286 | How to Automate Anything with Browser Agents & Zero Code with Chris Daigle

0:00Hello and welcome to another live episode of the Leveraging AI podcast, the podcast that shares practical ethical ways to leverage AI to improve efficiency, grow your business and advance your career. This is Isar Maitis, your host. And in today's episode, we are going to dive into one of the coolest capabilities AI gives us today. And that capability is actually a huge unlock to anybody who wants to start building slightly more complex AI solutions than they're building today. So if you're just getting started on your journey of building advanced AI solutions, or even if you got started already and you're building a few things and you're doubling with it, you know that there are technical aspects that you just have to figure out.

0:47These things can be setting up an OAuth 2 to whatever system or creating a project on Google Cloud to host whatever it is that you're trying to do or setting up webhooks to different tools that you want to connect to. And to be fair, I don't have a clue what an OAuth 2 means. I know roughly what is required in order to create a Google Cloud project. And I definitely do not know how to set up webhooks, but I have done this multiple times. And the way I'm doing this is I'm allowing AI to do it for me. So if you're not a highly technical person, these things probably scare you. Like, I don't know what that means and I don't want to deal with it.

1:27But in many cases, once you start building more advanced solutions, it becomes a necessity. The good news is today, our guest, Chris Daigle, is going to show you exactly how to do this. He will take you step by step through the different things and how to use AI. and to be more specific, how to use an AI browser in order to set up literally any technical thing you need to set up in order for your AI automation and solution to work. Now, Chris is the founder and the CEO of Chief AI Officer. And in the past three years or so, he has been helping individuals and teams and companies develop strategies and the tactics on how to implement AI successfully in their businesses.

2:13Now, because of that, Chris has a very solid understanding in both the strategy and the tactics. And this is why today's episode is going to be a goldmine to show you exactly how to combine these two things together, the strategy around it, as well as the tactics and how to make it work. And like I said, on things that are usually big showstoppers to people who are not technical people. And this is why I'm very, very excited about this episode. And I'm excited to welcome Chris to the show. Chris, welcome to Leveraging AI.

2:47Chris Daigle:Thank you so much for the honor and opportunity. So for those of you, yeah, for those of you listening, I just want to let you know, Isar was the first like AI person that I met in my career. I don't know if you remember when I was in Orlando. I did. We met face to face in a nice restaurant in Orlando. Yeah. Yeah. Seasons 52, I guess it was. But anyway, so we've got some we've got some history. Yes. Yes. So, yeah, you know, you and I do very similar things. We help companies in slightly different ways, but but to achieve the same goal, right on how to become an AI first company and how to be grow a business around AI and how to remove blockers from whatever you're doing right now.

3:31by leveraging AI to do that. And that's why I think the topic we picked today is really, really awesome because I think a lot of people, whether they're solopreneurs or in big, large businesses, these are huge showstoppers. Like, I just don't know how to do this thing. I don't even know what it means. And yes, you can go and watch 25 YouTube videos and try to follow them step-by-step. In many cases, it will work, but it will take you three to four hours instead of letting AI do it in five minutes for you.

3:57Chris Daigle:Yeah, for sure. And I think that it is important that this does apply to the solos who want to be building their own stuff or product or whatever, or the executive all the way up the chain that's like, I just don't like doing this part of my job. It'd be cool if I had an app or an automation that did this, right? So I'm not a technical person. We don't define the role of a chief, as a chief AI officer, as a technical person, I see it more as an operational person, right? So that's the lens that I'm sharing this with you. And we don't have a, I've got a small tech team, I guess, for our own company.

4:31Chris Daigle:And we've created this culture of training people to be their own builders, right? Now, not for something that's necessarily going to be attached to sensitive information or, you know, products that we're selling, but to self-serve on solutions when it comes to, this is a task in my job or, you know, the contribution that I want to make, the deliverable I want to produce. it would be cool if I, and maybe a GPT don't cut it, maybe a prop chain don't cut it, right? It's time to create something. So one of the first experiments that I did with this, that's going to result in this framework I'm going to share with you today, was we sponsor and speak at a lot of Vistage events.

5:13Chris Daigle:And if you're not familiar with Vistage, it's like a local CEO peer group. You know, people that take their businesses seriously enough to invest in their, you know, their peer group. And I'm very curious, like, on collecting data, what's going on on the front lines of AI like that, that I know what I know what Esar and I are up to, but we live in a bubble, man, like stuff that we're thinking about and talking about, like, you think everybody else knows the stuff, but they don't. So, you know, we started with a survey, a Google form, right? I mean, for an AI company that's, I guess it's a little like indie or whatever, but not odd brand for how I want to present.

