AI SEARCH OPTIMIZATION · SAN DIEGO
Get Named When the Answer Replaces the Search Results
ChatGPT, Google AI Overviews, Perplexity and Gemini increasingly answer instead of listing. We work on whether your business is the one they quote — and we measure it.
OUR OWN NUMBERS, NOT AN INDUSTRY STATISTIC
Impressions Down 44.6%. Clicks Up 9.0%.
That is the year-on-year picture across the client sites we manage in Google Search Console, and it is the clearest evidence we have that the shape of search has changed. Far fewer people are being shown a list of ten blue links. Average position improved from 18.1 to 9.5 and click-through nearly doubled from 2.07% to 4.08%, so the people who still click are further down the buying process.
The rest of that lost volume did not stop looking. It got answered — by a summary that read somebody's page and named somebody's business. This page is about being that somebody. We show you these numbers rather than a vendor statistic because you can ask us to pull the same report for your own property.
WHAT THE WORK ACTUALLY IS
You Cannot Optimise for a Model. You Can Be the Best Source.
There is no ranking dial inside ChatGPT and anybody selling you one is guessing. What these systems do is retrieve, read and summarise sources, then attribute. So the job is to be the most retrievable, most quotable, most obviously credible source on a specific question.
In practice that is unglamorous: clear factual claims a model can lift without hedging, specific numbers instead of adjectives, structured data that says exactly what you are and where, an entity footprint that agrees with itself across your site, your listings, your profiles and third-party mentions, and pages that load for a crawler that will not wait.
Most of it also happens to be good SEO. The difference is that ranking rewards being on the page and citation rewards being unambiguous.
WHAT WE DO
AI Search Optimization Services
Also sold elsewhere as GEO, AEO or LLM optimization. The acronym keeps changing; the work below does not.
Answer Engine Audit
We run the prompts your buyers actually use across ChatGPT, Perplexity, Gemini and AI Overviews, and record who gets named, what gets said about you, and which sources the answers are pulling from.
Entity & Schema Work
Organization, Service, LocalBusiness, FAQ and Person markup implemented properly, plus making your name, address, category and claims consistent everywhere a model might read them. Contradictions are why you get skipped.
Quotable Content
Pages built around extractable answers: a definition in the first paragraph, real figures, comparisons stated plainly. Produced with our content team so it also ranks conventionally.
Citation Monitoring
Tracked over time: which prompts name you, which name a competitor, and whether the sentence about you is accurate. Being described wrongly is a different problem from being absent, and it needs a different fix.
Third-Party Signal
Models weight what other sites say about you. Editorial coverage, directories, review platforms and industry listings — run with PR — are often the fastest route into an answer.
Retrieval & Technical
Server-rendered content, sane HTML structure, fast responses and crawler access. Retrieval systems are less patient than Googlebot — content that arrives only after JavaScript often is not read at all.
HOW AN ENGAGEMENT RUNS
The Formula for AI Search Optimization
Four stages on a monthly loop. The first and fourth are the same test, which is the only reason the middle two can be judged.
01. Baseline
A recorded set of buyer prompts run across the major assistants, so we know who is being named today before we touch anything.
02. Fix the Foundation
Schema, entity consistency, crawlability and the factual claims on your site. Most gaps are here and most are unglamorous.
03. Publish Quotable Answers
New and rewritten pages that answer the prompts directly, with the specifics a model needs in order to cite you rather than hedge.
04. Re-run & Report
The same prompt set, monthly. You see movement, or you see that there was none and what we think that means.
WHAT WE WILL NOT CLAIM
The Honest Limits of This Work
This is a young discipline and the market is full of people pretending otherwise.
Nobody can guarantee a citation. Answers vary by phrasing, by user, by model version and by week. Anyone quoting you a guaranteed placement rate in ChatGPT is selling you a number they cannot control.
Measurement is imperfect. There is no Search Console for AI assistants. We measure by running a fixed prompt set on a schedule and recording results — a sample, honestly labelled as one, not a complete picture of what every user sees.
It is not a replacement for SEO. The underlying signals overlap heavily, and organic search still drives most of the traffic we are accountable for. We sell this alongside SEO, not instead of it. If someone offers to move your budget out of search entirely, ask them what happens to your revenue while they are being right in theory.
