How to Build an AI "Second Brain" for Google Ads

The first step on this journey. I've sort of described this as a journey from ads to AI. The first couple of steps on that bridge are really automating the creation of reports.

Whether you're a business owner running your own Google Ads campaigns, or whether you're an agency or freelancer who regularly works inside of Google Ads, I can guarantee you that over the past two years, you've really thought about how you can use AI to improve the systems that you use for optimizing your Google Ads campaigns. And that's why in this video, I invited Mike Rhodes from ads to AI, where we can start to break down how you can go beyond just a couple of simple scripts to get some reporting for Google Ads, to really start building a powerhouse team of AI agents to improve your efficiency, reporting and decision making when it comes to Google Ads. But you can also use this in all different areas of your business. So it is an absolute pleasure to have Mike Rhodes back here on my YouTube channel. So sit back and relax and enjoy this interview. Welcome back, Mike. It is amazing to have you. Many people would know Mike Rhodes. I don't really need to give you an introduction, but for people who don't know, he's obviously from ads to AI and a real leader in this space. So what I wanted to talk about today is Mike would really like to get to a lot of people. AI has been around long enough now. People gone through the modes of scripts and those things. But if you're just coming in, where's a great place for people to start? And more importantly, what tools should people have to start their AI journey with?

Productivity, I think the biggest tool you can bring to the table, frankly, is, is your curiosity and your imagination because really, imagination is the only constraint now. It's like we all have access to the world's largest Lego kit and you thought there were 20 pieces or 100 pieces. There's 4,000 pieces in here and it's growing every day. All you have to know is what pieces exist and then it's up to you to put those pieces together in different ways. So I don't think the model really matters all of that much. So you've got ChatGPT, Claude, AI and Gemini are all kind of there or thereabouts in terms of intelligence. They're all kind of the same. But if you're a business that runs on Google Docs and Google Sheets, then Gemini is probably going to be you go to just because of how well integrated into the Google system it is built in down the side of Sheets, Gmail, calendar, it's just in everything. They've done a really good job of that. I tend to use Claude all day, every day in my thing I call the brain. We'll worry about that another time. ChatGPT. I don't know, I just. It doesn't quite have that sort of quirky personality for me that the others do for images. I think Gemini's ahead there with Nano Banana. Chat GPT I've got a decent one for voice. I do use the Chat GPT on my phone, the Chat GPT app on my phone to chat with while I'm walking. That's probably the best one for voice. So it kind of depends on your use case. But the thing that really elevates the output that you get from AI is the context that you give it. So think of the AI as the thing in the middle of the archery target. It's those next rings around on the archery target that really matter here. So it's all of the memories that you give it, the client communications, the call transcripts, the information about your business and you how you like to work, your preferences. If it were talking Google Ads, it's the data that you can bring to the table or giving it access to pull in whatever data it needs whenever it needs it to answer the question that you've asked it. So giving it access to tools. If you use asana in your business, it can access asana and read in stuff and update stuff, slides, calendar, Gmail. That's what really starts to make the difference context.

So I guess probably the way that I talking to people when they do ask this, and I'm assuming that you would say the same thing, is that really don't get caught up on any of those main three. What's more important is the systems, the data and what you're asking it to do, the context you're giving it, all three can do it. When you get further down the road, you might find some little nuances between. But any of those three you're going to, you're going to really get there

and get you into the game preference. At that point you said the word systems. I think that's key. You want to start thinking in terms of what's the AI system that I'm using or creating the AI model. It's not really the model in and of itself, it's what you put around it and how you choose to then use it.

