In Dynamics 365 Customer Service, Knowledge Search and Co‑Pilot currently use separate mechanisms to determine which articles are surfaced. Knowledge Search relies on user‑selected filters, whereas Co‑Pilot uses AI‑driven retrieval based on intent and metadata. Because these systems operate independently, users often experience misalignment between the articles surfaced by Knowledge Search and those returned by Co‑Pilot.

As a result:

• Co‑Pilot may return articles outside the filtered parameters set by the user.
• Co‑Pilot does not reliably infer context directly from the Case record.
• Agents must manually refine or rewrite their prompts to enforce the filters already visible in the UI.

This reduces efficiency and undermines user trust in Co‑Pilot’s accuracy.

Business Justification

1. Improved Retrieval Relevance

Agents expect Co‑Pilot to adhere to the same filters applied in Knowledge Search. Aligning these systems ensures more accurate article suggestions and eliminates surprise or irrelevant results.

2. Context-Aware Assistance

Co‑Pilot should automatically interpret Case details such as product type, scenario, scheme, keywords, or customer attributes. This enhances personalised retrieval and reduces manual effort.

3. Workflow Efficiency

By synchronising filters and context automatically, agents avoid repeating steps or adjusting prompts multiple times, improving speed and user adoption.

4. Consistency Across Tools

Implementing automated filters ensures the Knowledge Search interface, Co‑Pilot chat experience, and in‑case Knowledge panel all operate from the same contextual baseline.

Requested Enhancement

1. Automated Filter Mirroring

Co‑Pilot should automatically adopt active Knowledge Search filters, including:

• Topic or category
• Product, skill, or process filters
• Date or relevance filters
• Saved/pinned filter sets
• User‑defined view preferences

2. Intelligent Context Inference from Case Data

Co‑Pilot should apply structured inference rules that read from the Case record, such as:

• Case title and description
• Case classification fields (product, type, reason, scheme, channel)
• Any applied tags or metadata
• Past agent behaviour on similar cases

These contextual elements should automatically influence article ranking or filtering.

3. Admin Configuration Panel

Introduce settings that allow administrators to manage:

• Which Knowledge Search filters sync with Co‑Pilot
• Which Case fields Co‑Pilot considers “high‑priority” context
• Behaviour for specific roles, teams, or business units
• Whether Co‑Pilot defaults to AI‑only mode, filter‑only mode, or a hybrid mode

4. Filter Transparency for Users

Co‑Pilot should display a small visual indicator describing:

• Which filters were applied
• Which Case fields were inferred
• Why the surfaced articles match the user’s situation

This provides clarity and trust for agents.

This enhancement enables Co‑Pilot to intelligently honour user‑selected filters and context derived from active Case details. By unifying Knowledge Search and Co‑Pilot behaviour, agents gain more relevant results, reduced manual effort, and a smoother, more intuitive workflow.