The 4 Most URGENT Ad Agency AI Automations for 2026

Here are 4 areas that you should be considering using AI and automations for your ad agency or freelance business. Hey, I'm Depesh Mandalia, and I've been running my agency since 2017, and before that freelancing and contracting. And in that time, I've grown my business from $5K and $10K a month up to $250K and $300K a month in revenue. But I've done it the hard way. By learning, by failing and learning and failing multiple times to bring you the guidance on how to skip the problems and start executing on the opportunities. So today I'm going to be focusing in on the 4 areas, the basic 4 areas that you absolutely should be considering AI and automation for your business. Now, AI and automation are not the same thing. Automation, to give you an example, is Zapier, is Make.com, is NA10, and all these apps that can technically do things for you without you having to do the thing. So going way back to 2017, 2018, I was using Zapier for many things, even sending out contracts and updating Slack and things like that. So automation has been around for a lot longer than AI has been around, and I'm talking about mainstream. AI, on the other hand, is the ability to use intelligent machines to actually do things for you. And here's the magic for you. When you combine AI with automation, then it's game over. Like, I'm talking about replacing roles in your business, enhancing roles in your business, giving you more headspace to do the things that you need to do. To run your business and ultimately to take you out of your business. Like, this is the number one problem that I see. This goes across all agents, all business owners, not just agency owners, but the number one problem is how do you get yourself out of your business and actually start running it? This is the difference between you owning a business that works for you and you having a job, which you call your business. Let's be honest, a lot of people do have a business, but it's a job. You can't separate yourself from the business, in which case it's a job. If you're working for an employer, you're bound by contracts, you're bound by time, and your money is limited. That's a job. So if you work Monday to Friday, 9 to 5, you get paid. If you're running a business where the same applies, it's basically a job, which means that even if it is your own business, but you're working 9 to 5 and If you don't work, you don't get paid. That's basically where a lot of people get stuck. So especially when you're in the 10k, 20k a month stage, and whether you're a freelancer, and you know, you're doing everything yourself, and maybe you get a bit of help, or if you've got an agency, and you're now getting 30, 40, 50k a month, but still, if you're involved day to day, and you cannot step back, ultimately, you still have a job, and it's very hard for you to actually grow your business. The things you need to do on your business get neglected because, especially as an agency owner, you have people in your business and you have clients to serve. Generally, they will always take priority over you, your personal health, your business health, and your growth ambitions. Hey, I just wanted to quickly interrupt this to say that we're running a workshop in a few days' time. The link is in the description. Where we'll be breaking down exactly how I build a CALM agency. So if you think about the chaos of early stage growth, of growth and instability that comes with the unknowns, I'm going to be showing you how we've built our agency towards CALM. CALM is an acronym— C-A-L-M: Clients, Automation, Lifetime Value, and Monetization. And they're the things that helped me to enjoy my agency, to enjoy the profits, to enjoy working with the clients, to enjoy working with my team, and enjoy the experience of actually growing a business and not being chained down by it, getting the benefits of it as well. So the link is in the description below, or it might be above depending on where you're watching this. Go ahead, sign up. If you can't make the live, you can grab the replay, and I'll see you inside. So the 4 things I really want you to focus on for your growth. And this is an enabler. This is not just do these 4 things and things will be cool. No, these 4 areas of your business, you have to work on them to automate and to enhance with AI. And I'm telling you this as someone who's been involved in the AI game now since 2022. It's been transformational for me. It means in my case, I grew my agency to 30, 35 people. I don't need that many people now. Many of the roles are honestly just redundant. And we're talking about Copywriting roles, technical roles, creating landing pages, analyzing ad results, onboarding clients, analyzing client results, deciding what to do. Like, AI can do so much of that. And when you enhance that with automations, you then end up with a business that is like a machine where things are just, you've got multiple cogs in your machine and they just work. So the first thing I want you to focus on is marketing. So if we look at your flow of business, you have to get leads in. That's one of the most obvious things. Then you get to a point where those leads are flowing and now you need to convert those leads. Then once you convert those leads, you know, then need to figure out how do I pipeline that whilst serving clients whilst doing the rest. Now, when it comes to AI and automations, what we have done with our business is number one, lead generation uses AI. So when we take leads into our funnel, we are assessing them with AI. We're enhancing the data with AI. We're putting the data into our own database. We're scoring that data because bear in mind, I've been running my business as an agency since 2017. We have, I think, close to 10,000 paying customers. On our database. They're not all agency