Written by Roman ChornovolSource inclusion is the measure of whether a brand's evidence, claims, and perspectives are structurally woven into the frame an AI model builds about a category, not merely referenced in passing.
Also known as: source inclusion
What Source Inclusion Is
When an AI engine answers a category question, it doesn’t retrieve facts neutrally. It adopts a narrative frame built from the sources it has learned to trust. Source inclusion measures whether your content, data, and voices are among the inputs that build that frame, or whether they sit outside it entirely.
This is a sharper diagnostic than a mention count. A brand can appear in an AI answer, a brief citation, a name in a list, without any of its framing, evidence, or positioning having shaped the answer’s logic. True source inclusion means the model’s understanding of the category reflects your evidence: your definitions, your distinctions, your point of view. Absent that, you are decoration on someone else’s narrative.
Why It Matters for AI Search
AI search doesn’t rank; it recommends. The recommendation follows from a learned consensus about the category. That consensus is built upstream, from the sources the model absorbed before any query arrives. If your sources are excluded from that consensus-formation layer, no amount of downstream optimization changes whose story the model tells.
This is why source inclusion is an upstream metric. Mentions are a symptom; source inclusion is a cause. Brands that win AI recommendations do so because their evidence is embedded in the frame the model adopted, not because they appeared in enough answers after the fact. Chasing visibility without auditing source inclusion means treating the symptom while the cause compounds.
Source Inclusion and Narrative Share
Mavel treats source inclusion as the mechanism behind narrative share: the measure of whose story the model actually tells about a category. A high mention rate with low source inclusion signals a dangerous gap. The brand is present in AI answers but absent from the frame shaping them, which means a competitor’s narrative is doing the explanatory work.
Understanding source inclusion means tracing the actual sources and citations an AI draws on when it constructs its category answer. From there, you identify where your evidence is woven in versus where it’s absent or overridden. That audit converts a vague visibility score into a specific, actionable finding: which sources carry your frame into the model, which don’t, and where the narrative gap lives. That is the difference between a dashboard that describes a problem and intelligence that locates its cause.