Active vs inactive
- Active: included in assistant grounding for live responses.
- Inactive: retained for reuse but ignored at runtime.
- Use inactive state as a safe rollback path during experiments.
Assistant Contexts
Control what factual context your assistant carries into every answer
Use context entries to add targeted reference material, toggle it safely, and keep assistant behavior aligned with operations.
Context entry system
Assistant Contexts let teams store factual entries that are provided as reference material during answer generation. Each entry has a title and content field, lives in the assistant settings library, and can optionally be linked to knowledge gap resolution workflows for traceability.
Title and content fields
Each entry is a focused pair of title and content, designed for quick authoring and clear retrieval context.
Editable context library
Manage all context entries from assistant settings with a centralized, searchable library view.
Knowledge gap linkage
Optionally link entries to resolved knowledge gaps so teams can trace why context was added.
Active and inactive control
Each context entry can be toggled active or inactive so teams can test messaging safely and phase out stale entries. Active entries are included in assistant grounding for live responses, while inactive entries are retained for reuse but ignored at runtime — providing a safe rollback path during experiments.
Per-entry active toggle
Switch individual entries on or off without deleting them, enabling controlled rollout of new context.
Immediate influence control
Changes to active status take effect immediately, giving teams real-time control over assistant behavior.
Safe rollback path
Use inactive state to preserve context entries during experiments — re-enable them if response quality regresses.
Composition strategy support
Contexts work best when entries are scoped, non-overlapping, and clearly owned by a team for maintenance. Keep each entry focused on a single topic, assign clear ownership to avoid conflicting instructions, and schedule periodic reviews to prune stale or redundant entries.
Context scope by topic
Write entries that cover one specific topic each, avoiding broad entries that overlap with other knowledge assets.
Ownership and conflict reduction
Assign clear ownership per topic to prevent contradictory context from appearing in assistant responses.
Periodic review hygiene
Schedule quarterly context audits to remove stale entries and keep the context library lean and effective.
Entry state and composition strategy
Use context entries as scoped operational guidance. Keep them short, non-overlapping, and actively managed.
Overlap avoidance
- Do not duplicate the same statement across contexts, snippets, and pages.
- Keep one owner per topic and review quarterly for conflicts.
- Move long policy text into pages or custom answers, then reference it.
2. What it is
Assistant Contexts are modular entries of factual guidance attached to the assistant. Active entries are included as reference material in behavior grounding.
3. When to use it
- Adding operational facts that do not fit page/snippet structures.
- Encoding support-team constraints or escalation preferences.
- Resolving knowledge gaps with targeted context additions.
4. When not to use it
- For long-form user-facing guides (use Knowledge Pages).
- For exact legally sensitive language (use Custom Answers).
- For baseline product framing (use Product Profile).
5. Setup steps in app
- Open Assistant settings and go to Behavior and Context.
- Create a context entry with clear title and scoped content.
- Set entry active and save.
- Test assistant output on targeted scenarios.
- Toggle inactive if response behavior regresses or conflicts arise.
6. Best use cases
- Escalation guidance for sensitive account situations.
- Operational rules for support boundaries and next steps.
- Temporary campaign-specific context without rewriting full docs.
7. Common pitfalls and how to avoid them
- Pitfall: overlapping entries with conflicting instructions. Avoid with clear ownership and dedupe reviews.
- Pitfall: overloading contexts with long content. Avoid by keeping entries focused and linking to pages where needed.
- Pitfall: leaving stale entries active. Avoid with quarterly context audits.
8. Success metrics to track
- Reduction in answers that violate support policy or tone guidelines.
- Decrease in recurring gaps tied to missing operational context.
- Faster resolution for scenarios targeted by context entries.
9. Related capabilities
- Product Profile for baseline business framing.
- Knowledge Pages for structured long-form support guidance.
- Custom Answers for deterministic exact responses.
- Knowledge Gaps for identifying where context is missing.
Which option should you use?
Use this matrix to choose the right knowledge option for each support intent.
| Intent | Best option | Why | Link |
|---|---|---|---|
| Exact policy wording | Custom Answers | Deterministic response for high-risk phrasing. | /custom-answers |
| Short operational facts | Snippets | Fast to publish and update. | /snippets |
| Multi-step guides | Knowledge Pages | Better for long-form structure and context. | /knowledge-pages |
| Synced external docs | Knowledge Sources | Best for URL crawl and document ingestion workflows. | /knowledge-sources |
| Baseline product framing | Product Profile | Keeps answers aligned with core positioning. | /product-profile |
| Operational constraints and fallback rules | Assistant Contexts | Adds scoped behavioral context without rewriting docs. | /assistant-contexts |
FAQ
10. Assistant context FAQs
How are Assistant Contexts different from Product Profile?
Product Profile defines baseline business framing; Assistant Contexts add modular operational facts and constraints. Learn more
Can inactive entries be kept for later reuse?
Yes. Inactive entries remain stored but are not used as active reference material until re-enabled. Learn more
Should I duplicate the same fact in pages, snippets, and contexts?
Prefer one source of truth per fact and cross-reference where needed to avoid contradictory answers. Learn more
Launch a conversation layer your whole GTM team can use.
Unify grounded AI, web and WhatsApp coverage, human handoff, and conversation analytics without stitching together separate tools.