AI search shifts competition from clicks to delegated decisions

The delegation boundary is now the real competition

Jason Barnard’s sharpest point is that AI search is no longer just a retrieval problem. It is a delegation problem: how much of the decision-making journey a user hands over to the machine. That boundary matters because the old search habit was simple, you asked, scanned, compared, clicked, and decided. In AI search, the system can narrow the field, recommend a choice, and in some cases complete the transaction before the user ever reaches a website.

That changes the scoreboard. A click on a blue link is no longer the only meaningful win, and sometimes not even the best one. The new prize is being the option the system feels confident enough to carry forward on the user’s behalf.

Why this is different from classic SEO

Traditional SEO rewarded visibility in a results page. AI search rewards something more like machine trust. If a model can parse your product data, understand your category fit, verify your claims, and see enough supporting signals to feel safe, you are more likely to be surfaced as the answer, not just a result.

That is why Barnard frames the issue as a spectrum. Some people will still browse, compare, and click. Others will provide a natural-language prompt, add context, and let the engine do the heavy lifting. The strategic question is no longer only “How do we rank?” It is “At what point does the user stop evaluating and let the machine choose?”

For brands, that means the real contest is for algorithmic confidence. The machine has to believe your data is clean, your offer is understandable, your reputation is durable, and your product is the safe recommendation.

Google has already moved the experience closer to action

Google has been pushing this direction hard. At I/O 2024, it said AI Overviews were rolling out to everyone in the United States and were designed to take more of the legwork out of searching. Google also said those AI Overviews had already been used billions of times in Search Labs before the broader rollout, which tells you the company was not treating this as a side experiment. By October 2024, Google said AI Overviews had expanded to more than 100 countries.

That is the practical shift Barnard is talking about. Google is not just answering queries faster. It is compressing the path from question to conclusion, and that shortens the space where a user might have clicked out to compare alternatives. When the search engine handles more of the work itself, the brand has fewer chances to win the old-fashioned way.

OpenAI and Microsoft are making the delegate-and-buy model feel normal

OpenAI has pushed the model even further into commerce. In September 2025, it said more than 700 million people turn to ChatGPT each week for everyday tasks, including finding products they love. In the same move, it launched Instant Checkout in ChatGPT, starting with U.S. Etsy sellers and saying over a million Shopify merchants were coming soon.

By March 2026, OpenAI said ChatGPT shopping was expanding again, with richer visual discovery, side-by-side comparisons, and merchant integration through the Agentic Commerce Protocol. That matters because it moves the platform from suggestion to shopping workflow. The assistant is not just surfacing options, it is helping organize them, compare them, and increasingly route the user toward purchase.

Microsoft is clearly chasing the same logic with Copilot Shopping. Its materials describe a system that can compare prices, track products, and make purchases with in-line checkout. Once a platform can compare, track, and transact, the delegation boundary gets much thinner. The machine is not merely helping the user think. It is beginning to act.

Publishers and marketers are feeling the traffic squeeze

The backlash has been immediate. In May 2024, the News/Media Alliance asked the FTC and DOJ to investigate Google’s generative AI Overviews, arguing that the feature could starve publishers of traffic. That complaint gets at the economic core of the issue: if the user gets enough of the answer in the interface, the original source may never receive the visit it used to depend on.

Pew Research Center added a useful reality check in 2025. It found that Google users were less likely to click links when an AI summary appeared, and they very rarely clicked cited sources. That is the clearest evidence yet that AI summaries change behavior, not just layout. The user’s journey stops earlier, and that means the old traffic funnel is getting squeezed from the top.

For brands and publishers, that is the hard truth. Visibility is no longer only about ranking below the fold. It is about being selected inside the answer layer, where the user may never leave the platform.

What makes an AI confident enough to recommend you

If the delegation boundary is the new strategic question, then the editorial question for brands becomes very practical: what signals make a machine comfortable enough to hand the decision to you?

The answer is not hype. It is structure, clarity, and trust.

• Machine-readable facts matter. Product names, specs, pricing, availability, and category fit need to be easy for systems to parse without guesswork.

• Sourceable claims matter. If a model can trace a claim back to a reliable, consistent source, it is more likely to repeat it.

• Authority matters. Brands with clear expertise, recognizable proof points, and stable reputations are easier to delegate to than noisy outliers.

• Transaction readiness matters. If the platform wants to complete the purchase inside the interface, the brand has to be ready for that workflow.

• Consistency matters. Conflicting data across feeds, pages, or merchant listings can make a system hesitate, and hesitation kills delegation.

Barnard’s framing is useful because it makes the brand challenge feel less mystical. You are not trying to impress a chatbot with clever marketing language. You are trying to make your business legible enough that the system can safely carry the user across the line.

The regulatory backdrop is catching up

The governance side is moving too. The EU AI Act, Regulation (EU) 2024/1689, was adopted in June 2024 and published in July 2024. It establishes a legal framework for AI systems in the European Union and includes transparency rules, provider and deployer obligations, and provisions aimed at trustworthy AI.

That matters because delegation is not just a UX issue. Once AI systems influence recommendations, comparisons, and purchases, questions about transparency and accountability become unavoidable. If an AI steers a user toward one product over another, the stakes are no longer only commercial. They are also about disclosure, responsibility, and the rules governing how much decision-making power a system can wield.

The new playbook is built for handoff, not just discovery

The old search game was built for discovery. The new one is built for handoff. Google’s AI Overviews, OpenAI’s shopping features, and Microsoft’s Copilot all point to the same direction: platforms want to do more of the legwork, not less. The practical effect is that brands must compete for a place inside the decision, not just inside the results.

That is the big shift hidden inside the delegation boundary. Winning now means being the brand the machine trusts enough to recommend, summarize, and sometimes transact on behalf of the user. In AI search, the click is becoming just one possible outcome, and not always the one that matters most.