Are You Treating AI as Mission-Critical — Or Just Dabbling in It? With Paolo Vidali | Ep #793

Are You Treating AI as Mission-Critical — Or Just Dabbling in It? With Paolo Vidali | Ep #793 Speaker 2: Hey, welcome to the show, Paolo. How are you doing? Speaker 1: Thanks so much for having me. I'm doing great. Speaker 2: Yeah, excited to have you on. Tell us who you are, what you do. Speaker 1: Sure. So my name is Paolo Vidali. I have an agency called Hidden Gears. We focus solely on e-commerce and within that actually Shopify. Been doing design, development, digital marketing for growing Shopify businesses for the last 10 years and have a small distributed team across the US and Europe. I have been developing for Shopify myself since 2009. I've been in digital marketing doing PPC since about before. I have had this sort of long trajectory of working in digital and seeing things evolve. Speaker 2: Awesome. How did you start your agency? Are you an accidental agency owner? Speaker 1: I think it was accidental that I ended up doing it full time perhaps, or I backed into it a little bit. At the time I was working for a digital marketing agency, doing a lot of B2B, a lot of lead gen. They didn't touch e-commerce at all, so they actually allowed me to do this kind of And I've been working in hidden gears, you know, named freelancing as side work for a couple of years. And then at a certain point, I realized, you know, what I was spending my nights and weekends doing was actually something that was scalable that I could work for myself and try that out. And, you know, I was in a fortunate time in my life that I was just really low cost of living enough to New York. It felt like I could take the risk, I could take the jump and worst case, you know, I could go back to doing something else for someone else. But yeah, I really enjoyed the challenge of it and ended up growing it into a team. You know, very intentionally, we've been a smaller organization, even though we've been around for a long time, just because we like just having really intimate You know, kind of relationships with the clients, having really close bonds with team members and yeah, it's been fun. Speaker 2: Awesome. And AI has been the talk of the town for the past couple of years and it's really changing a lot of the landscape. You know, you think of like content agencies that used to write all the content. Now, you know, someone in AI can write or like when I I'm drafting up an email or copy for a landing page or need an image. You know, it's amazing what you can do. How is your agency using AI? Speaker 1: Yeah, so it's a great question. We actually sort of had to figure out how we wanted to approach it. I think as someone who is a hybrid marketer and developer, I knew of a lot of SaaS companies calling things AI and never really believe it until a certain point where it became apparent like, oh no, this is really happening, right? For us, we actually decided to come up with an AI policy, both for staff and contractors in terms of what to use, how to use it, and sort of how to involve us in the process of figuring out, hey, it's cool if we can experiment with something. If we're going to do anything with client work, how are we actually going to handle that? And then also one for clients, here's the way that we're using AI and you could opt out. But in some cases, Shopify, Meta, Google, obviously all these platforms are just building it in whether you like it or not. And so we definitely did a lot of R&D just for our own internal use, nothing billable, Really, that would be considered more than a rough draft. And then as things have evolved over the last couple of years, we found really interesting, very specific use cases where it's just saved massive amounts of time. For instance, like all text generation for accessibility and for SEO to work with an e-commerce client who has a database of honestly, even just hundreds, but sometimes tens of thousands of SKUs. Even if they take it on internally, to task anyone with that, it just would take weeks and weeks and weeks and weeks. And the ability for AI to ingest and transcribe images is pretty incredible. So that was something where at first we thought, well, this would just be a service that we provide and maybe we'll mark that up a little bit. And in the end, it's like, Here's a tool. Here's how to do it. It actually does a really good job. You still need to obviously spot check it, but it's just wildly more efficient. I think where there's now a question among agency owners is how you actually bill and price for things that you're using AI for. I'm doing market research and I can create a report that would have taken me 20 hours to do. But maybe it takes two hours with prompts and refinements and drafting and editing and formatting all that stuff. Do you charge two hours? Do you charge 20 hours? Do you charge 10 hours? How do you even make that value proposition? And I think the thing that we've been grappling with is with any major technological shift, there's going to be a period where people think, oh, well, that means just everyone can do it. The reality is, Still taking tons of experience and very limited, you know, specific knowledge set and saying, we have such depth in this area that if we use AI, it's going to be different if you're using AI or a random person is using AI or even a client is using it, right? How do we value that ability to put in our input and get different things out? I grew up, my mom was a librarian and introduced me to, you know, AltaVista and Excite.com and things in the 90s and what Boolean searching was. And to this day, even though I've been, I consider myself more or less a digital native, she can still out Google and find information because, you know, she has an entire depth of knowledge and career in finding