5:55Chris Daigle:But it works. It works. It's effective. It's free anymore. Yeah. And as they say, version one is better than version nine. Agreed. But I wanted to not just capture the information because then it's stored in a spreadsheet on a, you know, and I got to go, what drive folder was that in? And, oh, I got to share it with somebody else. And the, the extraction of intelligence from the multiple entries was cumbersome. So what I wanted to do was, but, and, and also like, it was like, thank you. At the end of the result, there was no real value for the, the participant other than just to, you know, participate in the survey and the survey of collecting the data.

6:36Chris Daigle:So I said, may it be great if we had some sort of like a, an opportunity audit or a risk assessment or something where we could ask those same questions, but there was logic behind it. And there was, you know, based on all the choices that they made, that there was a recommendation that was, you know, oh, that's pretty good. You know, let me take that back to the team kind of thing that they could get on the spot. Yeah. Okay. I'm not building that in a Google form. So I found initially one of the easiest places to build was Google AI studio for me. Right. And do we do screen shares in this? Are we just talking through it?

7:14Chris Daigle:Yeah, 100%. So I'll say two things before you jump into the screen share. One, for those of you who are joining us live, first of all, thank you. There's a bunch of people both on LinkedIn and on the Zoom call. So first, introduce yourself, say where you're from, what are you trying to get out of this so we have a better idea on who you are. And obviously, if you have any questions to Chris or myself, please write them in the chat. I'm monitoring both the Zoom chat and the LinkedIn chat. So I'll be able to answer these questions or deliver the right questions to Chris at the right time. If you're not here live, the question is why?

7:49We do this every single Thursday with really amazing people like Chris who share really amazing knowledge. And if you are not here, well, you can still listen to us on the podcast or watch us on YouTube, but you can't ask any questions. You can't engage in network with other people. And so come and join us. We do this every Thursday at noon Eastern time. and there's a link in the show notes. So if you're just listening to this after the fact, you can use the link to join us in future episodes. The last thing that I will say, and then I'll give it back to Chris, is we are going to explain everything that's on the screen.

8:22So if you're not watching the screen right now because you're driving the car, walking the dog, washing the dishes, whatever it is that you're doing when you are listening to this podcast, it's fine. We're still going to explain everything that's on the screen. Yep.

8:35Chris Daigle:And I'll second that endorsement. Like I'm an AI professional. I'm busy. I know this stuff, but I listened to Isar's podcast. So I appreciate that. Thank you. Yeah, for sure, man. There's high signal, low noise in this environment. Okay. So I'm going to share my screen and I'm going to show you what you're going to see is essentially the front end of Google AI studio. We're a Google workspace business. So for us, this is, I mean, it's free, right? Well, you're paying for it somehow, but it's not like an extra charge or anything like that. So if you have a Google Gmail account or whatever, if your business runs on Google Workspace and you're not using this, that's bananas.

9:15Chris Daigle:Okay. So what we see here is essentially the front page of Google AI Studio. And like most large language model or generative AI environments, it's got the dialogue box where I can put in my prompt, right? So very easy. And it's got the, it almost looks like a custom GPT with the suggestions and all that kind of stuff, or maybe a gym, more appropriate. So let me take you into some of the stuff that we've built. So where are we? My apps, there we go. So I'll run you through an example. I'll say one thing, just one thing before we dive in. So most people in the Google universe probably know Gemini.

9:55Think about this as more of the geeky side of Gemini, right? So you can do the same thing you can do in Gemini, you can do here, but you can do a lot more, such as develop entire applications or analyze entire videos or do more detailed things than you can do in the regular Gemini. And the only difference is that it actually uses the API, which means you're going to pay for tokens versus in Gemini, it's kind of like included in your membership. So that's the only difference, but you can do really advanced, cool, amazing stuff with a user interface that looks like a regular chat. Mm-hmm.

10:34Chris Daigle:So for me, going back to that AI opportunity audit, let's call it, right? And just let's fast forward. The tool that we're using is now like a standard part of our sales process. So it started with a Google form and a good idea. I had this idea that, you know, I wanted to be able to ingest the information from the people, use AI to process it like on the spot and give them a recommendation. Hey, based on your answers, It looks like you could benefit from here or, hey, there's risk here, whatever the case might be. We've now turned that into a kind of like part of our discovery call where we run this.

11:13Chris Daigle:So it wasn't just like a cutesy vibe coded one off kind of thing. It's it's now got commercial application. But, man, I've never coded anything. You know, I had a cursor account maybe in September of 2024. And it was cool watching like all the codes show up, but I couldn't ever get anything to work. It was a glimpse. And I think if you're involved in AI, you don't get discouraged when something doesn't work because you know that it's just a matter of time before it's working really, really well. So I was excited to see that stuff. But again, I'm not a coder. I was like, I'll get some use out of it.

11:53So with Google AI Studio, I just kind of went in there.

