AGENCY OWNER INSIGHTS / KENT LEWISGO YOUR OWN WAY / KJL

AI MADE THE OUTPUT FASTER. WHO OWNS THE RESULT?

A practical way to distinguish automated production, professional judgment and accountability when pricing agency work.

AI has made an old agency pricing argument harder to avoid. If the software can produce a useful first draft in minutes, why is the fee the same? It's a fair question. I wrote about this pressure in NextNW, and I think the wrong response is to defend the hours as if nothing changed.

The better question is what the client is buying after production gets faster. A polished deliverable isn't a business decision, and a decision without someone accountable for the consequences isn't much of a service. You can automate parts of the work; you cannot automate away responsibility.

KJL / FOUNDER FIELD NOTESPRODUCTIONSpeed and costJUDGMENTChoices and riskACCOUNTABILITYWho stands behind itThe invoice should reflect what the client actually buys.
AI-era fee sanity check — an original conceptual illustration, not a measured benchmark.
Diagram explained: A sustainable agency fee covers production, professional judgment and accountability. Automation may reduce production cost without removing the other two.

AI-era fee sanity check

Try this with your own figures or decisions. It runs in your browser; nothing is submitted.

For reflection and planning only. Illustrative scores and estimates are not forecasts, legal, accounting or investment advice.

Now add AI. A job that required substantial production time last year may take a fraction of that time today. The client knows it. The agency knows it. Both sides can feel the old conversation becoming awkward.

I wrote for NextNW in June 2026 about what I called agency pricing's "Kodak moment." The core challenge is familiar: if agencies sell hours while technology reduces those hours, they'll eventually have a serious commercial problem.

But I want to explore the question behind the price conversation. If a client can generate an output quickly, why should they pay an agency at all?

My answer: because a deliverable is not the same thing as a decision, and a decision is not the same thing as an accountable outcome.

Output is cheapening. Judgment isn't automatically improving.

AI can help create drafts, summarize research, generate options and accelerate analysis. That's useful. I am not interested in romanticizing slow work because it used to command a larger fee.

The danger is that more output can look like more value even when nobody has checked whether the output is accurate, relevant or right for the client's business.

An agency can now produce twenty campaign concepts where it once produced five. Great. Which one should the client actually use? How does the recommendation fit the audience and the economics? What evidence supports it? What happens if an assumption proves wrong? Who owns the consequences?

Those are the questions buyers are still willing to pay experienced people to answer.

Separate three layers of the engagement

When considering how to package an AI-assisted service, I'd separate the work into three layers. This is a proposed way to frame a discussion with owners, not a claim that Anvil used an identical model.

Production. The assets, research, reporting or implementation. AI can change the time and cost of this layer materially. Be honest about that instead of disguising automation as heroic manual labor.

Professional judgment. Defining the problem, choosing among options, testing assumptions and recognizing where the technology is unreliable. This layer may become more valuable as the volume of possible outputs increases.

Accountability. Setting expectations, protecting sensitive information, checking accuracy, handling the consequences of an error and revisiting the decision when results arrive. Someone has to own this layer regardless of how fast the first draft appeared.

If the agency's offer consists entirely of interchangeable production, price pressure is logical. If the agency's actual value lies in the second and third layers, the pricing conversation needs to make that visible.

Don't pretend every client needs value-based pricing

There's a lot of enthusiasm for pricing exclusively on results. I understand the appeal. When work becomes more efficient, why should the agency earn less for producing an equally good outcome?

But outcomes are often influenced by variables beyond the agency's control. A marketing firm can improve a funnel and still be affected by the client's product, sales process, pricing, inventory or economic conditions. Promising to own results you don't control is not a sophisticated pricing strategy. It's a risk-management problem.

Different engagement types deserve different structures. A fixed-fee project can be sensible when scope is clear. A retainer can support ongoing strategic work. Time-based consulting can still make sense when a buyer is purchasing access to expertise or an uncertain discovery process. Performance incentives may work when attribution and responsibilities are genuinely understood.

The answer is a pricing model you can defend—commercially and ethically—not whichever label is fashionable this year.

A pricing conversation worth having with your leadership team

Choose one service that AI has meaningfully changed. Ask:

  • What portion of the client's value comes from production, and what comes from judgment or risk reduction?
  • Has automation actually lowered fully loaded delivery cost, once review, quality control and governance are included?
  • What would the buyer do without us? A DIY tool, an employee, a cheaper agency—or nothing?
  • Which outcomes can we influence, and which belong to the client's team or outside conditions?
  • Can we explain our fee in one clear paragraph without talking only about hours?

If you can't answer the fifth question, the market will eventually answer it for you. Probably at a lower price than you'd like.

The positioning issue disguised as a technology issue

I have lived through enough waves of digital marketing technology to be wary of both extremes: "this changes everything" and "this won't change our business." Usually both are too simple.

AI does change what agencies can produce and how quickly. It also gives clients new expectations. The opportunity for independent firms is not to insist that every old service deserves its old pricing. It's to demonstrate where experience, creativity, context and accountability produce a better business decision.

Clients aren't obliged to finance an agency's inefficiency. Agencies aren't obliged to price their judgment as though it were an interchangeable software subscription. There is room between those positions for a healthier relationship.

If AI is changing your staffing, service mix or pricing faster than your business model can keep up, that's a good decision to bring to agency-founder advisory.

What I'd do next

I wouldn't use a universal rule to replace every hourly engagement with value pricing. Sometimes the scope is clear, sometimes uncertainty is high and sometimes the client controls most of the outcome. Choose a model that allocates risk honestly, then make sure the fee covers the real cost of review and judgment.

Try the cost model below on one service before changing your rate card. Include senior review time and a realistic allowance for quality control. If the numbers only work when you pretend those costs are zero, you don't have an AI efficiency gain yet; you have a spreadsheet problem.