Basis Technologies' 2026 Advertising Agency Report landed on April 28 and the numbers do not flatter the industry. Across 200+ agency executives, 87% expect the traditional agency model to collapse within 3–5 years. Among senior decision-makers the figure is 92%. AI usage sits at over 99% of agencies, daily AI use jumped from 38.6% to 58.2% in a year, and 46% are already deploying agentic AI in live workflows. Forty percent have conducted layoffs in the past twelve months. Optimism about digital advertising's future fell below 50% for the first time the survey has run.
Five questions worth sitting with
Take a paid campaign through to launch, the way it actually happens this quarter, and answer these in order.
1. If 99% of agencies use AI daily and 46% deploy agentic AI, what is the human still examining before the ad ships? Not approving — examining. There is a difference between reading a brief and clicking "publish." The first is a control. The second is paperwork.
2. When creative volume per account doubles and headcount falls 40%, what exactly is being scaled? Output, certainly. But output of what — alignment, or volume? A factory that doubles production while cutting QA is not twice as productive. It is twice as fast at producing whatever it produces, including the defects. Digiday's coverage of the same data framed the trend as agencies "turning creators into test labs" — execution moved out of the studio and into agentic loops, with no clear answer on who reviews the output before it ships.
3. If the creative governance layer was the first thing cut, was it ever treated as a control? Or was it always a cost that survived only because nobody had measured what it caught? A control whose value is invisible is indistinguishable from a cost — until something it would have caught reaches a customer.
4. If a Claude or ChatGPT agent writes the ad copy and a separate team owns the landing page, who has read both? The agent does not read the page. The landing page team does not see the ad variants. The media buyer sees performance metrics but not the creative-to-page comparison. The single person who used to hold both in their head is no longer on the payroll.
5. The agency model is forecast to collapse — but what survives the collapse? Whatever produces a result the buyer cannot get from the AI directly. Execution at volume is no longer that. Diagnosis, governance, and proof of post-click experience might be.
These are not rhetorical. The instinct is to answer them quickly. The Socratic move is to refuse to.
The unexamined campaign
Socrates' move was always to expose the gap between what we claim to know and what we actually know. The unexamined life is not worth living, he said, because an unexamined life is one we have only assumed is good. The same logic applies, with less drama, to a paid campaign.
Most performance teams assume their ads and landing pages are aligned. The assumption is structural rather than evidenced. The ad is approved in one tool. The page is approved in another. The two are rarely placed side by side. When they are, it is usually because CPA has drifted upward and someone is hunting for the cause.
The Basis numbers make the assumption riskier than it has ever been. With 46% of agencies running agentic workflows, ad variants are now produced faster than any human can read them. With 40% of agencies cutting headcount, the people who would have read them are gone. The gap between "ad shipped" and "landing page checked against the ad" has widened from a few hours to, in many cases, never.
This is what AdAlign measures. Ad-to-page congruence on a 1–10 scale across visual match, content consistency, and tone continuity. Not whether you assume the page delivers on the ad — whether it actually does. The signs of misalignment tend to look like rising CPA, falling Quality Score, and a creative team being told their work is the problem when the page is.
The philosophical claim and the operational one converge here. Examined alignment leads to predictable performance. Assumed alignment leads to the kind of campaign post-mortem where everyone is in a meeting trying to figure out which lever moved.
What survives the collapse
If the Basis prediction is correct — and 92% of senior decision-makers think it is — the work that survives is the work the buyer cannot extract from a model. A Claude agent will draft the ad. A ChatGPT agent will iterate the variants. The buyer will run them. What the buyer cannot do, alone, is verify that the page receiving the click is congruent with the ad that produced it. Not at the volume modern campaigns now run.
This is the specific thing AdAlign was built for, and it is also the specific thing the agency governance gap leaves uncovered. We wrote about how this changes the next 6–12 months in the future of ad-to-page governance, and about a tactical 48-hour fix in the navigation drift article. Both are companion pieces to this one.
The questions above do not have neat answers. They are not supposed to. The point is to notice that the industry has been answering them by default — assuming that the QA layer was redundant, that volume scales linearly with quality, that platforms will catch the drift. The Basis report says the assumption has stopped being safe.
Run an examined audit
Run an alignment audit on one live campaign. Three free analyses, no setup. You will see the specific visual, content, and tone gaps between your ad and your landing page on a 1–10 scale. Whether the result is reassuring or not, it will be examined rather than assumed. Run your free audit.
### Frequently asked questions
What is ad-to-page congruence? Ad-to-page congruence is the degree to which a landing page delivers on the specific promise made in the ad that drove the click. It is measured across three dimensions: visual match (does the imagery and design echo the ad), content consistency (does the headline, value prop, and CTA match), and tone continuity (does the page sound like the same brand the ad introduced). AdAlign scores it on a 1–10 scale.
Why are agencies predicting their own collapse? The Basis 2026 report finds 87% of agency professionals — and 92% of senior decision-makers — believe the traditional agency model will collapse within 3–5 years. The drivers cited include over 99% AI adoption, 46% deploying agentic AI in workflows, 65.3% losing work to client in-housing, and 40% conducting layoffs in the past twelve months.
How does AI in agencies affect landing page alignment? AI tools generate ad creative faster than human teams can review it, while landing pages are usually built and updated on a separate cycle by a different team. The result is that creative volume rises while alignment review shrinks. Misalignment between ad and page becomes structural rather than occasional.
What does "examined alignment" mean? Examined alignment means systematically scoring whether your ads and landing pages match — visually, in content, and in tone — rather than assuming they do. The Socratic framing is borrowed from "the unexamined life is not worth living": an ad campaign whose alignment has not been checked is one whose performance you are guessing at, not measuring.
How do I check if my ads and landing pages are aligned? Upload an ad screenshot or paste an ad library URL, enter the landing page URL, and AdAlign returns a 1–10 congruence score across visual, content, and tone, plus a list of specific mismatches and prioritized fixes. The free tier includes three analyses with no setup.