrage and made it ubiquitous. Now anybody can generate as much text as they want instantly. So I think the flood that will come from this, both in video as well as text, the whole world of SEO and all these things is going to be flooded with mediocre content. So I think the big delta in the next couple of years is going to be: what human-first things can you use that are still uniquely human? Something like cold email, where before it was more labor-intensive to supposedly customize each cold email to reach the desired person and have some kind of relatability. 

Taylor Hendricksen (07:56): Now that’s all agentic. And so the amount of cold email that’s going to come through is going to flood inboxes and make it basically a useless channel in a lot of ways. I’ll probably get a lot of hate from people who still use cold email and it still is absolutely effective, just like SEO is absolutely effective and same with most of these marketing channels. But the human-first premium on top of it is going to be a bigger and bigger deal the farther we go. So I think we’re pretty close to passing the Turing test from video. You can still tell AI video for the most part, but a couple of times I’ve been like, “Wait, that definitely fooled me,” which is interesting. 

Rob Walling (08:26): Yeah. And so what happens when that happens? When we get to the point where even video generation, like this video right now, you and I both know we’re humans. If we took the Voight-Kampff test, we would know we’re humans. But what if we could generate video just like this, at this level of quality? How do you think that changes the game? Should founders use it or should they stay away? Is keeping it authentically human meaning a human does it, or what are your thoughts? 

Taylor Hendricksen (08:57): I think the humans need to do it. People have a really good sense of whether something’s legit or not. Our BS detector is high. Whether it’s an offer or something like that, people can generally tell. So when everybody else takes the lazy route and hides behind not going in front of camera, or just has the AI do it, spinning up a sexy avatar to be the pitch person instead of actually doing it themselves, I think it’s just going to be another cheap thing. What’s going to separate you from every other person who could spin up the same avatar? So the human-first premium is real. It’s like when computers first beat humans at chess, back in the 80s or 90s, and chess isn’t dead. 

Taylor Hendricksen (09:33): It’s just now kind of a different game with how it’s played, but it still has that human-first level to it. 

Rob Walling (09:38): It’s interesting because a lot of times we’re asking the question, can we tell the difference between if AI did it versus if it’s authentically human? I don’t think that’s the right question. I think the right question is: is the content any good? Because if you show me two essays or blog posts or marketing landing pages or two podcast episodes, one is human and the other is AI, but they’re both mediocre, does it even matter at that point? And that’s the thing I think we’re forgetting. I have yet to see flat AI just generate something great. If I say “write a blog post on this topic,” it’s a solid five out of 10. It’s mediocre. Now if I say, “I’m Rob Walling and I’m going to dictate a new framework that I thought of,” and I talk for 10 minutes into ChatGPT and say “turn that into a blog post,” that post will at least have my phrasing. 

Rob Walling (10:37): It’ll have my concept that isn’t on the internet yet. And that’s the human aspect of it. I’m going to let you talk, but I’m fired up about this because I don’t think it’s a detection issue anymore. I think it’s just a quality thing. And I have not seen AI YouTube videos that are just AI voiceover on stills impress me. Even if the voice is good, I’m like, “This is just boring. This is flat.” It’s like a mayonnaise sandwich. It’s just a very bland thing, which I think comes back to what you’re saying about humans having to do that. 

Taylor Hendricksen (11:11): Oh, absolutely. You see the videos on YouTube that are clearly bulk-generated, throwing them out there to try to get views. You skip them the second you feel it. You can feel the AI voice, whatever it is. So in general, that’s what’s going to be the biggest differentiator: doing things a little bit more thoughtfully to stay on top of this mass of AI-generated content that’s going to flood the space. You can’t just give something to AI and say, “Hey, write me a blog post or write me sales copy or do this whole thing for me.” I’ve seen a couple of AI tools recently that just say, “Put your website in and it’ll automatically do the go-to-market for you.” 

Taylor Hendricksen (11:46): And that’s a cute idea, but in practice it all falls flat. Every time I have it go through and do it, it’s a really good enabler on the backend. Operationally, being able to do the research and compile the stuff to get it to a point where you can see the nuance and the details, market research or persona research, it could be incredibly effective. But by the time you then take that and have it try to spit out the actual headlines or the actual human-first things, it’s not able to grasp the nuance. It’s very good at “next best thing” or grouping things together and bulk-approaching it. 

