An architect asks an AI assistant how to put an early price on a house extension. The assistant might offer a rough figure, recommend a specialist estimating service, or use a connected tool to produce an estimate within the conversation. Whether an outside business becomes involved depends on the response.
This is an easy part of AI discovery to miss. Before a supplier is brought in, the conversation has to reach a point where its help is needed. Someone asking what an extension might cost could be satisfied with a general answer. Someone needing an estimate from drawings has given the assistant a more specific problem, which might lead it towards a specialist.
For businesses, that raises a question: under what circumstances does AI bring us into the conversation? AI discoverability optimisation (AIDO) means helping AI find a business, understand what it offers and select it when it fits the need.
Search engine optimisation still provides much of the groundwork. Pages need to be accessible and useful, with clear information about what a business offers. Google says its existing SEO principles also apply to its AI search features.
Being absent from an answer can have several explanations. The assistant may have missed the business, misunderstood what it does, or found something better suited to the request. It may have provided enough help itself. Each possibility calls for a different response, so counting mentions alone leaves much of the story unexplained.
There is a revealing test here. Ask an assistant to suggest services for a particular job. Then introduce a business it left out and ask whether it could help. It may assess the business favourably once named. That suggests it could have been overlooked, and gives us something to investigate. The assessment still needs checking against what the business can actually deliver.
At SR3H, these questions are shaping Signal, our work on AI visibility and analytics: understanding where customer needs create opportunities for businesses. We start with situations that lead people to seek help, then use repeatable question batches to investigate how businesses are represented. Comparing different systems, contexts and times can help assess which observations hold. This could show which patterns hold: when AI brings in a specialist, who it considers and where a suitable option seems to be missing. Understanding how often real customers encounter those situations requires further evidence.
Update, 29 September 2026: AIDO Labs is now Signal.
The next step is to investigate what could improve the outcome. Perhaps a service needs a clearer explanation, stronger evidence, or a way for AI to use it directly. Sometimes the product itself needs to change. We would record what changed and repeat the tests, using evidence from enquiries, trials and purchases to judge whether it made a useful difference.
The opportunity may begin several questions before anyone asks for a supplier. Understanding those earlier questions could change both how a business explains its work and what it chooses to offer.