Being found is more than a ranking
A business publishes helpful articles. Some appear in conventional search, while an AI answer cites different sources. Is the business invisible, or is it measuring the wrong thing?
First establish what is being observed. Source presence, actual visitors and business value are three separate questions.
What Bing is adding
On 13 October 2026, Bing introduced Topics Portfolio in beta within its AI Performance report. It groups topics for which a website appears as a source in the AI search activity being measured. Sufficient data is required. [1]
Microsoft warns that figures are approximate and AI-generated topic classifications may be wrong. This is not a complete measurement of every AI system, nor a guarantee that smaller websites already have a report. [1]
Evaluate three questions separately
I would initially treat a citation as evidence that a piece of content was referenced. Whether someone then clicks through needs separate investigation. A subsequent enquiry or order is another layer again.
Keeping those apart avoids misleading success claims. More mentions must not automatically be described as an equal increase in visitors. Equally, the absence of a visible reference in one answer does not establish that an entire site is generally excluded.
Organise content around readers' questions
Editorially, I would check whether an article answers a particular question, makes its information date clear and supports its claims. Understandable headings and terminology first benefit the people the site is intended for.
I would not stuff text with keywords to satisfy supposed AI rules. It is more useful to identify recurring unanswered questions and unnecessary competition between existing articles. A clear topic overview may help more than ten near-duplicates.
Observe changes in a traceable way
My approach would be deliberately limited: record a baseline, make a justified editorial change and later examine available data under comparable conditions. A single question to an assistant is not a reliable market-share test.
Ordinary website analytics remain important too. They do not automatically reveal every use of a text in an AI answer. Missing referrer information should not be replaced with invented attribution. A measurement gap remains a gap.
My conclusion
AI visibility deserves attention, not another round of unsupported promises. I would bring useful content, technical accessibility and honest measurement together. These suggestions do not guarantee inclusion in AI answers. This article contains no measurement of TNSystem's AI visibility.
Sources and context
Research date: 24 October 2026. Linked vendor and project information is not an independent product test. Examples and recommendations are identified as such.
[1] Bing: Topics Portfolio beta, 13 October 2026
Read source ↗