Hey, sorry to interrupt. We're going to have you right back to the interview with Mike Rhodes very, very quickly. But as Mike just said in this interview that he is actually giving away a free script writer. And to access that, all you need to do is to follow that link in the description below. And when you're in the description below as well, we'll have some extra information there for you about some of the different tools that you can utilize so that you can use AI greater inside of your Google Ads campaigns. So just follow that link in the description below. But right now let's jump back into the interview with Mike. And I know, I know just us personally inside defined Digital Academy and my flow, we've started moving more and more across to Gemini. But that's just because everything that we have lives on Google Docs and Google Sheet. And as you said before, it's just a good integration for us. But once again, I'm not. Yeah, we're definitely not saying that that's

the only one just for Google Ads pros because obviously a lot of people that watch this channel are playing with Google Ads. The first step on this journey. I've sort of described this as a journey from ads to AI. The first couple of steps on that bridge are really automating the creation of reports. Most people really could automate a bunch of reports. They probably save five hours a week. And so the tool for that is Google Ads Scripts, which I know we've talked about before, but using Google Ads Scripts and you don't have to learn to code, you don't have to learn to write. All three of those models we just talked about can help you write those scripts. But again, you have to give it context because yeah, out of the box you ask it for something, it will come up with some sort of fairly decent thing which may or may not work. So I've got this sort of 20 page document that describes what a good script is and how to write a good script. It's got a bunch of examples in there. That just means the AI gets it right first time more often. I started giving that away a couple of years ago. Now I've built a tool. It's the script Write a tool, I give it away for free. And you can use this so that all you have to do is type in one sentence the script that you want to create, the outcome that you want. It then bolts on this 20 pages and you can then take that to whichever AI tool you prefer to use and out will pop a script out the other side of that.

And just for anyone listening is as much of that script writing tool, it is something he's giving away for free. If you want to access that, just follow the link in the description below. We'll put, I'll put a link in that where you can access it. And while you there as well is that this conversation that we're having, we actually had a much larger conversation inside of Define Digital Academy, which is a masterclass which Mike shared. If you want more information about how you can access that inside of a Google Ads Mastery program, just also go into the link below. But let's get back into where we're talking from before. So where I'd really like to take this conversation to because I know a lot of people watching this are going to be coming in from that Google Ads management. Obviously we're using this as the starting process to really look to where you can, you know, automate everything. But let's keep it on Google Ads. So for the business owner, this could be an actual business owner running their own ads, could be a freelancer, could be an agency, someone living in the agency world. If they still see themselves in the glue in terms of their Google Ads processes, like they're the ones that doing all the everything optimizations, reporting all of those things. What would be the first steps that they should take to start building out some of these AI systems to save time and efficiency?

So my mental model for this is sort of a hierarchy of going from data to information to insight to action. Because your clients actually don't want information. They certainly don't want you to give them a Google Ads dashboard full of stuff. What they want is the insight from looking at that information, applying your intelligence to it and giving them those insights. They don't really want the insights. What they want is for you to go take action based on those insights. And sometimes they need to take action. But actually there's one more step. They don't really want the actions. What they want is to know the business outcome. What's going to happen when you take those actions based on those insights, based on that information. That's what they really want. Basically. Of all these things we could do, what's going to move the needle the most in my business? And so we need to start if we are still doing everything manually, we just need to start by going from data to information first. So that's Google Ads scripts. How can we use scripts to pull the data out that maybe we're still doing manually? Format it nicely, make it look nice, put it into a Google spreadsheet. The next step then, then we can start to use AI to go from information to insight we can actually have the AI given enough of the rules and the context and all the stuff we talked about before. Once we teach it about the account, it understands the constraints, the limitations, our preference. It could even understand every bit of communication you've ever had with your client, every email, every slack message that is possible today to give it that context. So now it understands what's going on, it understands change history. Now it can give you really useful information about maybe what you should do next or why a particular metric dropped or a suggestion for a new campaign type. If you ask a general AI out of the box, you got a chat GPT today and say what should I do in my Google Ads account? How should I optimize? It'll give you a very plausible but horribly wrong sounding answer because it doesn't know all of that context about the account. But given enough context, then the answers become useful. I'm not saying you should then action that immediately without checking it, but it gives you 80% of the insight. It solves the blank page problem, gives you a really useful starting point for you to then go apply your knowledge and wisdom too and say, well, actually for this account I won't do that. But those other three ideas are really, really good. Yeah, bonus points for that one that you don't want to do because of the thing that you just thought of. You feed that back into the machine and say I'm not going to do that because I forgot to tell you about this, this and this. Great, now I've got more context. I won't make that silly suggestion next time. And it just gets better and better over time. So it learns from those interactions if you keep giving it more context.

