How Agencies Increase Profit Margins with AI (2026 Guide)

Really excited to chat today with Garrett. Garrett is our go-to-market advisor at Rize. He's been immensely helpful, but he also runs his own consulting practice and is really connected in the agency space and knows a lot about what's changing and what's happening. I'll let you introduce yourself, Garrett, and then we can get into some questions.

Sounds good. Thanks Macgill, appreciate you having me on. My background is mostly working with early stage B2B software startups. I was a CMO and founding marketer for over a decade and I've been consulting full-time for about three and a half years. What's really interesting is because my background is mostly in software, I thought I'd be mostly working with software companies, and that's true — I'm working with Rize and others. But about half my clients today have ended up being agencies. I do a lot of work with early stage agencies around the one to two million in revenue range. The problems I help with are really similar across both software and agencies: building marketing and sales foundations — tight audience, tight offer, strong messaging, and clarity on which channels to focus on.

That resonates with a lot of the ways you've helped us this past year. For this topic we wanted to focus on AI and how it's changing things in the agency space. Have you seen any of those core marketing and sales foundations change?

A couple of things that have been top of mind. One is the old agency model is starting to die. The old agency model was hire more bodies to get more capacity, sell that capacity, make sure you have good margins. That's how the classic SEO, paid ad, and execution agencies were operating. AI has made agencies more efficient with their time — they can get more done — but it's also lowered the barrier for execution-only agencies because clients can now get AI to do 80% of the execution. It's a race to the bottom if you're an execution agency that focuses on growing by hiring people and increasing capacity. Successful agencies going forward have to be a strategic partner, not just an execution arm. You can't just scale through hiring more bodies. You have to think about the tech including AI that you use. You have to think about what's unique about your process.

How are some of the ways you've seen agencies make that shift from being more execution to being more strategic?

We're seeing a huge boom, myself included, with fractionals, consultants, and people going out on their own. I call it the micro agencies. A small operations team, a strategic founder, and a lot of AI. Honestly, those micro agencies are competing heavily with the heavier people-backed agencies today, and they have better margins, so they're more sustainable. The fractional play has definitely influenced the agency world since we're seeing so many more fractionals and consultants going solo.

Do you think the age of the bigger agencies is dying away and the future is more fractional and freelance consulting work across the board?

I think so. I also think the day of the full-service agency is rapidly dying — where you just hire one agency that does everything. Larger companies that are slower to adopt new trends will continue hiring full-service agencies for a while, but especially for smaller and mid-market companies, it makes no sense to hire a full-service agency when you can hire a bunch of specialists who are experts at what they do. If you want to stand out in the agency world today, you have to find your niche. Usually the best way to niche is by the problem you solve.

What's one area that you think is here to stay — a core value that you don't think AI is going to make changes to in the agency world?

People are nervous about AI replacing knowledge workers. I don't see that happening anytime soon. Some of that will happen at the lower end of execution — that's what it's good at. But any good AI output is about a couple of things: the quality of the inputs you put into it, the right prompt or guidance, and then the expertise to see the output and know if it's actually good quality. There's always going to be room for deep expertise and the right inputs used in conjunction with AI. We've all seen people who aren't copywriters think they produced something great with AI because they don't have the expertise to judge it. If you focus on being that strategic partner as an agency founder, helping with those two ends of the spectrum with AI, you're going to be okay.

From what you're seeing in the agency world, is there an area where agencies are overestimating AI — where they think it can do more than it actually can?

A friend of mine posted on LinkedIn — he said we have replaced so much of our execution with AI, and the punchline was it's only taken 150% of the time that it used to. There's truth to that. We're still figuring out the right way to use AI efficiently. You have to find the right balance. If you try to go all in on AI today and replace all your human-proven processes with AI, it breaks. You have to find strategic places to start introducing AI without blowing everything up. When you try to go all in, it ends up taking 150% of the time it used to, just to manage it. Have a healthy skepticism and approach it strategically.

What are some ways you've seen agencies approach integrating AI well?

You can't really integrate AI unless you have an established process that's been proven. If you're doing completely custom work every time and reinventing on the fly, with no standardization, it's going to be really difficult to implement AI successfully. It has to start with some sort of standardization. I have a proven process and now I use AI to help speed up that process. For me, I've started building out a lot in Claude code, including my own agents built on my processes and context. When I'm doing website messaging and positioning, I give the right input — usually customer interview data or sales call data — and it knows my framework and spits out something that's 80% of the way there. I refine it and it speeds up my process. It really comes down to: do you have a process? Do you have any sort of standardization?

We found internally at Rize that was something we were thinking a lot about. We had to reflect every couple weeks — how can we implement AI more, where do we feel time is being wasted? You default to your existing processes, so you have to be intentional.

It's an interesting balance. You have to spend some time learning what's new, but you don't want to go too far and blow everything up so it breaks next week.

What's one thing agencies are still doing manually that will feel completely insane in 12 months?

