Marketing Has Never Been Better — Or Worse. Why Agencies Have to Change.

The best and worst time in marketing, at once

The upside is obvious once you have used the tools seriously. Research, page structure, drafting, internal linking, technical cleanup — work that used to sit in a queue for weeks now moves in days. Output went up and the ceiling on what a small agency can deliver went up with it.

The downside is that the destination changed while everyone was celebrating the speed. Customers increasingly ask an AI assistant instead of scanning a page of links. That answer is assembled from sources the AI can read, understand, and trust. If your site does not communicate clearly to those systems, you are not in the answer, and being absent from the answer is the same as being invisible.

What AI actually changed about being found

Ranking is no longer the finish line. An AI assistant does not hand back ten blue links. It writes a short recommendation, names a few businesses, and moves on. To get named, your site has to answer the question a real person asked, in plain language, in a structure a machine can lift cleanly.

• —Questions have to be answered on the page, not implied by it
• —Every service and every town you serve needs real depth, not a single thin list
• —Facts — hours, service area, pricing approach, credentials — have to be consistent everywhere
• —Structure matters: clear headings, direct answers, no burying the point in paragraph nine

This is the work behind Generative AI Optimization, and it is why the conversation with a business owner has changed. The question is no longer only where do I rank. It is: when someone asks out loud for a company like mine, does the answer include me?

The fair question: you are using AI, so shouldn't it be cheaper?

Business owners ask this constantly, and it deserves a straight answer instead of a defensive one. Sometimes yes, dramatically. Sometimes no, and here are two real examples from our own work that sit at opposite ends.

Example one: the retrofit

A client asked us to improve the SEO and AEO on the existing pages of their Duda website. Good goal. Even with AI helping at every step, that job ran five to seven hours per page. The content had to be audited, rewritten inside the constraints of the existing platform, restructured by hand, checked against what the page already ranked for so nothing was lost, and then fixed page by page in an editor that was never built for that kind of surgery.

Example two: the new build

On our current build process, a full page with SEO and AEO baked in from the first draft costs $50. Not five to seven hours of labor — a page produced correctly the first time, because the structure, the question coverage, and the internal linking are part of how the page is created rather than something bolted on afterward.

Same tools. Same people. Radically different cost. The difference is not AI versus no AI. It is building it right from the start versus repairing something that was not.

Where the savings are real, and where they are not

Being honest about this matters, because a blanket promise that AI makes everything cheap is how clients end up disappointed.

Where the savings are real

• —New websites and new pages, built with SEO and AEO from the start
• —Ongoing content and blogs, where the research and drafting load used to dominate the budget
• —Research, planning, and page architecture
• —Reporting and analysis that used to be manual

Where the savings are not

• —Retrofitting old pages on a platform that fights you
• —Cleaning up a site that was built badly, by a person or by AI
• —Anything that requires a human judgment call on brand, claims, or strategy
• —Fixing conflicting business facts scattered across a dozen places

So when a client asks whether AI makes it cheaper, the accurate answer is: it makes the right work far cheaper, and leaves the wrong work exactly as expensive as it always was. Which is a strong argument for rebuilding rather than patching.

AI built, with human guardrails — and a warning

Here is the part no agency selling AI likes to say out loud: not all AI websites are good, and a bad one can actively hurt your marketing. We have seen sites full of pages that say nothing, invented details no one at the business would ever claim, five pages that are the same page with a town name swapped, and no real answer to a single question a customer actually asks.

AI output is only as good as the prompt behind it. That is the whole game. A generic prompt produces a generic site, and a generic site does not get cited, does not convert, and quietly damages trust with both customers and the systems reading it.

Our prompts are backed by more than twenty years of marketing experience. What goes in — the structure, the question set, the local context, the claims we will and will not make, the way a service page has to convert — is the part that took two decades to learn. AI executes it fast. A human reviews every build before it goes anywhere near a customer. AI built, human guarded, in that order.

Why agencies have to move, not just can

Even if none of the above were true, the math on the other side of the table would force this change. Local businesses are absorbing higher costs on everything — labor, materials, insurance, ad auctions that get more expensive every year — while more competitors chase the same customers.

In that environment a marketing retainer that has not moved in five years is not a partnership. It is the easiest line on their P&L to question, and they are right to question it. Agencies can either reduce the number themselves, based on real efficiency, or wait to be cancelled.

Reducing our fees is not charity and it is not desperation. It is what honest pricing looks like when the cost of delivery drops by an order of magnitude.

What the new model looks like in real numbers

One client was paying $3,000 per month for marketing. That is $36,000 a year, every year, with no finish line.

We put them on our new program. For roughly $10,000 covering twelve months, they got a 100-page website, two managed Google Ads campaigns, and two rounds of blogs. Their annual marketing cost went from $36,000 to about $10,000 — an immediate $26,000 back in their pocket.

And the results were better than what they had before, because the build itself was structured for how customers and AI systems actually find a local business, not just prettier than the last one.

That is the structural difference. The old model charges you monthly for an asset you never finish owning. The new model builds the asset, hands it over, and keeps the ongoing cost small enough that the client keeps their own money.

An open invitation to other agencies

We would like more agencies to copy this. Not because we enjoy competition, but because the alternative is worse for everyone. Local businesses are closing while paying retainers that consume the margin they needed to survive. Every agency that keeps charging 2015 prices for 2026 production is taking a slice of that.

If you run an agency: run the numbers on what your work actually costs to deliver now. If the honest answer is a fraction of what you invoice, lower the invoice and win on volume, retention, and results. Clients who keep their money stay in business, and clients who stay in business stay clients.

Where to start if you are the business owner

Start with your own numbers. What are you paying monthly, what did you get for it in the last ninety days, and what would it cost to simply build the thing properly once and own it?

That is the conversation a strategy session is for. We look at what you are spending, what your search and AI visibility actually looks like, and what the honest cost of fixing it is. If staying where you are is the better answer, we will tell you that too.