Yeah, I know for me that's, that's what really, I really see it. Many of you watch this channel know that last time we had Mike on, we released a script that Mike, when I say we did it in partnership, I very much gave you the information and you, you and your team created which was the, the smart bidding script. But what I found with that, and this is what I want to pick up from there, is that that saved me so much time, but not only time, mental energy because it would give me a red light or a ready or a weight information as to whether we're happy to go from maximized clicks to smart bidding or whether it should change CPO the target rice and the way that after we started using internally we would see that if it gave it a red light, the red light was 100% right. And they say no, it's not ready yet. So I didn't have to go there. If it gave the green light, we would go back and check that. I even found it wasn't me. Saving time was saving mental energy because I hadn't been looking at six weeks of data across multiple campaigns. I was just going to the ones that, that the AI, the system we've created went bang. You need to look at these campaigns across the. We were running. I remember we'd wake up and do like we could do five accounts. It was probably across 40 different campaigns. So it's more accounts that. But it's four different campaigns and it would highlight the two campaigns that I need to look at. It saved so much time. But it wasn't just the time. It was the mental energy that I found that we were making better decisions and ultimately getting better results for those campaigns.

And that's just one aspect of what you would look at.

Right.

Because then, yeah, you can do another one of budget pacing. I was just chatting to an agency a couple of days ago for them, lots of smaller accounts. They were going every account, checking through all of their tabs. I know they've got all of their tabs that stay open on their browser all the time. And every morning they go through all of these things and look at everything. And I much prefer, I don't want to have to go anywhere for information. I want it brought to me and simplified and management by exception. If everything's working, show me a green light. Great. Like you say, mental load, just way down and only tell me about the stuff where I need to get involved right now. As long as everything else is in the green, Fantastic. It just. You come into work. Okay, I've got energy and I still have energy at 11am I haven't been dragged down into the weeds doing all of this busy work that is so easy to automate. Once you know it's possible again, you don't have to understand how. You just have to know that it's possible and then go chat to an AI or you know, join your group to find out how to do that. The how is is is very, very simple really because we just, we can ask AI to, to build most of that for us given enough context.

We, we had. One of the key things that is obviously very important for Google Ads is all about copywriting not only inside of your ad copies, but also what's happening on the landing page. We had a guy last year who had a copyright, came, did a masterclass, spoke on the channel, did all these things and he said he's, he's created an automated system so research that used to take him about two days because he was very, very deep into copywriting. So he would do research and he would research, chat forums, Facebook groups, reviews, a whole thing to really get down to not only what are real users saying, what are the unique selling points, competitor analysis. And he said that would take his team two days to do. Did. Now on one night when he gets the client, he runs the script and then in the morning it's ready to go when he's fresh and that's when he can now actually get to the strategic thinking of this is all the research. Now I'm going to put my client knowledge and my knowledge on top of it. So I'm just getting straight to the creative side. And I think for me that's the one thing that I've really found as we've started to integrate more and more into it gone that you know, the changes they took like the analytical research brand, the creative brain AI is really, really great at taking that analytic research for you that load so that you can really get into the strategic creative decisions that really move the account forward.

Yeah, I call this being extra human. The goal here is not to replace humans, to get rid of people, to fire juniors, to stop hiring. It is about getting the humans to not have to do that busy work to reduce the mental load, to save all of that, to bring all of that wisdom and expertise and knowledge, but to bring it to bear on the important stuff. Not on fitting around with a Google sheet, changing column width and bolding stuff and, and pulling data in manually. That's a horrible way to spend time in any job. Nobody should be doing that. You should be minting people.

No, no, no. It's not only a hard way of doing it is that it's actually really, as long as you've got that system set up correctly, it's not going to make the same level of mistakes and it's going to follow the system. So it's actually something that can be more reliable for those tasks?

