Marketers have a new number on the dashboard. AI visibility: how often ChatGPT or Gemini names your brand when a customer asks for a recommendation.

Good. It is the right thing to measure. More than 900-million people use ChatGPT every week, according to OpenAI. Independent analysis suggests the large majority of AI-mode searches now end without anyone clicking anything at all. Gartner expects brands to lose half their organic search traffic by 2028. If you are not in the answer, you don't have the opportunity to even be considered by your customers.

For any brand with more than one location, a single brand-level AI visibility score based on a few keywords is close to useless. Here is why. Ask an assistant where to eat in Durbanville and it does not consult a national brand ranking. It assembles an answer from what it can verify about the businesses near Durbanville. Ask the same question in Sandton and you get a different answer, built from different data, against a different competitive set. There is no national AI result to rank in. There are as many results as you have trading areas.

This becomes a challenge for marketing, the stores missing from the answer are losing walk-ins to a competitor three blocks away, and nobody upstairs knows which stores they are.

So what does better measurement look like?

Measure per location, not per brand. Every outlet gets its own score, against the brands main competitors as well as the local competitors in their neighborhood.

Use contextual queries, not near me. Best burgers in Durbanville is what people type. Best burgers near me is what tools default to, and it tells you nothing you can act on.

Then aggregate, per keyword. Track best coffee across 200 locations and you want two views: the score for each location, and one number for that keyword across the entire footprint.

We built this into our own AI Visibility Tracker because clients kept asking, "which of my stores are missing, and where". Each location is scored against its local competitive set, results roll up per keyword to a brand-level view, and we have created an algorithm to set the weighting between being mentioned at all and where they appear in the answer and if the brands website is being linked to the answer. This is the first South Africa platform that gives per-location AI visibility, together with an aggregated brand roll up specifically for multi-location brands.

But measurement is the easy half, and this is the part the industry is getting wrong.

A score tells you a store is invisible in Bellville. It does not fix the wrong trading hours on Apple, the eleven unanswered reviews on Google, or the address information that isn't formatted correctly on Bing. AI assistants are not doing anything mystical. They are reading the same location data, reviews and local content this industry has been managing for a decade, and declining to recommend businesses they cannot verify. Accuracy of your info across multiple channels has become more important than ever.

Which is good news, if you spent that decade getting the fundamentals right. The brands that win AI answers will not be the ones with the cleverest AI strategy. They will be the ones whose data is already accurate, reviews scores were impressive and they have localised content — store by store, before anyone thought to ask a chatbot where to have lunch.

Start there. Then measure it properly, one suburb at a time.

For more information, visit www.socialplaces.io. You can also visit Social Places on Facebook, LinkedIn, X, or on Instagram.

*Image courtesy of contributor