An Agency CEO's Unfiltered Take on AI

We asked Kara Redman, CEO of Backroom—a Baltimore-based agency that's been navigating the AI shift without the hype or the panic—to share her unfiltered take on how AI is actually playing out in the agency world.

What came back was brutally honest. No agency doublespeak. No "AI transformation" buzzwords. Just a practicing CEO explaining what's actually working, what's broken, and why most agencies are setting themselves up to collapse.

You can listen to Kara's full audio responses below, or keep reading for the highlights and our exchange on what this means for marketers.

This one's longer than usual. Worth the read if you're working with agencies, running one, or just trying to figure out where the human/AI boundary actually belongs.

» Kara’s running a newsletter, The Saturday Paper. Sign up here.

She's right. We've watched this play out across dozens of companies. They buy the promise of "10x content output" without asking whether their current content even works.

Kara: "If your core inputs are mediocre, AI just scales mediocrity. It doesn't make it better for you."

This is the 2010 "content is king" playbook all over again. Back then, everyone needed a blog. Legal firms, CPAs, B2B tech companies—they all started churning out white-labeled content that was pre-approved but completely generic.

We created a digital landfill. And we're about to do it again, just faster.

Kara: "I see clients spending money on these automated engines, and I'm watching them produce more content that doesn't move the needle. The minority that are using AI for deeper audience understanding—pattern detection, resonance tracking, behavioral clustering—those are the ones getting real advantage."

Organizations that couldn't create valuable content before AI won't suddenly get better because they added automation. They'll just create more noise.

Kara: "Clients expect more with less now. They think they can do creative work themselves because they have a chatbot to brainstorm with. But smart clients see where the line exists—where AI speeds up execution but can't replace strategic thinking."

This is the tension every marketer is feeling right now. If a tool can help you brainstorm, write copy, or generate variations, why pay an agency?

We think about this in terms of the taste gap. We wrote an entire series of content around this problem (start here). AI has democratized production capacity, but it hasn't democratized judgment. The ability to know what's worth making, what resonates, what cuts through—that's becoming the premium skill.

And most agencies haven't built their value proposition around that judgment. They've built it around execution hours.

Quick word from our sponsor today:

Remember when software made business simpler?

Today, the average company runs 100+ apps—each with its own logins, data, and headaches. HR can’t find employee info. IT fights security blind spots. Finance reconciles numbers instead of planning growth.

Our State of Software Sprawl report reveals the true cost of “Software as a Disservice” (SaaD)—and how much time, money, and sanity it’s draining from your teams.

Read the full report and see how your company stacks up→

The future of work is unified. Don’t get SaaD. Get Rippling.

Stop SaaD in its tracks

What survives? Agencies with strategic IP. A strong point of view. Research that understands what gap you fill in the market. If you've built a business around that, AI helps you scale. If you're reacting to client needs or doing the same old, AI just exposes the cracks.

This is why we keep saying AI skills are management skills. The agencies winning right now aren't the ones with the best prompt libraries. They're the ones who know how to redesign workflows, where to inject automation, and where to protect human judgment.

They've done the hard work of defining what they're actually selling beyond deliverables.

This hits hard because we see it everywhere. Companies want AI to solve problems that aren't technology problems.

Kara: "It's a readiness gap, not a technology gap. If you have infrastructure to support it, AI becomes a competitive advantage. If you're using it to close gaps in your internal infrastructure, it won't work."

Here's the brutal truth: If your strategy is unclear, your messaging is inconsistent, and your team doesn't have decision-making frameworks, adding AI just amplifies the chaos.

We've watched marketing teams spend six figures on AI tools when their actual problem was that nobody agreed on who the customer was.

Oh, btw, try our AI Readiness Assessment, purposely created by marketers for marketers, and get your score. Try it here.

Kara: "Right now, most clients are using AI to churn out mediocre content faster. Turning AI into a glorified copy machine is the default. The minority using it for deeper audience understanding—those are the ones getting real advantage."

The question every marketer should ask: are we using AI to get smarter about customers, or just to produce more stuff faster?

This is the pattern recognition advantage. Where else? Of course, creative testing. Kara’s team has seen AI outperform humans in diagnosing why a creative resonated or flopped. Feed it structured engagement data plus qualitative patterns, and it spots signals earlier than teams typically do."

This is where we see the future going too. Right now, we're spending money upfront on focus groups, testing, and paid spend experiments. AI's pattern recognition can help predict performance before we spend the money.

Not perfectly. But better than guessing.

And where does it fail?

She's absolutely right. The moment you automate your voice, you sound like everyone else who's automating their voice. Anything requiring judgment—final creative, client communication, strategic decisions. “If we try to automate nuance, it's going to bite us later,” she adds perfectly.

This is the boundary. Use AI for synthesis, pattern recognition, and baseline work. Keep humans for framing, judgment, and anything that requires reading between the lines.

Big agency shops have complexity, committees, procurement, and rigid processes. We can pilot tools quickly, experiment with clients, and adapt workflows in days instead of quarters. We've seen this play out. Enterprise companies spend nine months getting approval to test a tool. Small agencies just started using it on Monday.

This matters for in-house teams too. If you can move fast, test small, and iterate based on what works, you have an advantage over organizations that need consensus before trying anything.

“One: Taste becomes the differentiator. Execution can be automated, but taste, style, the creative influence and social aesthetic we use—that's human.

Two: Strategic abstraction. The ability to move from messy business reality to clear positioning is a premium skill. Three: Systems thinking. Knowing how to combine human judgment with machine automation will separate top agencies."

This is the core of the taste gap framework. When everyone can produce at scale, curation becomes premium. And what gets automated?

"Production, reporting, research, baseline creative. What stays human: framing, decision-making, storytelling, leadership, brand voice."

This is the division of labor we're all figuring out. Not "what can AI do?" but "where do humans add disproportionate value?"

The agencies—and marketing teams—that answer that question clearly will win.

“We're actually considering writing this into our contracts—here's what information we might input into AI, here's what we use it for."

This is smart. Because right now, you don't know if you're paying agency rates for ChatGPT output.

Kara: “Refuse to tolerate agencies selling AI magic. If they don't have a point of view, a framework, accountability for results, the ability to pick up on nuance from clients—if they're just plugging in notes and saying AI will create magic—that's unacceptable."

Here's our version: If an agency can't explain where their judgment matters and where automation helps, they're not an AI-fluent agency. They're just an agency that uses tools.

The ones worth working with can articulate the boundary. They know what they're protecting and what they're accelerating.

The Bottom Line

The stress test is here. AI isn't going to save broken agency models or unclear strategies. It's going to expose them.

What survives: strategic IP, clear point of view, and the discipline to know where the human/AI boundary belongs.

What collapses: volume plays, busy work disguised as strategy, and anyone who can't articulate what humans are actually contributing.

If you're working with agencies, demand transparency. If you're running one, get clear on what you're selling beyond deliverables.

And if you're a marketer trying to figure out where you fit, remember: taste, strategic abstraction, and systems thinking. Those are the skills that matter when execution gets cheap.

The future isn't about who can produce more. It's about who can judge better.

Want more from Kara?

She publishes The Saturday Paper, a weekly must-read on brand strategy and creative leadership. If you're looking for an agency that's actually thought this through, check out her work at Backroom. Plus, follow her on LinkedIn.

See you on Monday,

— Peter & Torsten

Loved this edition? Share it and unlock rewards!

Our sponsor today:

Software sprawl is draining your team’s time, money, and sanity. Our State of Software Sprawl report exposes the true cost of “Software as a Disservice” and why unified systems are the future.

Stop SaaD in its tracks