Well, yeah, I mean we've all hired people that have made mistakes. I make plenty myself all the time. It can be more reliable, but it takes time. You've got to onboard it. Let's, it's, you know, on day one, it is the dumbest employee you'll ever have. But as you train it, as you teach it new systems, as you correct it along the way and give it feedback, it can change those systems. So it then remembers your new way of Doing it and gets it perfect every single time. There's, there's a balance here between AI and automation. Automation is if this, then that do this thing exactly the same way every time. So we can use AI to create the automation but then also to run the automation. So when the automation gets a bit stuck, it can have a chat. Instead of coming back to you, it comes back to the AI and says, oh, I got a bit stuck here, what should I do? The AI can make the decision and then allow the automation to carry on so it doesn't keep getting broken. And then at the end it pops back up to you and says, is everything you asked for.

So I think this is, I think it's the fourth time you've been on my channel now and the first, especially the first two times we're talking about your, you know, your now famous performance Max script. So there was very much scripts around doing individual tasks or data sets around individual things. What's the difference between. So some people, a lot of people watching this going, yeah, I use scripts but what's the difference between scripts and processes for one off tasks versus building What I've heard you, you referenced in the masterclass, we ran a second brain that connects all of your marketing data together.

It's night and day, so you can have scripts that, that automate parts of what you do, but you're the glue that still sticks all of those things together. You're getting a thing that runs and then you go look at the sheet and you decide what to do and then you go over to somewhere else and you put some notes to Asana and you tell the client about it in Slack and then you dive into the Google Ads account and take some action. The second brain can do really whatever you can dream up anything that you do sitting at a computer. I now spend all day, every day inside of my brain. Now again, I'm not actively managing Google Ads accounts day to day, so it is way, way, way more than Google Ads. Creating content, drafting emails, adding items to my calendar. I was reading in my membership area this morning, members using it to manage the healthcare of their elderly parents, plan inheritance, do expense items, plan photography trips with detailed information about where to go and what to see and how to use the camera that they want to take and it's, it's giving them step by step, it's teaching them anything that you want to learn. Really at this point it's a wonderful time to be a curious person. But all of those tasks that you do in your business, sitting in front of a computer you were mentioning about hiring when we were chatting earlier. It can write the PD for you, it can write the job ad, it can write the interview questions to ask in the interview. Then you have the interview and you record it. You give it the transcript and you say, one, is this a good person? Are they a good fit? Two, can you design a little audition test that will take them half an hour for them to prove that they know what they just said they did? And then also, can you. Now I'm going to hire this person, tailor an onboarding guide specifically for this person for this job to help them onboard into the business so that they understand or whatever you can dream up. It can help with.

Partly what we're doing behind the interview process is we're also creating a grading scale on a chart which gives each candidate an actual score. So it's reviewing the transcript, then reviewing. We've obviously got. With. I won't say what assigned scores we've got for the response, but we've got it from there. And it's giving us a number of who they feel is the more ideal candidate to move on to the next level.

So you're turning data into useful information so that you can apply all of your wisdom to have the insight. Let's say you were some crazy business that's hiring 100 people a week. You would then have more data and more context. You'd feed that into your second brain. It would be able to have those insights for you. Given enough context. It sees the patterns in all the stuff that you've done. So now you don't need to create the score for you, the human. It could then start saying, no, I know some people will go, well, this sounds dystopian to me. Now the AI is deciding who to hire. Obviously, you can put a human in the loop wherever you want, whatever you feel comfortable with. But the more trust that you build in the outputs from the machine, the more you allow it to carry on and do that part of the task itself. Next time you don't. At the beginning, you say, show me what you've done. And you give it feedback. Show me what you've done. But once it does it perfectly, 10, 20 times in a row, then you say, I don't need to look at this anymore. Could you just carry on and just let me know if something breaks?

Yeah. Now we're still. I mean, ultimately the still final decision is with me and the rest of the team.

