Preparing for AI

You're probably hearing a lot about AI right now.

How it's going to change everything. How you need to adapt or get left behind.

And if you're like most I talk to, there's a low-level anxiety underneath all of it. A sense that something big is coming and you're not ready. That everyone else is figuring this out while you're still trying to keep the lights on and the clients happy.

I want to tell you something that might help: it's not as complicated as it looks.

But there is pre-work.

And the agencies that do that pre-work now will be the ones who capture the gains when AI becomes capable of doing meaningful work. The ones who skip it will be scrambling later - or they'll try to implement AI into a mess and get nowhere.

This is what I mean…

Right now, AI can do simple tasks. It can automate certain basic things. But the gains aren't significant yet. We can't get a ton of real work done through it - not the kind of work that moves the needle on profitability or capacity.

That will change. Probably faster than any of us expect.

Obviously, you can't just plug AI into your agency and expect it to work.

AI needs structure. It needs clarity. It needs to know exactly what the workflow is, what the inputs are, what the outputs should look like, what good looks like vs bad.

If you don't have that clarity for your human team right now, you definitely won't have it for AI.

What’s beneficial is: the pre-work for AI is exactly the same as the pre-work for running a great agency. The things you need to prepare for AI are the things you should already have in place for your people.

Let me break down what that pre-work looks like.

First: RACI matrices for every service you deliver.

RACI stands for Responsible, Accountable, Consulted, and Informed. For every step in every process - quoting, scheduling, creative brief development, copywriting, design, whatever - you need to know who's doing the work, who has veto power, who needs to give input, and who just needs to be kept in the loop.

This seems basic. Most agencies don't have it.

As AI develops, it's going to be able to do certain things before others. Just like you have different roles in your agency that do different things, there will be different AIs that do different things.

The RACI tells you exactly where to slot them in.

For example, maybe right now your copywriter is responsible for all the writing and accountable for the quality. But in six months, maybe an AI writer becomes responsible - it does the bulk of the work - while your copywriter shifts to accountable. They're editing, quality-assuring, making sure it's strategically aligned. The AI does 80% of the work; the human ensures the 20% that matters.

You can do this across your entire operation. Design AI responsible for visual development, art director still accountable for strategic alignment and client fit. Quoting AI responsible for generating estimates, account manager accountable for accuracy and client relationship.

But you can only make these swaps if you have the RACI built. If you know exactly what each step is and who owns it. Without that map, you're just guessing where AI might fit - and guessing doesn't work.

Second: SOPs for every step.

Every item in your RACI could potentially have a standard operating procedure behind it. Documentation for how that step gets done.

An SOP needs a few things: the goal of that step, who it's for, the methodology if needed, step-by-step instructions, a checklist, a flowchart, and ideally a video of someone executing it with verbal explanation of the context and considerations.

This is important for your people. It creates clarity. It makes accountability possible. If someone doesn't know how to do something, you can't hold them accountable for doing it wrong.

But it's even more important for AI.

When you eventually want to automate a step, the first thing any AI engineer is going to ask you is: what is the workflow to get this work done? They need that static workflow because that's what the AI works off of. If you don't have clear workflows documented, they cannot engineer the automation. It's impossible.

The agencies that have their SOPs built will be able to plug AI in almost immediately when the technology is ready. The agencies that don't will spend months doing the documentation work they should have done years ago.

Third: data warehousing.

AI learns from examples. It needs to see what good looks like. What were the inputs? What were the considerations? What was the output? It needs many examples - not just one - so it can learn from different angles.

Which means you need a strategy for storing your finished work products in a way that AI can eventually learn from.

All your quoting information goes in a quoting folder - the considerations, the inputs, the outputs, multiple examples. All your copywriting goes in a copywriting folder. All your design work goes in a design folder. Every service, every step, organized and accessible.

You should also be recording things you're probably not recording right now. Internal brainstorming meetings where you discuss creative direction. Client presentations where you show work and get feedback. The back-and-forth conversations that reveal why certain decisions were made.

AI needs to learn from as many angles as possible. The richer your data, the better it will perform when you eventually train it on your specific way of doing things.

I can’t stress enough how vital the boring operational work is right now.

This work pays off immediately, regardless of AI.

If you build your RACI and your SOPs and your dashboards, you will see tangible gains from your team right now. Increased clarity. Increased accountability. Better efficiency. Fewer things falling through the cracks. People knowing what they're supposed to do and how to do it.

The pre-work for AI is just good operational work.

It's the stuff you should be doing anyway. AI is just another reason to finally do it.

One more thing worth mentioning.

Until the hallucination problem is fixed - until AI stops confidently producing things that are wrong - a human always needs to be accountable for the work AI does.

AI can be responsible in the RACI. It can do the bulk of the work. But someone on your team needs to be accountable for quality assurance. Someone needs to check what the AI produced and make sure it's right.

This is important to understand because it shapes how you think about AI's role.

It's not replacing your team. It's multiplying them. Your copywriter isn't going away - they're becoming an editor and QA function for an AI that can produce ten times the output.

Your designer isn't disappearing - they're becoming the strategic layer that ensures AI-generated work actually fits the brand and the client.

The humans shift from doing the work to ensuring the work is right.

That's a different skill set.

And if you start developing that skill set now, you’ll be ready when AI is ready.

If you’d prefer to learn more about this through video, you can click below to watch me break it down 👇

Nick

P.S.

If you want help thinking through what this looks like for your agency - the RACI work, the SOP development, the operational infrastructure that prepares you for AI while improving your business today - reply with "AI" and let's talk.