Evidence, language, and local trust

A teaching composite from one tram route: three clinics described the same service in three different ways, and one note in my book had a suburb name squeezed into the margin. One clinic used a clinical phrase nobody searched. One used a vague comfort line that could have belonged to a day spa. One buried the useful detail in a receptionist's answer on the phone. I wrote them down because the pattern bothered me: capable local businesses were doing the work, but their public language could not carry the facts far enough.

I am from Adelaide, and I have spent 17 years around the practical end of local visibility: local SEO audits, service-page architecture, review mining, directory cleanup, small-agency strategy, and editorial training for owner-operated firms. I have worked with the sort of businesses where the owner still knows which staff member handles the difficult jobs, which suburbs bring the wrong enquiries, and which service needs three sentences of explanation before anyone feels safe booking. That mess is not a flaw. It is the material.

These days I am strongest where messy business reality meets structured language. What does a plumber actually do when the job is not the neat version on the service page? What is a patient afraid to ask? Which claims can an answer engine safely repeat because the evidence is visible, corroborated, and specific? I keep a hand-built answer ledger that tracks how AI systems describe the same business category across suburbs, seasons, and intent types. It is slow work. Good. Fast summaries are often where local businesses get flattened. I treat generative engine optimization as a craft of evidence: service pages, FAQs, reviews, business profiles, staff bios, and case notes all need to agree without sounding like they came from the same tin of paint.