Conversational Help
Summarize
Summary of Conversational Help
Conversational Help is a generative AI-powered skill integrated into the ServiceNow Otto panel, designed to provide users with quick, contextual answers to their questions. This skill is enabled by default for users with appropriate roles and requires no additional subscription as part of generative AI entitlements. It appears as theGet Helpoption within the Otto panel and can be accessed either by selecting the skill or submitting a query directly.
Show less
Key Features
- Generative AI Integration: Uses advanced AI to interpret and respond to user queries by searching relevant ServiceNow product documentation.
- Automatic Skill Discovery: The system identifies the most appropriate skill to answer a query based on keywords, enhancing accuracy.
- Customizable Skills List: The skills displayed can be tailored in collaboration with integration teams to suit organizational needs.
- Centralized AI Search: Queries are routed to a central ServiceNow instance that searches synchronized product documentation via Instance Data Replication (IDR) from Now Support instances.
- Relevance Optimization: Retrieved content chunks are re-ranked using the Now Assist Q&A capability to improve answer relevance.
- Cache Utilization: Frequently asked queries are cached to provide faster responses when applicable.
- Release-Specific Results: Answers are generated based on the version of the user’s current ServiceNow instance, ensuring information is accurate and up-to-date.
- Control via AI Admin Hub: Administrators can disable the feature if needed through AI Admin Hub Settings.
How It Works
- User submits a question through the Get Help skill or directly in the Otto panel.
- The query is sent to the central ServiceNow instance where AI Search scans the knowledge table containing synchronized product documentation.
- Top matching content chunks are sent to Now Assist Q&A for re-ranking to optimize relevance.
- If the query exists in the cache, the cached answer is returned immediately.
- If uncached, the query and re-ranked content are sent to the Now LLM (Large Language Model) which generates a meaningful and contextual answer.
- The response is displayed within the Otto panel, reflecting the documentation and features of the user’s current release version.
Benefits for ServiceNow Customers
- Enables faster and more accurate self-service support within the ServiceNow platform.
- Improves user productivity by providing contextual help without leaving the Otto panel.
- Ensures help content is aligned with the specific ServiceNow release version, reducing confusion from outdated information.
- Offers flexibility to customize skills and manage feature availability through administrative controls.
This skill uses Generative AI application capabilities to provide answers to the questions on the ServiceNow Otto panel.
- Select Get Help skill and submit a query to use the help option in the ServiceNow Otto panel.
- Submit your query directly in the ServiceNow Otto panel. It will perform discovery and find the most appropriate skill among the listed, based on the keywords you entered. This discovered skill is used to retrieve answers for your
queries.Note:The available skills displayed in the ServiceNow Otto panel can be customized in consultation with the integration teams.
How Conversational Help works
Your query goes through the following steps to retrieve the best response.- AI search
The query is sent to central ServiceNow instance, where AI Search will search the knowledge table that stores content from product doc content.
We use Instance Data Replication (IDR) that synchronizes product documentation content from Now Support instances.Note:Multiple central ServiceNow instances are deployed across three regions: US East or US West, EMEA, and APJC. - The top matching chunks (or records) along with the user’s query, are sent to the Now Assist Q&A Genius Result configuration to generate a meaningful, contextual answer.
- Re-ranking
The Now Assist Q&A, an out of box (OOB) capability, re-ranks the retrieved chunks to ensure better relevance.
- Cache lookup
- If the user query exists in the cache, the pre-stored answer is returned.
- If the query is not cached, the query and re-ranked chunks are sent to Now LLM to generate a meaningful answer.
For more information, see Fetch end points in Now Assist Conversational Help skills.