The five fears that keep owners on the sidelines, what clients actually want, and four steps to start leading this month.
By Jeremy Swiller, ANA-Credentialed AI Training Instructor. Founder, Flux+Form. Published October 28, 2026. Download the PDF
Independent agency owners hold back on AI for five reasons: they fear losing relevance, losing fees, weakening the work, taking on risk, and trusting the wrong people. Each fear is reasonable, and each has an answer. Clients want their agency to lead on AI with clear guardrails, and the people on your team are already using it, with your guidance or without it.
At Lenovo, I started using AI on real work before anyone had given permission. In 2023, ChatGPT helped me understand AI capabilities (unrecognizable when compared with today). It was good for competitive research and bouncing ideas.
Nobody had written a policy yet. In fact, I had numerous conversations with our information security teams about how marketing could use AI. We took calls with OpenAI, and we wouldn’t even allow them to use their AI meeting recorder because we couldn’t guarantee that it would be “safe.”
But I had a deadline and a tool that helped me hit it, and there was no rule to follow.
I tell that story because some of the people on your team are probably doing the same thing right now. They are good at their jobs, they are under pressure, and they found something that saves them an hour. The question for an agency owner is whether that happens with your guidance or without it.
This paper is about why so many independent agency owners hold back, why those reasons make sense, and where to start.
Most owners I talk with are anxious about AI, unsure how to lead on it, and already paying for it. Very few are fighting it. The research backs that up.
The ground is moving. In Agency Core 2026, research on small and mid-size agencies from White Label IQ and Audience Audit, 51% of agency leaders strongly agree the market is changing dramatically, up from 35% a year earlier.
AI costs more than it earns. Forrester and the 4A’s found in 2026 that 61% of agencies still treat AI as a cost of doing business. Only 6% sell it to clients as a line of business.
Each of these is a reasonable response to a real problem. So let’s take the fears one at a time.
This is the big one, and it’s fair. If a client can get a passable first draft from a chat window, what is the agency for?
Clients answered that question in AMI’s 2026 Agency Edge research, a survey of 400 client decision-makers. 86% agree the best agencies combine technology with human creativity. 37% want their agency to teach them how to use AI well.
That’s an invitation to lead. Clients want the agency in the middle of this, and many of them are waiting to be shown how.
Relevance comes from judgment: knowing what a brand should say, what the work should feel like, and what will move people. AI raises the value of that judgment, because someone has to decide which of the hundred options is right.
Owners worry that every hour AI saves becomes an hour the client stops paying for. Forrester reported in 2026 that nine in ten US agencies use AI mainly to cut costs, which invites exactly that conversation.
The client data points somewhere more hopeful. In the same Agency Edge research, 52% of clients said they want better strategy from AI. Only 29% said they want lower fees.
Clients pay agencies for thinking. A team that uses AI to explore more routes, pressure-test strategy faster and spend more time on the ideas that matter is delivering more of what clients value most. When the conversation turns to “what did AI let us do better,” the fee conversation changes with it.
The agencies that lose on price are usually the ones that let the client define what AI is for.
Creative founders feel this one in their gut. They have spent careers protecting craft, and much of what AI produces looks generic.
That’s a fair reading of untrained use. Slop comes from people who take the first output and ship it. Good AI work comes from people who know how to brief it, push it, combine it with their own thinking and throw most of it away.
Craft is a skill your team already has. Training teaches them to apply that same standard to a new tool. The creative director’s eye matters more in this world, because there is more raw material to judge.
Business-side partners worry about risk, and they’re right to. The useful question is which risk, because each one has its own answer.
Client data training someone else’s model. The most common worry, and the most solvable. Business and enterprise plans from the major AI providers keep your data out of model training in their terms. Choose tools for their agreements, set them up once, and the worry has a paper trail.
Work that falls short. Nobody has a stronger reason to ship polished work than an agency. Synthetic production has reached the point where most viewers can’t tell it from a practical shoot, so the real question is who signs off before it leaves the building. A creative director already does that job.
Rights and likeness. Who owns AI-assisted work, and does it borrow from someone else’s? A rights check before anything runs handles it, the same way you clear music or talent today.
Clients who want none of it. Some will say so. Ask, write the answer down and honor it.
The bigger risk is the one nobody is managing. In IPREX’s 2026 survey of staff at its communications agencies, 64% said they need more AI training, and 61% said they have little or no visibility into what their clients expect on AI. People are using these tools anyway, each with their own habits and their own idea of what’s safe.
A ban pushes all of that out of sight. Guardrails put it in the open, where you can manage it.
The AI training market is crowded with tool vendors, tech consultants and people who discovered AI last year. Most of them have never sat in a client meeting, defended a campaign or run an agency budget.
Owners are right to be skeptical. Whoever trains your team should understand how agencies make money, what clients actually ask for and what good work looks like.
Set the fears beside what clients say, and a clear picture shows up. From the 2026 Agency Edge research:
89% expect their agency to stay fully accountable for any AI used on their work.
82% say transparency about AI use increases their trust.
65% say it’s essential that agencies explain how they protect client information when using AI.
Clients want a partner who has thought this through, can explain it plainly and stands behind the result. That’s a leadership role, and independent agencies are well placed to play it. You are closer to your clients than any holding company, and you can move faster.
Guardrails let your team use AI with confidence and let you tell clients exactly how it’s handled. A good starting set fits on one page and covers six things:
1. What goes in. Which client information can be used with AI tools, which can’t, and which needs the client’s sign-off first.
2. Which tools. A short approved list, with the settings that keep client data out of model training.
3. Who reviews. Every piece of AI-assisted work gets a human owner who checks it before it leaves the building. Accountability stays with the agency.
4. What you disclose. A plain-language statement for clients on how the agency uses AI, plus a way to honor clients who want none of it on their business.
5. Rights and originality. How you check AI-assisted work for likeness, trademark and copyright issues before it runs.
6. How it’s taught. Everyone learns the same rules and the same methods, so quality doesn’t depend on who happened to figure it out.
Once these are written down, the conversation with clients gets easier. You can tell them what you do, what you don’t do and who is responsible. That’s the transparency and accountability they’re asking for.
Leading on AI usually starts small:
Ask your team what they’re already doing. Anonymously, if that helps people be honest. You’ll learn more in a week than from any report.
Write down the guardrails. One page is enough to start.
Train together. A shared method beats a dozen private ones. Teams that learn together build the same standards and push each other’s work.
Bring it to clients. Show them your guardrails and one or two examples of better thinking. Let them see you leading.
Each of these steps takes weeks, and none requires a big purchase. What it takes is a decision to lead.
Related reading: How Should an Independent Ad Agency Adopt AI? and, in our dictionary, guardrails and AI safety.
Thirty minutes, no pitch, just a real conversation about whether the workshop is the right fit.