clients. A lot of them are people that come through lead magnets, people that bought low-ticket offers, people that bought one-off products, things like that. But we have such a rich database that if you go back 2 or 3 years, I would probably have downloaded spreadsheets and tried to analyze and maybe copy and paste between ChatGPT and try and make sense of things. Now, My database is owned by my own app, which was created through AI, and that's a totally different story. But the data is accessed by my AI assistant, and it can read it, and it can tell me things about the data. So now, for example, from those— so the 10,000 customers, they're all paying customers. Some of the lead magnets that we've run have converted into paying customers. So from that list of 10,000, give or take, the AI can identify not only who are my VIPs, but who are the potential to be VIPs and where are people on their journey. So previously I'd have to go and segment and run spreadsheets and lookups and things like that. I don't need to do that anymore. Furthermore, the AI is doing that autonomously for me and using automation, using daily checks and telling me what's going on with the data. I don't need to effectively worry about it too much. And also when a new lead comes through, let's say for example, someone inquires, we had an inquiry last week. I think the prospect was inquiring about one of our agency services. It came through in my app as a lead. The AI Head Lead scored it. What does that mean? When you're generating leads, you have the concept of MQL, Marketing Qualified Lead. Then you have a concept of SQL, Sales Qualified Lead. Difference between the two is here's a lead that's come through marketing and they look good versus here's a really good lead that I think we can convert into business. So this lead came through as an MQL. It was lead scored and the lead score looked okay. But what the AI was doing was first of all, matching up, does this prospect already exist in our database? This one did. Then what segment do they belong to? I think for this, I think this prospect had not purchased with us before, so it went into our kind of non-purchases data, but they had interacted. And I can see from the interaction when they last opened an email that I sent, which website pages, if any, they had visited, how long they've been in the database, how frequently they open emails, how recently they opened. Emails. So I can see all of that data and the AI is constructing that into a lead score. So if you imagine that's just one lead, if you've got 10, if you've got 100 marketing leads that come through, very quickly the AI can score and prioritize those for you. So that, let's say for example, you have a salesperson or you have an appointment setter, they can very easily go in and prioritize those scores for the leads for outreach. So then you can say we've got 100 leads that come in, It's going to focus on the top 10. It's going to try and get them to book a call. So that's the first part. Second part is automating the follow-ups as well. So once that lead comes in, we use what's called a human-in-the-loop system. So that means right now we don't fully automate the follow-up process. So if a prospect comes in, I don't want that AI to start automating all the back and forth. We tried that and it's not always as effective and in a year or two years' time, I imagine it will be a lot more autonomous. But right now, there are certain leads we want to maybe get on a call quicker. There are certain leads that we still need to nurture. Maybe we want to send them some documents, like some guide on the thing that they inquired about, and try and prime them up a bit more before we get them on a call. But the AI is also learning from our interactions. So the idea behind this is we're not just using AI and saying, hey, create a follow-up funnel and here's some use cases and here's some ways of doing it. We are doing this human-in-the-loop system to teach the AI so that AI can see the interactions. So in future, the base that the AI uses is not some instructions that we've written and said, in this scenario, here's how you do a follow-up. In this scenario, here's what you do. It's actually looking at data. So it has access to the back and forth emails that we're producing for prospects. Now let's say we forward it, fast forward 2 or 3 months, the AI will be able to look at a prospect that's come in, find similar prospects in our database, and then say, how was that nurtured? What kind of messages were put in? What segment were they? What was the conclusion? And then from that construct automatically the follow-up for this new prospect. And that means the data that has been developed by us by using this human-in-the-loop system where we are actually interjecting, the AI is learning from that. So this is the difference between if you go to some of these, if you look at some of these YouTube videos, they're like, you need to train it up with documents and examples and things like that. We're stepping beyond that and saying the examples are not written documents. But the examples are actually hard factual data in the database where you can see the flow of these prospects. So let's say, for example, someone comes in and they're in a particular RFM segment, Recency, Frequency, and Monetization segment, which says they are a high value prospect. The AI will go into the database and say, are there other high value prospects that have been through this process already? How were they handled? How did they close? If they did close, what was the process? How many emails were sent? So like the AI can do all that analysis and then decide for this particular person what needs to be done. AI can do that within maybe 20, 30 seconds. It's not difficult for us as a human to do that. It would take a bit longer. We'd have to download