Taylor Hendricksen (12:24): But this whole thing of how do you take something over here and connect it here, or really have a nuanced understanding of what a human is actually experiencing to then be able to relate to them? AI will be able to do a bulk offense at it, like a DDoS attack, overwhelming the systems. And that’s the whole thing: people are going to use AI to spin up a hundred ads each with micro-differences between the avatar or the hook or the phrase, and hopefully one of them is good. That’s the only way I can see AI potentially beating the human-first stuff in the short term, meaning one to five years. 

Taylor Hendricksen (12:59): But after that, you’ve got to have the human level into it because it’s about nuance and the things you don’t fully grasp. 

Rob Walling (13:05): I had Eric Ries on the show because he has a new book out, Incorruptible, and one of the things he said I thought was a very succinct description of what we’re talking about. The moment he said it, I knew it in my bones. He basically said AI is the average of the internet. It’s the average of everything. And so if you’re not a good copywriter, if you’re a two out of 10, and you have AI write copy, it’s going to write a five out of 10. And you’re going to be like, “Man, AI is so great.” But if you’re a six, seven, eight, nine out of 10, if you’re actually good at something, you look at AI and you’re like, “This is solidly mediocre. This is like what an intern would write.” 

Rob Walling (13:45): Same thing. And it’s the Dunning-Kruger effect: the less you know about something, the more you think you know. That’s why, if you’re an expert, if you’re a really good designer, you know AI is not a good designer. If you’re a really good copywriter, you know AI is not a good copywriter. If you’re a great musician and you listen to AI music, you know it’s not actually that good. But this is why you see these sentiments on the internet of folks saying, “AI can do this and tomorrow it’s going to be even better.” And it’s like, yeah, a little bit, but it’s not 10% better every month. So that’s what we’re getting at: it produces average work unless you give it above-average inputs and above-average refinement and revision, and you actually train it not to be average. 

Taylor Hendricksen (14:33): Which is going to be a million times more valuable, because the average is going to go from having a hundred people putting out posts to now a million people. So if you do a little bit more to sit on top of that, it’s just going to skyrocket your ability to do things. But if you’re a two out of 10 copywriter and you’re just a straight dev, yeah, you could use it to have a baseline, do persona research, write headlines that work, and get an average passable thing. Sometimes that’s fine. You don’t need to be a pro in everything. But having that expert-level task in whatever it is, especially selling for founders, you have to have that empathy and really believe in whatever it is, otherwise it’s just sales material, just a headline, and it won’t actually resonate with people. 

Rob Walling (15:13): AI is helping founders ship faster than ever, but at some point you’ve probably wondered, is my codebase actually ready to scale? Today’s sponsor Designli can help you answer that. Here’s their CEO, Keith Shields. 

Keith Shields (15:26): Thanks, Rob. We’re hearing from a lot of founders right now who are avoiding parts of their codebase because something might break. And honestly, that’s a sign that your product isn’t ready to scale yet. AI development leaves behind hidden security issues, fragile architecture, and features way more tangled than you’d expect. That’s not a knock on AI, it’s just reality. So we built the engineering intensive. In just two weeks, our senior engineers do a deep audit of your code, stress test your infrastructure, uncover vulnerabilities, and hand you a prioritized roadmap to get scale-ready. We give you full clarity on your product’s health, backed with a money-back guarantee. So if you want to keep your AI tooling, but add professional oversight from senior software experts, book your engineering intensive at designli.co/fortherestofus. That’s D-E-S-I-G-N-L-I.co/fortherestofus. 

Rob Walling (16:18): So I talked to a couple of the founders that you coach and asked, “What is Taylor exceptionally good at? What has he helped you specifically with?” And one of the founders told me that you’ve really pushed him to rethink sales and prospecting in the age of AI. You obviously come from performance marketing where the numbers can get pretty crazy. What do you see bootstrapped founders getting wrong about how fast they could actually grow? 

Taylor Hendricksen (16:42): I think it’s just this artificially placed ceiling that they put on themselves. “I could never imagine doubling my business in a week.” Once you get past that idea and limitation, you realize the only limitation is your own comfort. You don’t necessarily think this will work, or they’ve tried something like that and it failed in the past, so therefore they don’t try again. The biggest thing founders, especially SaaS founders, could take from the performance marketing world is: when something’s working, how can I 10X this in a very short amount of time? What would it take? And once you start to do that, you start reverse engineering: okay, what would it actually take? 

