Status: Original concept, first publication. Strategy Sandbox, jasonbarnard.com. Date: 16 May 2026.

Gate 10 (Won) is the tenth and final gate in the 10-gate AI Engine Pipeline, and the entire purpose of every preceding gate (Discovered, Selected, Crawled, Rendered, Indexed, Annotated, Recruited, Grounded, Displayed) is to deliver the brand to the moment of resolution where a real outcome happens in the real world: a sale, a booking, a subscription, a contract, the thing the brand actually wanted from the visibility chain.

For the past twenty-eight years the SEO industry has talked about the Won outcome as if it were a single event: someone clicked, someone bought, someone signed up. That framing was adequate when there was only one way to win, but the way the Won gate resolves has changed twice already since 2022, and the third resolution arrived in 2025 and changes the rules entirely.

There are three Won resolutions, and they belong to three different eras: the Imperfect Click of the Search Era, the Perfect Click of the Recommendation Era, and the Agentic Click of the Agent Era. All three resolutions co-exist in 2026. The first two are still the dominant patterns by volume. The third is the smallest by volume and the largest by strategic consequence, because the third changes who actually decides.

Imperfect Click: The Search Era Resolution

The Imperfect Click is the resolution every SEO professional has been optimising for since Google indexed its first page in 1998. The user types a query, the search engine returns a list of options ranked by some combination of relevance and authority signals, and the user makes the final decision by clicking one of the options.

The click is called imperfect because the human has to do most of the evaluation work. The engine narrows ten million possible pages down to ten visible results, but those ten results still represent ten different decisions the user has to weigh against each other. Which result is most relevant to my actual intent? Which source is most credible? Which one looks like a marketing page and which one looks like a useful answer? The cognitive load of the final filtering sits with the human.

For a brand, the Imperfect Click is the most generous of the three resolutions. The brand gets to be one of ten options, gets to write the meta description that appears in the results, and gets to design the landing page the user arrives on. There are many moments of intervention between the search and the sale, and every one of those moments is a chance to win the customer who hasn’t quite decided yet.

The Imperfect Click is also the most familiar, and familiarity is dangerous. Brands assume that what worked in the Imperfect Click era still works in the resolutions that came after, and they assume wrong.

Perfect Click: The Recommendation Era Resolution

The Perfect Click is what happens when the user asks ChatGPT, Perplexity, Claude, or any other assistive engine for a recommendation, and the engine returns a single answer (or a tight group of two or three answers) instead of a list of ten. The user looks at the recommendation, evaluates it briefly, and either accepts it or rejects it. The cognitive load of filtering has shifted from the human to the engine.

The click is called perfect because, from the user’s perspective, the engine has done the work the human used to do. There is no list to scan, no ten-result audit, no comparison between options on different pages. The recommendation arrives pre-filtered and pre-evaluated, and the user is asked to confirm or reject rather than to choose from a slate.

For a brand, the Perfect Click is dramatically less generous than the Imperfect Click. The brand is no longer one of ten. The brand is either the recommended option or it is not recommended at all. The middle ground (visible but not chosen) has been eliminated. There’s no meta description to optimise, because the engine wrote its own summary of the brand based on whatever it learned during training and grounding. There’s no landing page to design, because the user may not even visit the brand’s site before deciding.

The Perfect Click is where the Untrained Salesforce starts to cost real money. When seven AI engines are all making recommendations at scale, and each one either names the brand or names a competitor, the difference between trained and untrained representation becomes a direct revenue line.

Agentic Click: The Agent Era Resolution

The Agentic Click is the resolution that arrived in 2025 with the launch of agentic commerce platforms, and it is structurally different from the other two. The user does not click, and the user does not see the recommendation. The user delegated the decision to an agent, and the agent decided and executed without returning to the user for confirmation.

The pattern goes like this. A user tells their AI agent to book a flight to Singapore for Tuesday morning, optimised for cost and arrival time. The agent queries the network of bookable inventory, evaluates options against the user’s stated criteria, applies whatever stored preferences the user has accumulated over time, and books the flight. The user receives a notification that the flight is booked. The brand that received the booking was never seen by the human at any point in the funnel.

For a brand, the Agentic Click is the most adversarial of the three resolutions, because there is no moment of human intervention to influence. The agent’s decision is made before the brand has a chance to present itself. The agent’s criteria are set by the user (cost, time, schedule) and weighted by what the agent learned during training, and the brand has zero opportunity to argue its case during the decision itself.

The brand’s only intervention point is upstream: long before the agent runs the query, the brand had to make sure the agent’s training data, recommendation logic, and stored preferences already favoured the brand. Algorithmic preparation is the only intervention point that exists in the Agentic Click resolution.

Why the Third Resolution Changes Everything

For me, the strategic shift between the Perfect Click and the Agentic Click is the biggest single change in commercial communication since search engines existed. The Imperfect Click era gave the brand ten possible moments to win: search snippet, landing page, second visit, comparison page, retargeting, email, phone call, demo, proposal, contract. Every moment was an intervention opportunity, and the SEO industry built twenty-eight years of practice on the assumption that there would always be ten.

In the Perfect Click era, the number of moments collapsed to roughly two: the recommendation itself, and the user’s optional click-through to verify. Most of the intervention opportunities disappeared. In the Agentic Click era, the number of moments is exactly one, and it happens before the agent ever runs the query. The agent’s preferences, training data, and stored memory determine the outcome of every transaction the user delegates, and once the agent has decided, there is no further conversation. The brand either trained the agent successfully or it did not, and the consequence of failing to train the agent is not visible to the brand until revenue starts disappearing into competitors who did train the agent.

Assistive Agent Optimisation is the discipline that names what brands have to do in response. The work is not optimisation in the traditional SEO sense, because there is no SERP to rank in and no recommendation to influence at the moment of decision. The work is preparation: making the brand legible, credible, and recommendable to the agent at the layer of training data, integration, and stored user memory, long before any specific user query happens.

Three Resolutions, One Pipeline

All three resolutions co-exist in 2026, and the brand that wins is the brand that prepares for all three simultaneously. The Imperfect Click still drives the majority of commercial outcomes by volume, and the SEO disciplines that have worked for twenty-eight years still apply. The Perfect Click is growing fast and now drives a significant minority of high-intent commercial decisions in categories where AI recommendations are trusted, like travel, software, and professional services. The Agentic Click is small in 2026, but it is growing on a curve that suggests it will be a dominant pattern in five years, and the structural change it introduces means brands cannot wait until it is dominant to start preparing.

The brand that prepares for only the Imperfect Click wins less every year, the brand that prepares for only the Perfect Click wins more but loses ground in agentic transactions, and the brand that prepares for all three wins consistently across the transition.

The Won gate has always been the gate that mattered. The way the gate resolves is what has changed.