But it's.

It's taking me. Yeah, it's taking some of the early level emotional, now of do I like this person? Because we're feeling as a good connection and now the connection is important. But what I want getting the AI system to, to judge for us is are they actually skilled to complete the job? Because even if there's a really strong connection, I still want the best candidate. So it takes those biases out.

Yeah, that's really smart because we tend to hire people that we like and people that like us and often the right person for the job is your complete opposite. That's a very smart.

I don't want more people like me. You need the differences so that you can create a better, well rounded team. What I would like to finish up with is and it's very, very bold of anyone to look any more than a couple of months ahead. But I am going to ask you to think what do you think AI looks? And let's keep it within the realm of ad spend and Google Ads. You know what does AI look like at the end of 2026?

So I'm looking at what members are creating with their second brains and the stuff that they're building. Getting rid of notion, getting rid of super metrics, using the brain to do all of that stuff. I created these tools last year because I don't want to wait for Google to add more automation. That will only work on Google because it's. We should have to work across, right? Yeah. If the right thing to do is to lower the spend in Google and increase the spend in Meta, then Google probably aren't going to tell us that. But our own independent tools can.

Yeah.

What I'm trying to build is the self improving agency so that as you go through and give this thing feedback, it remembers it gets better next time and you can go up a level and start to do more and more and start to work on it, not in it, but just keep on working up. So this thing starts to learn. So I've got a second AI that looks at all of the output. It's essentially listening to me and if I say, whenever I say the word wrong, it knows that something's gone amiss, it keeps track of that and then it self improves, it builds a better system so it doesn't get it wrong next time. So right now I am still having to give it feedback more and more. It's self healing, self improving. That's where I think we'll get to by the end of this year it will do an amazing array of stuff again. You have to take the time to give it the context that's the bit people try and skip over because it does take time to do that. But all of these documents, all this context that lives on my computer, I haven't written any of those documents. I've chatted to the computer and asked it to write up the documentation, to format it properly, to stick it in the right place. I'm not aware of all of that. So it is the Ironman Jarvis suit. The only constraint is your imagination. So I think it will be doing more stuff, helping us be extra human, helping us just level up. A year is a long time in AI. I've never been as excited in business. I can't wait to see what is going to be possible. I feel like I have a team of six working for me every day. You were very generous before and said your team created this thing. I didn't, I just chatted to my. The AI built all of that.

Well, we're calling, we're calling them your team though. And that's where, that's where I think the language is going. Is it? Yeah. They are legitimately your team.

Not only is it a team, but it can also create new team members. Now, again, not as good as having a human. But just to illustrate this, I created something I call my Expert Council. I built one for Google Ads, one for marketing and one for the business as a whole. And I can convene an expert counter whenever I start. So my marketing one has people like Rory Sutherland, Seth Godin, Ogilvy, Eugene Schwartz. So dead people as well as alive people. Not as good as having Seth Godin and Rory Sutherland in the room. But I don't have a spare million dollars to get them in the room for me to chat to on a Tuesday afternoon. They're also available 24 7. And when I find someone else that I really like their take on stuff, I ask the brain to build a Persona with freely available information. I'm not using anything private. I'm not scraping people's books or courses or anything. You could do that if you have access to it. And I create a new Persona and go, let's add April Dunford to the Expert Council now.

Yeah.

And I can not only ask them a question and have them all deliberate and feed me back an answer I can then ask. Actually, let's have a follow up question with Seth Godin and can you ask him why he said these things and what else I could do around here? It's phenomenal. It's not as good as having a person there, but yeah, I have this extended team now of some of the best experts in the world on tap whenever I want that help.

No, it is amazing. Hey Mike, as always, it's an absolute pleasure having you here. So thank you again for joining us. And remember, if you do want to get access to that scriptwriter tool and also find extra information about where you can access the full masterclass with Mike, just follow that link in the description below. Hey Mike, thanks again. We'll definitely have you on in the coming weeks and months throughout 2026.

My pleasure, Aaron. Thanks for having me on.