Taylor Hendricksen (17:21): We had a project doing this for lead generation, and it was, “What would it take to get to 100,000 leads?” About 10 times bigger than what we were thinking. And then my business partner and I started going back and forth and realized, “Oh, all it really is, is we just need to crank the dial up to 11 on ad spend and have the backend operations that can handle it.” It was a very scalable deal. The only thing holding us back before was our own limitation. For the founders we’re talking to, SaaS can scale pretty quickly operationally. And especially the service-enabled ones, there’s a lot more headroom there to increase than they ever realized, but they always just get stuck at something comfortable, especially when their needs are met. 

Rob Walling (18:00): We’ve seen a lot of people in the TinySeed community: once they hit 20, 30, 40K MRR, life’s good. They don’t have that broke fire burning under them as much. So they kind of limit themselves naturally, or they’ll naturally get to the temperature of the room around them. If you’re in a group and everybody’s doing $50K and you start to do $100K, you kind of coast. Your goal is not to get to 500K, but just to sit there. So I think getting around people who are doing higher numbers and thinking bigger is key. Not just that ephemeral “think bigger” motivation, but actually, “Well, what would it take to do 10 times what we’re doing?” 

Taylor Hendricksen (18:36): And then what would it take to get up there? But at the same time, not just scaling for scaling’s sake either. I remember having this conversation: “What’s your goal for the business?” “Well, we really want to get to five million ARR.” “Why?” “Well, that just seems like a good number to hit.” But you were there before, you were doing two or three million a year, and now you’re back down to one or two. Before, you had twice the people, twice the headaches, and you were making the same amount of profit. So scaling for scaling’s sake is still not necessarily the best answer. 

Taylor Hendricksen (19:07): And obviously the TinySeed and MicroConf community is not about scaling just to get the Lambo and sit in front of a private jet. But I think it is important to think: why am I doing what I’m doing, and would it be better to scale more? It’s going to have different problems when you get there, but thinking bigger is probably the biggest thing you could bring. 

Rob Walling (19:25): Another founder that you coached told me that you’re really good at go-to-market and marketing and getting content out there. So if a founder buys into what you’ve said, “Maybe I do want to 2X or 5X this year,” and they’re ready to push hard, how do they actually get in front of people these days? What’s working? 

Taylor Hendricksen (19:46): I think the biggest thing is: scaling operations is about 10 times harder than scaling marketing. Marketing’s my hammer, so it’s obviously the thing I want to swing. But making sure you can fill the backend first. Assume SaaS is scalable. Getting out there in front of people is now going to be about really figuring out who your target demographic is. What are the core problems they’re actually trying to solve? And how can you get to the places where they’re actually talking about it? You’ve got to have a really deep understanding of who that market is, and the market may be very different than you actually think. You have to have an offer for getting in front of them that they really feel stupid saying no to, or isn’t a huge burden to get over, and then you have to solve their problem in a different way. 

Taylor Hendricksen (20:26): Starting with the avatar: who are you going after? A lot of people want to go after the most common ones because they don’t have creativity. They go after dentists and chiropractors. But they often don’t realize some of the non-sexy niches are the best places to be. Everyone wants to go after the startup or the FinTech, but the companies who are making money hand over fist and still have a very backwards view of technology, some of those are the best ones to go after. They’re the least critical on product delivery, timing, and speed. So really figuring out who you’re going after is an important starting point. 

Taylor Hendricksen (21:07): The other thing is really trying to figure out what problem you’re trying to solve. You’ve been talking about solving problems with software for decades. And I think a bigger thing now is not necessarily solving the problem they have right in front of them, because a lot of people aren’t waking up and saying, “I need a new procurement software because it’s inefficient.” But really trying to figure out how you’re solving one of their primary problems with a secondary solution. For the UI/UX design agency, they’re not necessarily waking up shopping for a new agency to redesign their SaaS website. They’re really trying to figure out how to look more modern, like they’re not about to go out of business. If somebody comes to your site and thinks, “I don’t even know if those people are still working,” or “The product is going to be bad because it looks so dated,” that’s the actual problem you’re solving. 

Taylor Hendricksen (21:59): Or for one thing we’re working on, a business for digital disbursement cards for corporate payouts. They’re not actually solving “how do I get someone a debit card instead of writing a check.” They’re figuring out how to retain the best workers in areas where churn is one of their biggest problems by making sure people get paid right away. So that’s where some of the positioning is really about: how do you solve the problem that actually keeps them up at night? Most of the time, founder products don’t actually solve that problem, but they definitely can. 