LOCAL ANSWERS, LOCAL SOURCES
“Best in San Diego” Is Now an Answer, Not a List
When somebody asks an assistant for the best law firm, restaurant or contractor in San Diego, they get three or four names and a sentence each. There is no page two. Being eleventh in Google was survivable; being unnamed in a four-item answer is not.
Local answers lean heavily on sources these systems trust for place-based questions — your Google Business Profile, review platforms, local editorial and directories — and they punish inconsistency between them harder than classic local SEO ever did. A wrong category or an old address is now a reason to be left out of the answer entirely.
We are at 770 First Avenue downtown, we work these signals for San Diego businesses every week, and we can show you what an assistant currently says about your category before you commit to anything.
We work with businesses across Downtown, La Jolla, Carlsbad, Encinitas, Chula Vista, Oceanside, Escondido and Coronado — the full list is on our service areas page.
Grow Your Brand With Our Marketing Formula
Discover how our expert strategies can elevate your business to the next level.
Not big on forms? Call us at (619) 995-8333
"*" indicates required fields
What our Clients are saying
By far the best marketing company in San Diego! They are on top of it- the social posts are on time and the website updates and SEO are effective. They built a WordPress website for us and we couldn’t have been happier- highly recommended
Formula Marketing has been a phenomenal company to work with. The entire team is professional and very talented. They are clearly passionate about the work they do and have helped our business create an amazing social media presence, which has ultimately increased our sales tremendously. I highly recommend working with Formula Marketing!
Best marketing company in San Diego. They helped take my business to the next level. With Formula Marketing SD I truly feel like I have an amazing team on my side helping me every step of the way.
QUESTIONS
AI Search Optimization FAQs
What is AI search optimization?
The work of getting your business named and quoted inside AI-generated answers — ChatGPT, Google AI Overviews, Perplexity, Gemini, Copilot — rather than only ranking in a list of links. It is also sold as GEO (generative engine optimization), AEO (answer engine optimization) and LLM optimization. Same work, different acronym.
Is this just SEO with a new name?
Substantially overlapping, but not identical. Ranking rewards relevance and authority on a page; citation rewards being unambiguous, structured and quotable at the sentence level. A page can rank fourth and never be quoted, and a page can be quoted constantly while sitting outside the top ten. Anyone who tells you they are the same thing is not measuring the second one.
Can you guarantee we will appear in ChatGPT?
No. Answers change by phrasing, by user history, by model version and by week, and no vendor controls any of that. What we can do is measure where you stand on a fixed prompt set, fix the reasons you are being skipped, and show you the movement on the same test.
How do you measure whether it is working?
A fixed set of buyer prompts, run across the major assistants on a monthly schedule, with results recorded: were you named, in what position, and was the description accurate. Plus referral traffic from assistant domains in analytics, which is real but still small for most businesses. We label the sample for what it is rather than dressing it up as a share-of-voice percentage.
How much does it cost?
Usually $1,500 to $4,000 a month as a layer on top of an SEO or content engagement, depending on how many prompts and markets we are tracking. As a standalone it rarely makes sense — most of the fixes are on your site, and you will want somebody accountable for conventional search at the same time.
How long does it take?
Schema and entity fixes can change how you are described within weeks, because that information is re-read relatively often. Getting newly named in a competitive answer usually takes a quarter or more, because it depends on third-party sources and content that has to exist and be trusted first.
Should we block AI crawlers instead?
That is a real strategic choice and it depends on your business model. If your content is the product — publishing, research, paid media — there is an argument for it. If you sell a service and want to be found, blocking retrieval means guaranteeing you are never the answer. We will talk it through honestly rather than assume.
Does schema markup really matter for this?
It matters more here than it does for ranking. Structured data removes ambiguity about what you are, where you operate and what you offer, and ambiguity is the most common reason a model quotes somebody else. It is also one of the few parts of this work that is entirely under your control.
We already rank well. Do we need this?
Ranking well is the strongest possible starting position and it is not sufficient on its own. Look at your own Search Console: if impressions are falling while position holds or improves, you are watching the same pattern we see across our portfolio. The demand is still there; it is being satisfied before the click.
Do you work with businesses outside San Diego?
Yes. The local signal work is San Diego-specific, but entity, schema, content and retrieval work is not tied to geography. Most of our clients are here; several are not.
Find Out What the Assistants Say About You
Send us five questions your customers ask before they buy. We will run them across ChatGPT, Perplexity and Google's AI answers and send you back exactly who gets named — whether or not it is you.