information, right? And some people are going to be better than others. And so I think that's the new A challenge we have is we can use AI in a lot of different ways, but actually how do we account for that? How do we decide how much value is created? And I think that's going to be kind of a line in the sand between agencies that aren't adapting. Speaker 2: So how are you guys charging for, you know, that let's use that example that you used for the two hour report, right? Are you charging two hours or are you charging what? What have taken without AI the 10 hours and the expertise? Speaker 1: Yeah, I think for us, it's been more on a sliding scale and I'm rounding down a bit, especially in these early days of figuring this out. But I still am adding a value-based equation of my assessment of the output. And whether it's good or not, whether it needs to be refined or not, how it gets refined, whether this is actually quality enough and accurate enough to use is still the value that we bring, where anyone can, you know, use garbage in, garbage out, or put something smart in, but have it I actually put something smart out and that our depth of expertise is really the question are how we use the data, how we input data, how that's sanitized, our assessment of data security, all of those things. It's something that's still this bill that we bring, right? But I do think We have already talked with folks who are like, I'm just not going to use it. Speaker 2: I think that's a huge mistake. Right? Speaker 1: Yeah. Cause I remember when responsive web kind of came out, right. And it's like, Oh, there's just agencies that never started building mobile sites or never figured out how to do that. And you always have to upscale and you always have to be changing. And I think, you know, platforms are trying to do more and more even with Shopify. There's so much more woozy wig auto configuration, things that. For some shops, their bread and butter might be doing maintenance updates that are now kind of obsoleted because the user can do those and there's no gate keeping, there's no barrier to entry technically. So I feel like similar with AI, it's like technically anyone could use this and put an agency out of a job. But on the other hand, no one else is doing as much every day in this field as we are and you're still going to be the subject matter expert, right? Speaker 2: Hey, agency owners, if you're sticking with one platform for every project, you're probably stuck in a growth bottleneck. More clients means more hires, which adds noise and cuts into profits. To break the loop, you need a more flexible tool that doesn't stretch your resources. Now, Wix Studio is a smart addition to your agency's toolkit. Intuitive by design, your team can quickly master the platform and focus on the work that matters the most. Now, with robust native business solutions, Like booking, e-com, and events, you can take on any project at any scale without adding cost to third-party plugins. Plus, Wix Studio is a low-maintenance platform, meaning you can redirect client budget toward real growth initiatives. Think more value for your clients, steady income stream, and stronger relationships. Scale your agency at wixstudio.com. Well, I think people need to adapt. It's kind of like you look back at farming in the 1800s, right? Or the early 1900s when the tractor came out. But John Deere used to sell wagons for what I remember. I'm not from the 1800s, but from the research that I know a lot of people may think I'm from the 1800s. You have to adapt. And I always believe in charging for expertise. I don't care if it's kind of like that old, that old saying, the refrigerator repairman comes, you know, to fix your refrigerator. That's not working. He literally takes two seconds and it's fixed. Then he gives you a thousand dollar invoice. And they're like, what the hell took you two seconds? Why are you charging me a thousand dollars? He's like, let me break down the invoice. You know, here's $100 to turning the screw, but here's $900 for knowing which screw to turn, right? Like, I don't care if I can do something really fast. That's why I hate when people charge by the hour. Because as you get more efficient with these tools, AI, you get smarter, that kind of stuff, you have better processes, you're going to make less profit. Yeah, you won't lose money, but you're making less profit. And they're not paying you for all that learning, that expertise. That's kind of why, like I always tell people, I'm like, pay me for my expertise, not my time. Speaker 1: Yeah, for sure. And I feel very similarly in that, you know, I've been developing for Shopify for so long that what takes me five minutes to do I've seen junior devs take an hour or two hours to do and, you know, what is the value in that? I mean, we still have a hybrid model in our agency because We do a lot of takeover projects, a lot of reverse engineering, a lot of debugging and sort of refactoring. So for us, it kind of de-risks us to do something hourly versus we thought this unraveling this, you know, API and what someone else did with it would take 20 hours and it takes 60. Right. But I do think that there's a time and a place where this is cannot You will undervalue yourself if you're only charging by the hour. Speaker 2: Yeah, I want to talk about the processes you guys put in place. You know, I was, I was chatting with our mastermind community and you know, AI is always a big topic that we talk about. And there were some agencies that have clients that the clients actually made them signs that you cannot use AI, right? These are more kind of government, healthcare, some of in those industries, like even from using a meeting recorder and using the transcription, they get like all that kind of stuff. So what was your procedures that you developed