Answers, including the inconvenient ones.

AI search visibility is how often an AI assistant names your business when someone asks it for a recommendation. It is not a ranking position. An answer engine typically returns two or three businesses in a sentence or two — either you are one of them, or you are absent. There is no second page to be on.

GEO is the practice of making a business's information extractable, consistent, and corroborated so that AI answer engines cite it. Where traditional SEO competes for a ranked list of links, GEO competes to be one of the handful of names an assistant states directly in its answer. The work is different because the mechanism is different: models synthesize an answer from sources they trust rather than returning a list for a person to sift.

SEO optimizes for a ranked list of clickable links; GEO optimizes to be named inside a generated answer. They overlap — a fast, crawlable, well-structured site helps both — but they diverge in three ways. GEO weights machine-readable facts (structured data) more heavily, it punishes inconsistency between your site and third-party listings far more aggressively, and it rewards independent corroboration over on-page keyword targeting.

Four: ChatGPT, Google Gemini, Perplexity, and Google AI Overviews. We track multiple platforms deliberately rather than optimizing for one, because citation behaviour on any single engine can change with a model or product update. A business that only appears in one engine's answers has a concentration risk, not a result.

We test a fixed set of roughly a dozen prompts written in the language your customers actually use, run them against all four answer engines, and record whether your business is named in each response. That produces a citation rate — for example, four of twelve prompts. We run the identical prompt set for two named competitors so the number has a reference point.

The same prompts are re-run every month for retainer clients, which is what turns a snapshot into a trend.

Foundation work typically starts moving citation rate within eight to twelve weeks. Structured data and listing corrections are picked up relatively quickly; authority signals like reviews, press mentions, and content depth compound more slowly over several months. Anyone promising results in two weeks is either guessing or manipulating something they shouldn't.

No, and no one honestly can. Answer engines are third-party systems whose selection logic is undisclosed and changes without notice. What we can commit to is measurable movement on the inputs those systems demonstrably use — extractable facts, consistent listings, and independent corroboration — and transparent monthly reporting on whether that movement is happening. If it isn't, you'll see it in the same report we do.

Usually not. Most sites need structured data added, existing high-value pages restructured to answer questions directly, and content served in the initial HTML rather than assembled by JavaScript after load. That is remediation work on the site you already have. We recommend a rebuild only when the existing platform makes those three things impossible.

The audit is a fixed one-time fee; GEO Foundation and Authority & Citation Growth are monthly retainers, and Agent-Readiness is a setup fee plus monthly monitoring. We quote after the audit rather than before, because the audit is what reveals whether you need a light foundation pass or a full rebuild of your listing and content base. Ask us for the current bands and we'll send them.

Established local and professional service businesses that are already competent and already invisible to AI. Concretely: two or more years in operation, an existing base of genuine reviews, a real website even if it's dated, and a competitor who is being recommended by AI assistants when you are not. That last point is usually what prompts the call.

Brand-new businesses with no review base, and anyone looking to buy a guaranteed placement. GEO amplifies real signals — reviews, longevity, genuine third-party mentions. If those signals don't exist yet, the honest advice is to build the underlying business first and come back in a year; we'll say so rather than take the retainer.

Agentic commerce is an AI assistant completing a purchase or booking on a customer's behalf, inside the chat, rather than sending them to your website. It is real but immature: major platforms have already introduced and revised their approach within the past year. Agent-readiness means structuring your catalog — product and service schema, accurate machine-readable feeds — so you are eligible when the protocols stabilise. We treat it as infrastructure, and we do not forecast sales from it.

No, because we exclude the tactics that would. We do not create fake reviews, invent locations or addresses, register domains resembling competitors', or misrepresent your credentials, licensing, or history to any AI system. Those tactics carry legal and reputational exposure, and they contradict the entire mechanism we're selling — an answer engine cites you because it trusts your information, and manipulation is what destroys that trust. Everything we do is work you could describe out loud to a customer.

Monthly, against three numbers in a deliberate order. Citation rate is the leading indicator — it moves first. AI-referral traffic and branded search activity are confirming indicators — they show the citations are reaching real people. Attributed new customers is the revenue number — customers who told you they found you through an AI assistant. Every report also includes the competitive comparison against the same two or three named competitors, so the number always has context.