Ask an assistant for a recommendation and it searches, reads what it can, decides what it trusts, and writes a short answer naming a few businesses. Four decisions happen in that sequence, and every one of them is decided by your website.
Two years ago the job was to rank in a list and win the click. Every instinct that came from that era works against you now. Here is what we find, over and over, on sites that were perfectly well made.
"We deliver world-class solutions tailored to your needs" cannot be quoted, checked or trusted. It says nothing a machine can carry into an answer. What you do, for whom, where, and starting at what price, can be.
A page about everything is the answer to nothing. When a question has a page of its own, with a heading that matches the question, the retrieval step has something to find.
Schema is the machine-readable version of your business. Most sites either have none or have whatever a plugin generated, which usually names the software and not the services you sell.
Page builders wrap content in layers of styling divs with headings used for size rather than structure. It looks right to a person and reads as noise to everything else.
A Google Business Profile that disagrees with the site, service areas listed in one place and not the other, an address from two moves ago. Contradictions get a business quietly dropped from consideration.
The most common finding of all. Nobody has asked the assistants what they say about the business, so nobody knows. It takes twenty minutes to find out, and it is usually the moment the conversation gets serious.
The industry has not settled on a name yet. You will hear all three, often from people selling the same work. We call it AI visibility because that is the outcome you are buying.
The oldest of the three. Emphasis on writing content that directly answers a question, so it can be lifted into a featured snippet or an assistant's reply.
The academic term. Emphasis on being retrieved and cited by systems that generate an answer rather than return a list. Same mechanics, different vocabulary.
The catch-all, and the one most agencies have adopted. Broad enough to cover the content, the schema, the technical work and the corroboration together.
Where they genuinely differ is scope, and the scope that matters is the widest one: content, structure, schema, speed and corroboration, done together. Anyone selling one of those five alone is selling a fifth of the job.
In order, and always starting with the same twenty minutes of looking rather than a proposal.
Nobody can guarantee an assistant will recommend you. Anyone who does is guessing with your money.
No one outside those companies controls the models, and the models change without notice. What can be controlled is whether your business is easy to find, easy to understand and easy to trust, and whether you are named more often this quarter than last. That is the work, and that is what the report shows.
Every site we build carries this structure from the first sitemap, not as an add-on afterwards. It is the reason our builds start with questions rather than a homepage layout.
Custom website design, from $2,997Content, schema upkeep, local and citation work, competitor gap analysis and the monthly answer report. Visibility fades as models refresh and competitors publish, so this is ongoing work by nature rather than a project.
SEO and AI visibility program, from $1,497/moWhether AI assistants name your business when someone asks a question you should be the answer to. It is measured by how often you appear in those answers, not by where you sit in a list of links. The industry also calls it AEO, GEO and AIO, and they describe the same work.
SEO earns a position in a list someone then chooses from. AI visibility earns a mention inside a written answer, where there is usually no list and no choice. The technical groundwork overlaps heavily, so good SEO helps, but ranking and being quoted are different outcomes measured in different ways.
Because it was written to rank, which means persuasive prose, thin service pages and no machine-readable facts. An answer engine needs the opposite. A site can be new, attractive and fast and still have none of what it looks for.
No, and be suspicious of anyone who says otherwise. Nobody outside those companies controls the models. What we control is whether you are findable, understandable and trustworthy to the systems assembling the answers, and whether you get named more often over time.
A fixed list of about twenty questions your buyers would ask, checked every month across the major assistants. We record whether you were named, who was named instead, and what the answer said about you. Same list every month, so the trend means something.
Not always, but often. If the site is structurally sound this work goes on top of it. If it was assembled in a page builder with unreadable markup and no schema, a rebuild costs less than fighting the platform every month. We tell you which case you are in before you spend anything.