the data and analyze it and things like that. So that's how we're using AI and automation technology to bring in those leads, to evaluate those leads and determine what needs to be done next. So that's my first tip, I guess, on the lead gen side. And that's not even talking about all the things that you probably see in YouTube, on Facebook groups and things like that, where people say, hey, you can now create an avatar of yourself and you can mimic your voice. And by the way, I am not doing this as an AI voice. This is me talking directly into this podcast mic and giving you my direct information on this. But in 6 months' time, 12 months' time, you know, a lot more, as the voice advances a lot more, these will be obviously driven by AI. And you can do the same as well. You might be sitting there and thinking, I don't really have time, or I don't have the confidence to go and do these things. Go and use the AI marketing apps. You can create an AI video avatar. You can recreate your voice with AI, which I don't think is quite there yet. Like, I've seen a lot of these and they, the videos are absolutely not there right now. Even the audio, like, I follow some of the bigger names and I know it's not them. I can just, it's just not great. Maybe you can't tell whether I'm AI or not, but there are certain things that AI doesn't quite get right about voice right now. But I don't want that to hold you back from doing your marketing. Here's the thing. When I first started doing my marketing in 2017, I hated being on camera. I had no confidence to just open up a mic and start talking like I'm doing right now. But I had to learn that. It took me a good few years to learn to be confident, to be okay with listening back to my voice and seeing myself on video. But you can bypass that, all of that. You just don't need to necessarily go through the same learnings. You can use the AI apps and start creating social posts. You can get some really good copy out of Claude, Gemini, ChatGPT. You can get great images created out of ChatGPT, out of Google Banana Nano. You can get great videos created. There's no reason not to be marketing. If I'm talking about marketing as the first layer, you absolutely need to get AI and automations in. There's so much more to it, but really for lead gen and lead scoring, I think if you're not doing it, you're losing out. You're absolutely losing so much opportunity on pattern recognition and scoring and things like that. So that's my first recommendation. The second part is, okay, you've got marketing, you've got leads, you've got clients coming in. If you are not using AI and automation for your clients, and for your client ad accounts and your client reports and things like that, you are seriously wasting time. And you're wasting resources because if you're not using AI and tech, you're using human power and While I'm not against hiring and having humans in the team, I think it's a good thing. I don't think we're anywhere near running a complete agency where humans are not required. Not because I don't think AI and automations can't replace, but it's like they can replace some of the day-to-day, but as a client, do I want a robotic business to work with? Probably not. So that's still where with my agency. I still want that human touch, but behind the scenes, so much is done through AI. For example, when it comes to something as simple as client onboarding, our AI, once we, so here's the process for you. When I'm on a sales call, I use an AI to transcribe the entire call. So I don't make any notes from the calls anymore. Once the transcription comes in, that gets passed to another AI. Which has been programmed to create proposals. That AI will then send me the proposal. I read through it. It takes me about 5 minutes to read through it and make any amends. I then send that to the client. Once that's sent, if the client agrees, then I have another AI slash automation, which will automatically send out the contract. Once that contract is sent out, the AI is— the automation is monitoring it. Once the contract's signed, then the automation together with the AI will generate an invoice and will wait and chase up that invoice to be paid. Once the, once the invoice is paid, that's when we trigger, and this is done automatically, the onboarding. So what does onboarding look like? Number 1, a welcome email is sent, that's sent automatically. Number 2, a Slack channel is created for the client and the client's invited. Automatically. And we request access to Ads Manager and things like that if we don't already have access. So all of that's automated. On top of that, the process of creating the campaign structures, the avatars, the ads, all of that is also automated as well. So very quickly, once the contract is signed and the payment's created, what used to take us maybe a day or two of chasing, of doing some research and things like that, is literally just done in minutes by AI. So that means we could literally start running the ads the same day if we really wanted to. But we still need to double-check and make sure everything's set up. So the time to actually launch is so much faster. And then the time to analyze, to process and things like that is also sped up through AI as well. I built my own AI app, which analyzes ad results, which actually goes and identifies opportunities for scaling. Opportunities for trimming wasted ad spend, opportunities to work angles even more that are working. So all the, all of the ads management strategic insights that we'd either have someone doing or myself would be going in to do are done by AI. On top of that, the reporting and things like that is AI and automation and things. So really, and it's not a case of Could I just use AI to replicate what you're doing? My AI is built on the way I like to run ads. So I've pre-built the processes