11:57Chris Daigle:and like with any of these tools now, I guess I don't use Lovable or Base or any of those. I don't need to use them that often. I just told it what I wanted, kind of like a prompt, right? And I gave it a link to the Google form and that sort of thing. And so that started a process, an iterative process of me and the brain, I guess Gemini 3.1 or whatever version we were using at the time, going back and forth in iteration. And like best practice, hell no. But when I first got started, I was like, oh, make it do this. Oh, and make it do this. There was no PRD, or if you're familiar with what that term means, there's no product plan.

12:37Chris Daigle:It was just as I was going through it, and I would encourage you to do the same thing, like eventually move to the professional stuff. But how you learn about using these things is by using these things. And if there's some intimidation about, well, I don't know how to use it, just go in there and be dumb. Hey, Jim and I, I don't know what I'm doing, but I'd like to build an app that does. I have no, you do that. And what happens is, you know, repetition is the mother of skill. You start to learn little bits here a little bit. Ooh, next time I'll do it this way. Oh, I should have asked for this.

13:07Chris Daigle:Right. So I went through that process and, but where I kept getting stuck was, okay, great. I'm collecting the information and I'm, I'm spitting out some info. But, and one of those ideas was, oh, let's send them a report. Seems pretty easy, right? Like, okay, we've already got the information on the screen. Let's send them a report. It's not, If I wanted it to be in PDF, it needed to be, there was stuff that was happening on Fire, whatever Google's database are, Firebase. And there was all these things that were Google developer tools. And I was just like, man, I cannot let, I'm on to something here.

13:46Chris Daigle:I can't let that stop. Now, go try and find talent on Upwork that knows how to use these AI tools. I mean, everybody claims they do, but like, you know, Isar and I, we have a discerning palette when it comes to talent on AI. So to find the right person on like Upwork or something like that ain't happening. I'm a startup. I'm not going to be able to pay US-based, you know, New York City vibe developer rates. So it was up to me. And one of my mentors a long time ago saying that I use all the time, if it is to be, it's up to me, right? So I had to take responsibility. if I wanted this thing to present my company well, if I wanted to capture the information so that I could action it, if I wanted to make sure that the prospects that were on sales calls with us were having an experience, like a wow experience, I was going to have to figure this out.

14:41Chris Daigle:And it was about that time that I started getting into the Comet browser from Perplexity. ChatGPT has Atlas. I don't use Atlas much. I'm a pro user or max or whatever it is at ChatGPT,$200 a month. So I'm a power user of ChatGPT, but I don't use their Atlas browser. So I've been a big fan of perplexity since the beginning. I've been teaching people about it since 2023 or whenever it came out pretty much. So when the browser came out, I was particularly interested. It makes sense. I could do my little perplexity stuff on deck. I don't have to necessarily context switch and go here and forget stuff.

15:20Chris Daigle:But it was this thing right here. If you can see my screen, it was this assistant button up here. And if you're not looking at the screen, it's a Chrome-based browser. So it looks very familiar. All your plugins that you have in your Chrome or whatever, they work here because it's Chrome-based. But Perplexity is giving this browser away for free. And one of the things that comes with it is in the upper right-hand corner, we'll call it a browser agent. It's listed as an assistant because most people are like, if I were to go to a YouTube page right now, let's try that actually, just so you can see.

15:55Chris Daigle:It's more like, you know, speeding up the process. But if I go to assistant when I'm on the, let's take a look at one.

16:05Okay.

16:06Chris Daigle:So the instant or instantly it's suggesting, you want me to summarize this video, extract key takeaways, scroll to the next interesting moment, right? Helpful for sure. But, um, if I want it to do something, if I want to, if I want the perplexity browser to take over my, uh, browser and perform actions for me, it can do that. Um, I'm not going to give away some of my secrets, but like, for instance, oops, I hope nobody from LinkedIn is watching, but I, I, I use, um, and I had, you know, a couple hundred messages in my DM that I just, I love LinkedIn. I love, you know, interacting on it, but the DM thing, I'm just, I'm bad at email.

16:51Chris Daigle:I'm bad at texts. I'm bad at DMs. So I've now created something to where the perplexity browser will take over this particular tab and it'll, it'll explore the messages that are coming in and prioritize them based on opportunity and that sort of thing. Very helpful because otherwise I just wouldn't be able to do it. But now I take 30 seconds in the morning, give it, give it its instructions and then off to work. I go, um, by doing that sort of thing, it hit me. It was like, Oh, wait a minute. I need the OAuth too. I need the PDF like engine tied into a hosted blah, blah, blah, blah, blah. Right.

17:29Chris Daigle:I didn't even know what those things meant. I mean, I understood conceptually, but I didn't know all the stuff. And as Isar pointed out earlier, my options are, you know, Plexit, I guess, go to Plexit. Hey, Plexit, how do I do this? Right. I could get the instructions from the AI studio. It was happy to give me those. Hey, give me those step by step. But even doing that, like the time just wasn't there. It was, it was, it ended up being projects that were like, let me get, you know, 45 minutes now. Oh, I got to do this call. Oh, let me try and get to it later and later and later and later. And it was just an open loop.

