The Solo Flight: Why Brands Are Ditching Their Agencies for AI, and What to Pack Before Takeoff

There’s a meeting happening in a lot of conference rooms right now. Budgets are under review. The agency retainer is on the table. And someone asks, “Why don’t we just see what AI can do?”
I’ve started calling this moment The Solo Flight. Brands are pulling back on agency support and taking the controls themselves, with an LLM in the copilot seat. Some of them will land just fine and some won’t. This post is for both.
This Isn’t a Rumor
Forrester’s 2026 B2B Brand and Communications Survey shows the shift clearly. Ninety-three percent of B2B marketers still use agencies in some form. But the share expecting agency budgets to grow dropped 13 points from last year, and nearly half now expect spending to stay flat. In digital marketing, the share expecting more agency spend fell from 51% to 31% in a single year.
Forrester analyst Karen Tran summed it up well: “B2B marketers are not walking away from agencies, but they are becoming far more disciplined about where agencies add value.”
So the planes aren’t grounded. Brands are just choosing more carefully who rides along.
The Case for Taking the Controls
Questioning every dollar you spend on an agency is exactly what a good operator should do.
AI is good at a lot of things. It can write fifty headline variations before your coffee cools. It can summarize reports, expand keyword lists, develop 6 months worth of organic content to post, and manage bids inside a structure you’ve already set. If your team can take some of that on and free up budget, you should.
Agencies have some looking in the mirror to do, too. The same Forrester research found that more than half of B2B leaders now weigh data strategy and AI readiness when choosing an agency, and only a small share are satisfied with what they’re getting. That’s on us.
Your Copilot Agrees With Everything You Say
Here is where my concern lies. Researchers at Stanford and Carnegie Mellon tested leading AI models and found they affirm users’ actions about 50% more often than humans do. The bigger surprise was how people reacted. They rated the agreeable answers as higher quality, trusted the model more, and said they were more likely to use it again.
That study looked at personal advice, not media plans. Now picture asking an LLM, “Should we cut Connected TV and move everything into search?” You’ve already told it what you want to hear. A tool built to be agreeable isn’t likely to tell you the premise is wrong.
A quality agency partner will. Sometimes that pushback is the most valuable thing you’re paying for.
Same Budget, a Third of the Return
In August, I wrote about a client who learned this the hard way. (Read the full story here.)
Here’s the short version. The client had an LLM review its paid advertising. It recommended moving spend into bottom-funnel paid search and cutting Connected TV and display retargeting. We advised against it. The client went ahead.
The search metrics looked great. The business results didn’t.
Full-funnel program: about $105,000 in spend, returning 26 to 1.
Search-only program: about $106,000 in spend, returning 8 to 1.
Cost per click went from around $3 to nearly $11.
Direct traffic fell by roughly half.
Same money. A third of the return.
I’ll repeat the caveat I gave then. This is correlation, not proof, and the deal mix shifted during the same period. But the pattern was hard to ignore. The upper funnel was warming up the audience, and that warmth kept search clicks cheap. When the upper funnel was cut, the cheap clicks went with it.
As I wrote then, “Automation is excellent at optimizing inside a lane. It is not yet good at deciding which lanes should exist.”
What a Crew Brings That a Copilot Can’t
AI can fly the plane on a clear day. A seasoned crew earns its keep when the weather turns. Here’s what that crew brings:
Context. We know your sales cycle, your seasonality, and what happened the last time someone tried this.
Accountability. A real person owns the outcome, not just the output.
Cross-channel judgment. We decide which lanes belong in the plan, not just how to tune each one.
Pattern recognition. Over six years, we’ve managed millions in ad spend across a lot of brands, and we’ve seen how these stories usually end.
Real efficiency. Count the hours your team spends learning, prompting, checking, and fixing. The agency is often the cost-saving move.
Your Preflight Checklist
If you’re going solo anyway, and plenty of you will, here’s how to give yourself the best odds.
Keep the funnel architecture with a person. Let AI tune the engine, and let someone accountable choose the route.
Test one lane, not the whole program. Run AI on a single channel or campaign and measure it against a control.
Watch the leading indicators. Cost per click, direct traffic, and branded search will warn you before a quarterly ROAS number does.
Give your AI the full picture. Include your attribution gaps, campaign intent, and sales cycle. Without that context, you’ll get confident answers to the wrong question.
Set your check-in date before takeoff. A quarter is enough to see costs move. It isn’t enough to see demand generation pay off.
We’ll Keep the Runway Lit
Fly the Solo Flight. Learn what AI can and can’t do for your business. You’ll come away with a sharper sense of where your marketing dollars work hardest, and that’s a win either way.
When you’re ready to bring a crew back on board, whether that’s next month or next year, agencies like ours will be here with the expertise, the efficiencies, and results that are already proven.
And if you’ve built a plan with AI and want a second set of eyes before you commit, just ask an agency for a consult for an experienced POV. No judgment, just an honest read on what you’re about to fly.
Safe travels. We’ll keep the runway lit.





Comments