and templates into AI. So what the client gets, they're still getting the expertise that we apply to our ad accounts. In the background, we're just processing it faster and leaner. So I'll give you an example. When you go to buy a car. If you bought a car 40, 50 years ago, the process behind that car production was very different to how it is now. There's, there is a lot more machinery. There are car plants where there really aren't humans doing anything. Now, if you compare, for example, I was watching a video on Porsche in Germany, how they create their cars is very handmade. There's like hand stitching and so much more that goes in. That's reflected in the price. You pay more because it's got that human touch. But does it matter? Does it matter if a machine stitched it? Does it matter if a machine put together the car? For some people it does, for some people it doesn't. So you could get a similar level of performance of a Porsche, but maybe not pay the Porsche prices. For example, you could get a Corvette. They're different classes of cars, I get that. But the Corvette as a sports car is tremendous. It's maybe not entirely comparable to Porsche. Porsche is a different driving experience, but like for one costs so much less than the other because they use a lot more automations and soon I'm sure AI as well. And that's something for you to ponder on for your agency as well, because this process of managing clients leaner, faster, more efficiently, and still get great results for them means that your profitability goes up. This isn't just a case of tech is cool, AI is cool, automation is cool, and it'd be cool to use this kind of stuff. No, this is profitability. This is hard line impact on your bottom line. So let's say, for example, you've been charging $3K a month, and that's been your fee for the last 3, 4 years. Your margin's been tight. There's always hiring and firing and all that kind of stuff that goes on. If all you do is install AI and automation into your client stack, your margin just instantly goes up, because you don't need as much people and as much oversight in terms of what's happening because AI and automation is doing it all. But your results are getting better and better. Now what do you do? Do you still stick to the 3K a month? Or maybe do you increase it because your performance is really good, you're faster at launching ads, you're faster analyzing, your results are even better and things like that. And your reputation is increasing, your clients are staying longer, more clients want to work with you. Naturally, you're gonna increase your prices. That's exactly what we did because we got to a point where we didn't have capacity for these new clients. And what do you do in that case? Do you actually increase the capacity or do you increase the prices? That's what we did. We went from 2 to 3K a month to 5K to 7.5K up to 10K and more a month in ads management because our reputation had started to precede us. People really wanted to work with us cuz they'd seen the case studies and people had talked to people and things like that. Because our bandwidth was limited and I didn't really want to necessarily bring in more people to manage those clients and plug into the AI and tech. I just wanted a smaller client base. And that's the difference between, let's say, 20 and 30 clients paying you $2K or $3K a month versus 10 clients paying you $10K a month. I know which one I would prefer to take, and the one we did take is fewer clients paying more. So they're just things that it opens up the possibilities when you are able to deliver great results, but still maintain a higher margin. It's like the— it's, I'm not even gonna say it's a game changer. I think it's, it's the must-have. You absolutely need to do this. And I learned this the hard way in the first few years of my agency. If you're not tracking the margin per client, the overhead costs and the profit per employee and things like that, then you're going to be losing a ton of money. And when you start to track it, then you realize actually, we're not maybe making as much money as we thought. So the third thing that I want you to focus on is then the CRM insight. So as you start to generate those leads, as those leads come in through either the marketing you're doing, referrals, or whatever it is, and as those clients come in, as clients leave and new ones coming, you're enriching your CRM data, customer relationship management data. And within that, you're going to have lots of insight into lifetime value, which clients are likely to stay, which ones stay longer, which ones pay you more, looking at retention curves. So how do typical clients stay? Like, how long do they stay? How much do they pay? If they pay more, do they stay more? Is it— what's the right price point? What's the right front foot in the door offer? How do I actually build my— do I do an audit first and do I charge for it? Is it free? There are so many things to answer there. What the upsell paths like. One of the best upsell paths that we ever had was in 20— 2019, I think between 2019 and '21, we invested heavily on a $97 ebook. At the time, we were just selling the ebook just to make some side money. So $97 book, I think the average order value was about $120 with upsells. And our kind of cost per acquisition was maybe $60, $70, $80. So we're making money per sale. And that was it. There were some email follow-ups. We didn't really worry too much about it. Selling courses, getting them to book a call. It was only a year or two later as we were analyzing and getting deeper into understanding upsell paths and looking at how people came into the agency that we realized that on the one hand, that $97 product had sold for, I don't know what it was, a million dollars worth of revenue. But that had already, that had also driven a million dollars