Configure AI Search contextual filter preprocessing for Recommended Actions
Configures AI Search to preprocess contextual inputs from Recommended Actions so search results are dynamically filtered and boosted based on the current record context.
Antes de Iniciar
Role required: admin
Por Que e Quando Desempenhar Esta Tarefa
When recommendations are generated, Recommended Actions passes contextual information to AI Search using the Additional Context payload. After receiving this payload, AI Search performs preprocessing before executing the search. Preprocessing is an internal preparation step that ensures contextual inputs are interpreted in a structured form, made available to AI Search dynamic filtering logic, and evaluated against search results during execution.
At this stage, AI Search does not return results. Instead, it prepares the contextual data so it can be applied consistently during filtering and ranking. Once preprocessing is complete, the prepared contextual data is used by AI Search dynamic filters. Preprocessing supports logic evaluated through the AisDynamicFilter extension point, including filter condition evaluation (preprocessing). This enables AI Search to determine whether a search result is relevant to the current context record.
Preprocessing results in CI‑based filtering, meaning that search results that do not match the contextual configuration item (CI) information are filtered out, and only results relevant to the current record context are included. In addition to filtering, preprocessing enables CI‑based boosting, which prioritizes results that are more relevant to the current context and ranks contextually aligned results higher in the search output.
Procedimento
Resultado
In the Search tab of the Recommended Actions context side panel, only the search results that match the exact context of the current record appear.