AI Recommends Your Competitor: Here’s How to Fix It

The latest episode of the Search Engine Journal podcast features Kris Jones, a search veteran who has been in the industry since the late 1990s and currently runs the agency LSEO after successfully building and selling Pepper Jam. Our conversation kept circling back to a single concept that determines whether AI systems will recommend your brand, and Kris delivered the answer in one word early on: synthesis. You can catch the full episode here.

AI Took Over The Comparison Step

Traditional search worked by presenting you with options, leaving the evaluation process entirely in your hands. You would open multiple tabs, weigh the alternatives, and make your own decision. Kris explained that AI systems have now assumed that comparison role and return a finished recommendation to the user. The old model gave you choices and required you to synthesize the information yourself. AI changed the game by performing that synthesis for you.

This places users much closer to the end of their research journey than a standard page of blue links ever could. If the model includes your brand, the buyer arrives already pre-sold. If it omits you, they never know you exist. I described this shift on the episode as moving from a card catalog to a concierge. A library used to hand you the 20 best resources in ranked order and send you off to read. Now something reads the top hundred sources for you and comes back with a single recommendation.

Kris acknowledged that this involves a genuine leap of faith. People are accepting these synthesized answers and acting on them without ever checking the underlying sources.

Why Brand Is The Bridge

When I asked Kris where these systems pull their information from, his answer was brand signals, specifically what other people are publishing about you. That reframes the entire nature of the work. You are no longer competing for a position on a results page. You are trying to become the answer itself, and the raw material for that answer lives on websites you do not control.

Kris also quantified how much of your existing program still matters. In his estimation, 70% to 80% of AI search optimization is fundamental SEO practice, with the remaining 20% to 30% covering newer third-party efforts. His advice to marketers was blunt: do not fire your SEO company. This aligns with how the industry arrived at this moment. Google spent years training everyone on what it rewards, from featured snippets to structured markup. As I noted on the episode, the shift over the past five years from writing for people to writing for what Google rewards is largely what trained these models.

Treat this as a layer on top of what you already fund. The budget shift is smaller than the vendor noise suggests, and most of it belongs in earned media.

The Awards Your Competitors Are Applying For

Kris identified the cheapest immediate step available to most businesses, one that plenty of marketers have quietly dismissed. Identify every credible third-party award and recognition available in your industry, then go get them. He knows the objection. Many of these programs are pay to play, and the same names seem to win every year. His response stuck with me: those are the companies that are showing up in the AI recommendations.

The same logic applies to comparison and best-of coverage on publications with real authority. Kris argued that when an established publisher in a category writes up the top platforms in that space, the piece becomes one of the strongest sources a model can draw on later. One thing worth saying plainly. Publishing your own “best agencies in our category” post with your name at the top is spam, and Kris called it that himself. The value here depends on the source being independent of you.

Brand Messaging Is The New NAP Consistency

Getting mentioned is not the finish line. Kris kept returning to the context of a mention, meaning what the surrounding content actually says about you. He compared it to the early days of local SEO, when we all learned that a business name, address, and phone number had to match everywhere they appeared. The same discipline now applies to your brand messaging. What does the web say your product does, what does it say you charge, and does any of that still match reality? What has changed is that AI search output now feeds a system answering customer questions directly, which raises the cost of leaving stale information out there.

When AI Overviews Gets Your Pricing Wrong

I have a concrete example of what that costs. I worked with a chain of adult education schools whose tuition figures were being pulled into Google AI Overviews from a Reddit thread that was roughly a decade old. The number showing up was about 30% below what the schools actually charged. The lost conversions were the obvious problem. The bigger mess was everything downstream. Prospective students called in confused, and some were annoyed, wanting to know why the price had gone up. Nobody had raised the price. A forum post from 10 years ago was setting false expectations. It takes publishing accurate current information in enough credible places that the AI models have multiple corroborating sources to work from instead of one stale thread.

Don’t Set Your Baseline From ChatGPT And Google

One tip I want people to take out of this conversation is to widen the list. When you get into your Google Analytics data, the click rate coming from ChatGPT will probably baseline lower than you expect. It grows from here, and so does the traffic from every other engine people are using. The surfaces keep multiplying. Kris mentioned using Grok in his Tesla, which makes obvious sense once you connect in-vehicle AI to navigation and local search. Meta AI is sitting right inside Instagram and Facebook in front of an enormous audience, and it is not an engine our industry talks about much.

Behavior is what is driving that. I rarely see the people around me typing into these apps. They hold a conversation with ChatGPT and take the answer, and as Kris pointed out, children especially use voice search. Kris uses voice AI regularly himself, and his read is that a spoken question carries the same intent as a typed one. The sources shaping those answers matter just as much either way. If you size your AI visibility budget against the referral traffic those engines send you today, you are working from a number that is small now and rising. Kris made the point that the benefits of this kind of work play out over time, but it is best to get started while AI search is still young. The benefits can be a force multiplier down the road.

What Should Marketers Tackle First?

Two things from this episode are worth acting on before your next planning cycle. The first is an audit. Ask an AI assistant to recommend a vendor in your category, read what it says about you, and check whether any of it is out of date. The second is the awards list you have been ignoring. Kris’s argument is that the companies collecting those recognitions are the ones the models cite. You can decide to go after them this quarter, and it does not require building anything new.

Traditional search is still here and still worth funding. What changed is that the comparison step moved, and the inputs to that comparison are sitting on other people’s websites.