Taylor Hendricksen (22:40): It’s kind of like Nike’s “Just Do It.” They’re selling generic T-shirts and clothing, but what they’re really selling is their ability to encourage you to do that thing you’re scared to do, through clothing and shoes. 

Rob Walling (22:52): Talk to me about some unusual marketing approaches. I’ve heard you mention direct mail and founder-led ugly talking-head videos. Is that working? Have you seen that helping people out these days? 

Taylor Hendricksen (23:06): Yeah, totally. In general, I’m very agnostic about what marketing channels are best for a business. “Best” is very relative based on who you’re going after and what you’re trying to accomplish. In the TinySeed community, in the B2B space in general, you have a target demographic you know you’re going after. Let’s say it’s chiropractors or a specific type of business. Going after that using paid media channels like Facebook and Google, it’s really hard to get that granular on targeting just chiropractors or just the office admin assistants for chiropractors. Cold email has obviously been a great channel for these founders. But most people now ignore most emails in their inbox, and I basically ignore all cold email and cold LinkedIn spam. 

Taylor Hendricksen (23:49): It’s just so egregious that they don’t really get on the platform anymore. So when founders need to get in front of the demographic they need, approaching the old crowded channels that used to work a couple years ago but now don’t, you just have to find a new channel and get in front of them in an uncrowded way. We’ve found that with direct mail, being able to mail something, not a cheap postcard that gets chucked as soon as it comes over the bin, but there are handwritten letter services, there’s stuff in a plastic garbage can that you stuff the letter into and stick in a package. We even got video mailers shipped over from China that you can upload a custom pitch video into, slip into something, and send it out to get in front of the demographic you really want at surprisingly low cost. 

Rob Walling (24:30): Even the video mailers with shipping, with labor and everything, are about $25 or so to get in front of somebody. So when they start thinking about that, they can actually approach their target market in marketing channels that are not crowded now. But you still have to marry that on the other side with the trust built, because as soon as they get that thing or see your ad, the consumer isn’t necessarily just clicking on the ad and going directly to your landing page. They’re immediately opening up a new tab, searching for your company, trying to figure out who you are, checking the reviews, and trying to figure out, “Should I be able to trust these people?” So one of the SaaS Institute founders started making videos on YouTube to build that trust and started posting her expertise across a number of platforms: YouTube, Facebook, LinkedIn. 

Taylor Hendricksen (25:12): And she’s basically now got a really good trust built up over that. She’s a known entity in the space. “Thought leadership” is such a phrase, but she’s basically known as a trusted entity in that space. But that itself isn’t necessarily going to drive a lot of the results. People think, “Okay, well I’ll have a YouTube channel and leads will just flow in.” Kind of. That’s usually part of it, but it’s the trust layer on top. If you’re really trying to go after specific people, those people aren’t necessarily searching for what you’re talking about. So going after them very specifically with direct mail or direct outreach of some kind, send the brownies in the mail, do something different that’ll get through the noise. Then once they go see you, the trust will be built, and they’re much more likely to convert. 

Rob Walling (25:54): We’re talking a little bit about distribution here, and as you drive traffic and generate interest, that attention has to land somewhere. So talk to me about the offer itself. When you think through an offer that you’re going to make, what makes someone actually say yes? 

Taylor Hendricksen (26:10): Yeah, totally. So if you put yourself in their shoes, they’re getting bombarded with different offers, pitches to hire services, spend money with them, or try something for free. That has gotten so saturated because their inbox is full of them. You have to figure out what is the thing that’s going to cut through, and most of the time that’s going to be the baby steps going into it. It’s really hard to come in and pitch somebody cold on a big $100,000 project. But if you can come to somebody and say, “Hey, we have this very specific thing that solves a very specific issue for you. You don’t have to replace what you’re doing already.” 

Taylor Hendricksen (26:44): “It’s going to take a very short amount of time to implement and see the results, and you’re not going to have to spend a bunch or take a big risk on whether it will work.” So for any company out there: what is the thing that’s going to make them feel stupid for saying no to? Alex Hormozi talks about offers all the time. It’s a really good place to start on figuring out how do you stack something that people feel stupid saying no to. And feeling stupid for saying no to isn’t necessarily just a free trial, because a lot of people are offering free trials now. It is: how do you solve the very specific issue that they have, or their primary issue with a secondary solution like we talked about earlier. 