worth of revenue to the agency. And we were not tracking it. So we didn't enhance it. We didn't push more into it. But what was happening was this, the $97 product was solving a problem for the prospect there and then. And these are people running their own ads or running ads for clients. Then 2 or 3 months later, after getting good results, they're like, how else could this person or this business help me? And some of them were clients that were running their own ads, but got to a stage where they now needed someone else to run their ads. Now, where else are they gonna go? If they've spent 2, 3 months working on your process and your frameworks, and they've got great results, and actually that $97 has turned into $5K, $10K, in, in one case they actually made someone a million dollars, where are they gonna invest that into? They're gonna come back to you. And that's exactly what happened. As the prospect comes in, they spend $97, make money, and they've had great results, they're gonna come back to your agency. There's no pitch. They literally just come in and say, we want you to run your— run our ads. There's no, we're looking at 3 or 4 options and things like that. And that's what we start to understand about the upsell path. Now, fast forward to now where we are with our AI and CRM insights, and we've built our own CRM. We've taken data, all our sales data, all of our prospects' data, that I can now see the upsell path clearly. The AI can tell me exactly what's working and how to shift people. I can also see front-end offers that maybe didn't work as well. So we had another offer, $97, we launched in 2021. For probably last— for the few years, I understood that was driving a lot of calls, and it did. And I understood it was driving a lot of business. In reality, the lifetime value of that offer from 2021 was nowhere near as high as the one from 2019. I didn't see that until recently when the AI told me that. So Sometimes you don't know what the right questions are to ask. Sometimes you don't know how to answer them in the right way as well. And we've had analysts and different people looking at it. But when you've got the AI analyzing your business and your processes and your upsell paths and your low tickets and your high tickets and your closing rates and how long people stay based on the front end offer, it just opens up so much more opportunity, which means I can now replicate what's working well. I can avoid what didn't work. And I can continue to learn from what we're doing as well. So that's the third side, which is CRM insight. And the fourth part I highly recommend you getting AI to help you out with is on your finances. So we're talking about your profit and loss analysis, talking about cash flow forecasting, margin growth, things like that. We have used a finance app called Xero. Yeah, it's called Xero for the longest time. I can't remember when we first installed it probably 2018 or something like that. And it's good. It creates static reports and gives you some insights and things. But it wasn't until probably 2023, maybe '24, I can't recall, that we started to push that data into AI. And it started to recognize patterns that we didn't see. It started to see wasted spend that we didn't see at the time. And it started to advise us on how to avoid the pitfalls of growing a business and not keeping an eye on some of the smaller things that maybe get overlooked, but also better financial planning as well and cash flow forecasting and being able to see where the business is going and when to take key actions. So as we went through some big growth in 2020 and '21, it was always difficult to plan forward on who the next hires need to be. What the salary levels need to be and when to start looking for these people as well, because you're on the one hand, you're growing and you've got revenue and you've got cash in bank. On the other hand, you need to invest on growth because if you want more clients, you need a bigger team and things like that. We had automation in '21, we didn't have AI. So all the decisions were based on if we need to grow to get more clients, we need to get more people. That with AI becomes a far easier conversation because there are times in your business where it's actually good to plateau, which means, for example, if you've hit $20K, and these are the— so first of all, a plateau is when you stabilize revenue. Growth isn't what it maybe is made out to be. If you're constantly in growth, then you're constantly in anxiety and stress. So don't be sitting there thinking growth is the only way forward. Sometimes plateauing and just holding revenue is actually far better for you. I learned that the long, the hard way as well. To give you an idea of where we plateaued and why it seemed like a struggle and why it shouldn't have. First of all, we plateaued around $20,000 to $30,000 a month. In the first probably 12 months, we were up and down in revenue, but we couldn't break through and it really stressed me out. Like I didn't know what I was doing wrong, why we couldn't break through and all that kind of stuff. In reality, what I should have done And what I ended up doing, probably took me a bit too long to do, was to plateau at that level and focus on just building up my cash reserves. Because in my mind, I was just hellbent on growth because I was looking at other people and how fast they were growing and they're sharing case studies and friends in my group were saying, hey, our agency is doing so well. We're at 50, 100, 150K a month. And I'm looking at that thinking, why am I not there yet? The reality is, everyone is on their own journey, everyone's on their own path. And therefore, you need to take reflection points based on where you are to figure out the next path forward. So when we hit 20 to 30k, what I should have done early on is to accept that. And rather than looking for more revenue, optimized on margin. Now at that point, I think our profit levels were about I'm going to hesitate to say 15%, maybe 20%, which means if we're making $30K a month in revenue, then between, give or take, I think, what, $1,500? I can't remember what the math— no, maybe it's $5K, give or take. Whatever the math works out, $5K was profit. Now at that time, if I had brought in, if my revenue was $30K a month for 6 months, and I just stabilize it and I'm bringing in 5K profit a month, which means the team's costs are paid, software's paid, taxes are paid, I'm paid, things like that. And we got 5K left over every month. After 6 months, we've got 30K of profit sitting in the bank account. 