for your agency around AI? Speaker 1: For your contractors and your employees, I haven't specifically sort of blocked any tools except deep sea because it's so obviously just feeding into something that's insecure. And, um, you know, but otherwise, uh, our policy is sort of, if on your own time with your own information, that is nothing to do with in gear that you want to play with AI and then report back and share, like, we're all about learning new things and there's going to be stuff that's just not on everyone's radar. You know, be an early adopter. That's cool. But for anything that's client data, it should not be put into any model that is not specifically sandboxed on a localized environment that says, you know, this data does not get fed into learning models and so on and so forth. I've used it a lot internally for My own copywriting and ideation, I've used it for proposals to refine them and reword things. How I think of it and how it comes out on the paper is not necessarily actually the best way to say something or the most intuitive. Found really interesting kind of depth there where it's like, okay, this is actually helping me say this better in a way that's going to be more digestible and therefore have a higher rate of return. Yeah, I mean, you mentioned content, agencies, copywriters in the beginning. We do e-commerce, SEO, and we certainly were writing copy for folks. And we were transparent in the fact of we are using AI in a process of drafting, but it's never the final product. And there's fact-checking that has to happen. There's grammar. There's the weird hallucinations where it'll go off and do other things. It really has to be something where we feel it's an additive process that is helping us save time, which then in theory I save the client money or improve the product rather than something that we're trying to do as a crutch or anything else. But I do think overall that the way people shop for, consider, and purchase content services and content marketing has now forever changed. That is absolutely different. We've definitely seen a decline in investment In SEO, because I think most people feel at this point on the client side, well, I can just go to GPT and tell it to write me ties on matters. Unknown Speaker: Why would I have you do that? Speaker 1: I still think the output. On a raw basis is just not there and it's not truly optimized in a way that will actually be helpful necessarily. But I understand that impetus and I do think that's affected the value proposition pretty significantly in that. Speaker 2: Yeah. And it's coming. AI gets better every, every single day. Last year when I was using ChatGPT. Yeah, I loved it. And then I hated it. Like it would go through different versions. I was like, this is complete garbage. And then, then it's even better. And it just, agencies need to really kind of adapt fairly quickly and kind of be, you know, scared of it. Like I'm constantly looking at new tools. I'm constantly telling my team, like, use it. Like this is, you know, one of our core values at Agency Mastery is just being resourceful. And if something used to take us an hour, if I can make it take us five minutes. We're all about saving time and not getting trapped in that prison that most agency owners get trapped in because there's so much to do. We don't know what to do. Speaker 1: Yeah. No, it's actually completely affected my schedule and how much time I'm spending on things. We've been doing something also where it's essentially creating a parallel track. Hey, we're going to spend the time, do things the way that we're used to doing them. I want to put that, whether that's a design, maybe it's a strategic report or a UX audit or something else and, you know, bill what we're going to bill. We're going to follow what we pitched and just crank on, do projects, but also we're going to take some extra time and just see what would happen if we actually did that using AI tool. What if we had the AI tool look at, you know, a screen recording or screenshots of the site or analyze how the card drawer interaction is or so forth and give feedback on that. Oh, does that change what we would have said? Should we actually now change something we did, but like have that as an option to sort of A, B test our own processes versus efficiencies? And sometimes they're like, yeah, this is, it came up with these 10 things and they're just not. They're technically true, but those are not things that move the needle for revenue, right? It's not wrong, but it's not adding value. Or there could be other things in order. Well, I never noticed that aspect. That's something that is actually adding value, right? And so I think that's an interesting way to go about it. If you do have the extra time is not necessarily throw everything out you've ever done, but to start doing internal comparisons and saying like, this is just Our R&D, we're not billing this to client. This is not something that other people are paying for. We're just saying we're going to test our assumptions on what we think we can do versus what AI can do and see if there's an efficiency. Speaker 2: Awesome. Well, cool. Well, Paolo, thanks so much for coming on the show and sharing how your agency uses AI. And for everyone listening, if you guys enjoyed this, make sure you share it out. And like the show, and if you guys want to be around other amazing agency owners and really be a part of a community that understands where you're at, where you're going, so you can actually get there a little bit faster and a little bit more confidence and have that support network, we'd love for you guys to chat with us about joining our community. Make sure you go to agencymastery.io, click the little scale button at the top right, and we'll have a conversation. And until next time, have a Swenk day.