I'll come at this sideways. I recently went to a Claude code class on how to set up your own chief of staff for your own business — a core agent that knows your business context, has access to your calendars, emails, knows who your clients are. The instructor said it's going to be crazy that if everyone in five years doesn't have their own AI chief of staff with all the context on their business, their unique models, and access to their connections. Where it gets really powerful is when everyone codifies their expertise into their own chief of staff agent, and those agents can interface and talk with other experts' agents. If my agent can tap into the unique expertise that your agent has, imagine the kind of future that could create. It could change work drastically over the next 3 to 5 years.

How do you see the structure of agencies changing with AI — with things like everyone having a chief of staff agent?

We talked about micro agencies. In the past, one of my first jobs in college was a small SEO agency where I was one of many students hired for cheap labor to crank out SEO title tags and meta descriptions. That model is just outdated today. It's way better to have a small micro team of an expert founder, a small operations team, and AI agents that help fill in the gaps on execution. I have a friend with a LinkedIn paid ads agency and his whole model is: I'm only hiring experts with a certain amount of experience, we all work directly with clients, and we use AI to fill gaps. None of the classic agency where you talk to an expert who sells you and then passes you off to a junior person still in college to execute. That's the model that's dying.

What parts of agency work won't be automated away? Where do you still need the human?

I like to think about the opposite — what parts are actually going to be removed. The challenge for people starting their careers is that the lower entry-level jobs, the grunt work where you pay your dues, that's the stuff being replaced by AI. The strategic side — knowing how to use AI, knowing what good output looks like on the back end — that's not going to be replaced. I have another client who hired a kid right out of school with no SEO background but was a go-getter and great with AI. His VP of marketing said: go learn how to do SEO and AEO using AI, get us to rank better in ChatGPT and Claude. Within a couple of months, he had made drastic improvements using AI to help him implement. That's the type of stuff where you can have a big impact going forward. It's not just doing the junior execution work anymore.

It's a good time to have high agency. How should agencies be thinking about time tracking in an AI world?

It's still vital. It goes back to profitability. You want to have visibility on where you are spending your time and where your employees are spending your time — even in the future, where your agents are spending time across which clients. That visibility is one of the most important things for a founder or business owner. If you can have realistic visibility on how your business is actually operating, where time is being spent, where the bottlenecks are — that's one of the biggest challenges. Not having that data is a huge problem because as an owner you need to spend a good amount of your time removing bottlenecks and optimizing. It's a crucial piece going forward for margins and profitability.

If there's a time tracker you want to recommend, feel free.

Rize is the one I've seen that is on the cusp of doing this the right way with the changing model and world. I've tried and used multiple time trackers. The thing that's awesome about Rize is people first come in and are like, wow, this is amazing, I can just do this automatically — I don't have to have that little start/stop clock in the corner of my screen. Then you start understanding all the layers deeper and it's really powerful and impactful. Highly recommend.

How are you thinking about utilization? Time tracking has been used so much for workload and utilization and AI is changing that.

The old model was hire bodies, throw them at the work, mark up that work. That was really the only leverage lever agency owners traditionally pulled. There's actually a lot more leverage levers they can pull — niching, deep expertise, AI as a form of execution leverage. The idea is: how do we make the most money in the least amount of time without trading time for money directly? You still want visibility into where time is being spent, whether it's you or the agents you've built internally. You need that visibility so you can know: am I making good margins on this? Am I actually profitable? It's similar to the past, but a different way of thinking about it. Now it's about agents and other forms of leverage, but you still need to track that spending to make sure you're profitable.

How do you think AI is going to change margins for agencies?

The ones that figure it out are going to drastically improve margins. The ones who are stuck in the old world of hiring and throwing bodies at the problem are going to see their margins shrink. If you're just an execution agency, that's being quickly replaced by AI. It's a race to the bottom — you're going to have to lower your price over time and your margins will shrink. You want to be the expert helping with some of the execution, but you have to have that expert lens in the work you do.

In the past 10 to 15 years, billable hours have more or less been replaced with retainers. Do you see billing changing with AI?

I definitely think we'll see it change. We've already started seeing more agencies and software become outcome-focused. It's not just paying per seat or per hour — it's paying per outcome. AI is one of the things starting to change that. It's less about the execution because the cost of execution is going down with AI. It's more about: can you help me actually achieve this result? Agencies can potentially change their entire model and cost structure with good execution that's a lot of AI, but still get paid really well by helping hit certain milestones and outcomes. I'm interested to see how this continues to evolve.

I've got a couple of hot takes for you. AI is going to kill blank inside agencies.

AI is going to kill time wasting inside agencies. If you have visibility into where time is being wasted with tools like Rize, you can eliminate that. There's a lot of time wasting and fluff that you can just eliminate when AI starts helping with the execution.

The biggest mistake agencies are making with AI is blank.

It's on two extremes. Either ignoring AI and not doing anything with it, or going all in and trying to redesign their business and fire everyone. Finding the middle ground is the key today.

In two years, agencies that don't do this will struggle.

Agencies that don't use AI at all are going to struggle drastically. But also agencies that try to use the old model of just hiring people and marking up as their main point of leverage — that's going to hurt them. And agencies that don't focus and become specialists are going to struggle. The days of the full-service agency are dying rapidly.

That's all the questions I have. Any last words?

This has been awesome. Appreciate you having me. Go check out Rize if you don't know what it is.