30K then gives you choices. Like, do I now go and bring in another hire? Whether that hire costs me 3K a month or 8K a month, whatever it is, I have a buffer. Having a cash buffer reduces anxiety, it reduces the stress of losing a client or missing out on opportunities and things like that. At that time, I didn't really have cash buffers. And my thing was, we're making 5k profit a month now, can afford to bring in another employee for 3k or 4k, and we'll still have 1k left. But the problem with that, and the problem it created for us was, okay, we're now making 5k profit a month, I bring in someone, maybe they pay 3 or 4K a month. We're now down to 1K profit a month. I now need more clients, even though I create more capacity, I need more clients up the revenue and increase the profit level. And so you're always in this position of, I need more money, I need more profit, and I need more people. And we got caught up in that. And I actually ended up making losses through this process. Fast forward to now and where AI is helping me out with cash flow forecasting, AI has access to the sales data and what's happening. And I also share kind of cash in bank and things like that. The AI can better mentor and support me on this. So even though I'm a business owner and I've had finance analysts in the team, we've hired a finance manager, a chief financial officer and things like that. Sometimes what happens is, and by the way, I've programmed my AI to be blunt and to be straightforward. I don't want any fluff. I don't want any I don't want to hurt your feelings kind of thing. Sometimes when humans are talking to humans, especially when you're the boss and your team are talking to you, they don't want to upset you. They don't want you to feel bad. They don't want you to feel like you made a really bad decision when in reality you probably did. They're the conversations that AI doesn't have a problem with. So it will tell you, no, you're making a mistake. Here's what you should do instead. And that makes a huge difference. It's just instantly available. So with, for example, the CRM app that I've developed. My AI has access to everything from marketing and sales to client management to P&L. So within the same app, it has everything, which I've built myself. I can then have a conversation with the AI and say, I'm thinking about hiring this new person or bringing this new product opportunity in or running this thing. How does it fit with everything we've got? And I can get direct feedback as well, based on what our profit and loss is right now. So what our margins are, And right now our margins, it could be anywhere between probably 30 and 50%, which is really good for agency business. And by the way, our revenue isn't just agency. We have memberships and other things as well. So our margin is slightly higher from that, but I can manage our profit and loss much better. I can look at our forecast and say, what's coming up? Like, for example, we might have big costs coming up in AI development or We're running an event or something like that. I need to be able to be aware of those things as well. So as you start to see your business as a machine and you start to see AI and automation as cogs, and you start to see yourself as someone who's outside of this machine, then I think it starts to make sense why you need to be tackling these step by step and bringing AI and automation in. And that's exactly what I've been doing over the last 2, 3 years, breaking it down step by step and saying, where do I need this tech the most? How can it enhance what I've got right now? What could it replace that I'm getting involved with? And what could it enhance with what I'm doing? They're the questions I've been posing to myself and working with ChatGPT to develop all the AI and automation that we have in the business right now. And really right now, we're a team that can operate with probably a bare minimum of 3 people. Probably if I had to trim it down to its core and say, could I maintain this business for the next 12, 24 months? Could we even grow it? 3, 4 people, easy. That question in 2021, I'm talking about 25, 30, 35 people agency. So that's the possibility for you. That's the opportunity for you. I like to run a small business where I have Ideally, no more than 7 to 8 people, people I can talk directly to, people that are cool to work with and things like that. I've run bigger teams. I don't enjoy it personally. Maybe you do. I like my independence of working from home, having my own office, flexible hours, time away from the business, and things like that. So you need to look at what's important for you, and then to map out how to actually develop that. Is it more people? Is it more processes? Is it more AI? Is it more automation? My path nowadays is to ask, can I build an AI automation for it? So as you're starting to think about your next level, are you at the plateau or growth stage? And what are you most stressed about in your day-to-day? And what could you replace with AI and automation? And how can you actually build an agency that works for you and not the other way around?