Taylor Hendricksen (27:17): Putting something in front of them that gives them a low-risk way to start testing this relationship with you. That could be something where they start the relationship, watch your founder talking-head videos for three months, six months, a year or two before they’re ever actually ready to interact with you as a business. But once they do, having that thing that makes it easy to say, “Okay, yeah, I’ll give it a try” versus committing to something big, I think should be in every founder’s playbook. The biggest problem with some of the software stuff is the replacement offer: “Come in, I’ll replace your existing internal systems.” That’s a really big ask. It’s a really big one-way door. 

Taylor Hendricksen (27:54): “If I try you and you don’t work, I’ve got to go back and revert to the old system. There’s got to be a lot of breakage. It’s going to be huge pain. It’s going to take months of migration.” All those things. 

Rob Walling (28:05): As we wrap up, I want to circle back to what we started with, which was talking about AI. Lest people think that we are anti-AI or don’t use AI or don’t think founders should use AI, we think none of those things are true. Throughout TinySeed and MicroConf, all of our team members have Claude and ChatGPT accounts and we are using a lot of it internally to automate a ton of stuff, and I know you are as well, and I know you’re seeing the founders that you coach use it too. So I wanted to say, we covered a lot of stuff today: positioning, scaling, distribution, offers. It’s a lot of plates to keep spinning. How are you using AI to make things faster? Or more specifically, how are you seeing the founders you coach use AI? 

Rob Walling (28:53): If it’s not what we said at the top, it’s not to generate the video, it’s not to generate the blog post. So how should we use AI? Because we all should be. It makes us faster. 

Taylor Hendricksen (29:03): Absolutely. I think there are two core pieces of it. One is the agentic stuff: how do I have this repeatable task that I’m already doing be an automated thing? It shouldn’t take four hours a week to do this thing. That’s one level of AI where especially the production and operational people should be absolutely leaning in heavily to whatever agentic framework it is. Whether you’ve got open source tools, Hermes, Claude, whatever it is, that’s an important thing: where am I wasting time on stuff that could free me up to do higher-value tasks? Then the other side is the non-repeatable things. The things you can’t necessarily just hook up an agent to go do research on. 

Taylor Hendricksen (29:41): And then I think it’s more this iterative process with AI as a thought partner. AI is really good at taking long rambling stuff from me, making sense of it, and going back and forth ping-pong style to really develop something out. So for me, it’s a lot of market research, a lot of demographic research, figuring out who we’re selling to. One of the companies we invested in is specifically going after truckers. So getting up to speed on the space, going out and pulling: “Go to Reddit and pull 50 comments based on this, get their actual verbatim language that you can use across the sales stuff. Go do deep demographic research.” Then build all that stuff up to fine-tune the AI for starting to get some initial rough drafts on copy. 

Taylor Hendricksen (30:21): So this kind of back-and-forth thought partnership thing. And making sure you have good guardrails on it too, because obviously the “Oh, you’re totally right” thing gets thrown in a lot. So having custom instructions like: don’t just placate me, really challenge the thoughts, don’t make stuff up, point out where I’m doing something wrong. I’ve developed some custom frameworks around this to hopefully give it some better rails. What I use it most for is how do you speed up the learning curve and compress that time down? For the trucking company we invested in, it’s a FinTech platform for truckers. Making jokes about Swift, the trucking company, is one of the best things to poke fun at. All these different things I would never have known, but AI basically goes out and does the market research to pull actual quotes from different places. 

Rob Walling (31:02): Taylor Hendricksen, as always, a fount of wisdom. I’m really appreciative of you coming on the show. You are a SaaS Institute coach, as I mentioned, and folks can hit saasinstitute.com if they’re curious about that program. It’s our premium paid coaching program for SaaS founders doing a million or more in ARR. So thanks for all that you’ve given back to the MicroConf and Startups for the Rest of Us community over the years, man. It’s great to have you on the show. 

Taylor Hendricksen (31:29): Thank you for putting everything together and thank you for having me on, Rob. Really appreciate it. 

Rob Walling (31:32): Thanks again to Taylor for taking the time out of his busy schedule to join me on the show, and thanks to you for listening this week and every week. This is Rob Walling signing off from episode 835.