Named seven times a month, in a city of two million people.

Four different versions of the same business.

The first thing the audit surfaced was an identity problem. The company name appeared four different ways across the web — with and without "Ltd", with an old trading name, and with two different suffixes on directory listings. Two phone numbers were in circulation. The address on the website did not match the one on Google Business Profile, because they had moved units eighteen months earlier and never updated it.

That inconsistency is fatal for local visibility. Search engines and AI assistants both work by resolving a business to a single entity and then gathering corroborating signals about it. If the signals point at four entities, none of them accumulates enough weight to be trusted — so none of them gets recommended.

On top of that: one page listing eighteen suburbs as plain text, no LocalBusiness schema, a Google Business Profile with three photos and no service list, and 140 reviews that had never been responded to.

Local intent is the highest-converting traffic there is.

Someone typing "emergency plumber near me" is not researching. They have a problem right now and they will call one of the first two or three names they see. That is why local intent converts several times harder than generic commercial search — and why the shortlist being three names deep matters so much.

AI assistants make that shortlist shorter, not longer. Ask ChatGPT or Gemini for a local recommendation and you get two or three businesses in a sentence, with no scrollable list underneath. Position four does not exist. The businesses that get named are the ones the model can corroborate from multiple independent sources that agree with each other.

Local SEO stopped being a ranking tactic the moment assistants started answering in prose. It is now the qualification round.

One identity, eighteen real pages, a profile that works.

We picked a single canonical NAP — exact legal name, one address, one tracked number — and pushed it to the website, Google Business Profile, Bing Places, Apple Business Connect and 23 directories, correcting or removing every conflicting listing we could reach.

The Google Business Profile was rebuilt properly: full service list with descriptions, correct primary and secondary categories, opening hours including the out-of-hours policy, 40 geotagged job photos, products for each service type, and weekly posts. Every one of the 140 existing reviews got a response, and we set up an SMS review request that fires after job completion.

The single suburbs page became eighteen genuine service-area pages — each with the engineers who cover it, response times, real completed jobs with photos, local landmarks and parking or access notes, plus its own LocalBusiness and areaServed markup tied back to a single organisation @id.

Ranked by what actually shifted placement.