18:06Chris Daigle:And it wasn't because I wasn't clear on what I wanted. It wasn't because I didn't have something that was viable or marketable. It was those technical skills. And I didn't have the budget or the staff, the team, the internal knowledge to be able to do those things. So it hit me. Hey, Gemini, or hey, AI Studio, I'm going to be using the Comet browser agent to do this for me. Can you give me these instructions in a way that I can copy and paste them to the browser agent and let it cook? And the entire technical world opened up to me. I'm being a little hyperbolic, but using that, it was able to set up the OAuth 2 and test it.

18:57Chris Daigle:And all the things that I would have paid a US-based person, I don't know, maybe a hundred bucks an hour or more to do, maybe more if I could get on their clock or hours of my time. And since then, now, anytime that I'm building something, and I don't build, I say I don't build a lot, but I probably build something every week. And for me, that's not a lot, but for others, and now my team is doing the same thing. Anybody who's on my team that has that idea, they already know how to do what we call think in AI, right? But now they're thinking in builds because they have access. Everybody in the company uses the comment browser here and the agent to help them build these things.

19:39Chris Daigle:And now with clients, we're building MVPs. Any of our chief AI officers are now building MVPs in a couple of hours. And before it was like, okay, get your requirements together. Let's get it over to the dev team. and there was just this big lag time. I have set stuff up on Cloudflare. I have, you know, everything inside of the Google API environment. I have done it, I guess. I mean, I've led the doing of through these agents. And here's the beautiful part about using this is that, okay, I've got this tab open. I want to do something with setting up a database so that I can store all of these answers and I can ship them out to people when they put in their email address.

20:27Chris Daigle:I can capture the opt-in and all that stuff. A lot harder than it sounds when you're building from scratch. but the the the gap between me having the idea me being able to launch something because of this particular tool has been compressed to a 30-minute session no kidding i mean now again am i saying that your first time out hell no like i spent hours and you will too if you get into this stuff the first few times it's hours and it's like, you know, breaks or whatever, but it's enough. Even, even as a brand new user, there's enough confirmation from the device, from the models, from the tools that you're like, you know what?

21:10Chris Daigle:I, I can do this. Like, there's just a couple of things that I didn't know, but now I know them. Let me try again. And then you're going to land it. Awesome. I want to pause you just for one second to add a few, a few things that I think are, First of all, this is fantastic, right? So just for those of you, as a quick summary and recap, you get to a point that there's a technical aspect in the process that you did not know how to do. And the AI in whatever tool you're using, again, in this case, we're talking about Google AI Studio, but this could have been Cloud, this could have been anything else.

21:41And it tells you, oh, now you need to set up this thing. I'm like, I don't know what this is. What Chris is doing is, okay, give me the instructions of exactly what needs to be done and what it needs to be set up to. and then he takes that instructions and he gives them to a browser-based AI tool that can take over the browser. And then it can actually do the thing. So it doesn't tell you what to do. And by the way, using perplexity a lot, it happens a lot. Like perplexity, the assistant in the browser said, oh, to do this, you need to do one, two, three, four, five, six. No, no, no, no, no. You do this.

22:14Like, oh, okay. And then he takes over the browser and actually goes and does the thing. A few things about variations of this, or I say one thing that is a variation of this, and the other is the concept that I think people have to think through. The variation of this, I build most of my stuff in Cloud. And by the way, same thing for Gemini. But Cloud has its own browser controller. So it's the same exact thing, but it's a Cloud extension. It opens on the side, but Claude Cowork and Claude Code know how to control it. What it saves me is copying and pasting the instructions or copying and pasting the results back to, in this particular case, Google AI Studio.

23:01So I just work with Claude Code or Claude Cowork and I say, hey, you know what? Go and do this in the browser. And then it, on its own, will open the browser, do the thing and figure out and set up everything. And when it gets stuck, it deals with it. I don't have to copy stuff back because the brain of Claude Cowork that is aware of the bigger thing that I'm building knows what the required output is. And it goes back and forth with the browser. And I don't even have to see the browser, which is fantastic. The only times you have to go in is like pop-up screens that it doesn't have access to.

23:33Like if you need to log into something, okay, you need to log into your Google account. I can't do that. So these kind of things are the only things that require your attention. but then you just go to that browser page that's opened by Claude and you go and do the thing. So this is kind of like a cheat to get this even slightly less work. The bigger thing here is I think the mindset and I'll explain what I mean. We are humans are the bottlenecks to most digital work today, period. Right? So if you need to do a process, whatever the process is, whether it's writing a proposal, analyzing information, creating a form for your customers, like all these kinds of things, you are the bottleneck.