AI skills

An AI Assist skill is a specific, named AI task — “summarize this incident,” “draft resolution notes” — rather than a general-purpose chat. Skills are grouped by the application they belong to, and most can be scoped to a set of tables.

The skill catalog

Application	Skill	What it does
IT Service Management	Incident Summarization	Summarize the record from its short description, work notes, comments, and activity.
IT Service Management	Resolution Notes Generation	Draft resolution notes from the record’s targeted fields when it moves to a Resolve state.
IT Service Management	Work Notes Generation	Draft a work note from recent activity on the record.
IT Service Management	Ask About a Record	Answer questions about the current record (timeline, impact, likely resolution).
IT Service Management	Change Summarization	Summarize a change request, its risk, plan, and approvals.
Customer Service	Case Summarization	Summarize a customer case from its details and interactions.
Customer Service	Chat Summarization	Summarize a live-agent chat conversation into the record.
Customer Service	Case Resolution Notes	Draft resolution notes for a customer case.
HR Service Management	HR Case Summarization	Summarize an HR case for the assigned HR agent.
Security Operations	Security Incident Summarization	Summarize a security incident, indicators, and remediation status.
Platform	Knowledge Article Generation	Generate a knowledge article from a resolved record.
App Builder	Code Generation	Generate or explain a script (server rule, client-side script, flow).

Most skills apply to a specific set of tables: Incident Summarization covers Incident, Problem, and Change Request; Case Summarization and its siblings cover the Customer Service case table; HR Case Summarization covers the HR case table; and so on. Code Generation is the exception — it’s a developer skill with no record table attached.

How a skill reaches an agent

Four of the skills above (Incident Summarization, Resolution Notes Generation, Work Notes Generation, and Ask About a Record) map onto a quick-action bar that appears whenever the AI Assist panel is opened in the context of a record:

• Summarize record — a summary generated from the record’s fields, work notes, and comments.
• Generate work notes — a draft work note from recent activity, for the agent to review and post.
• Generate resolution notes — a draft resolution summary from the record’s fields, meant for a record moving toward resolution.
• Ask about this record — a free-text question about the specific record open in the panel.

Every one of these is a draft: the panel always frames AI output as something to review, not something already applied.

Beyond the panel — other places AI Assist shapes structured output

A few other surfaces feed a description or a document to the same AI layer and get back structured data instead of a chat reply:

• AI-powered list filtering (“Ask AI”). Typing a natural-language request over a list turns it into a validated, structured filter for that table — or, if the request is ambiguous, a couple of short clarifying questions.
• App Designer. A plain-language app description is turned into a structured blueprint (tables, fields, states, forms, lists) that App Designer can then build.
• “Fill with lens” document extraction. A photographed or scanned document is read and its values are mapped onto the matching fields of a create form, transcribed faithfully rather than translated or guessed.

Activation

Skills are turned on or off, and browsed by application/category, from the Skills tab of the AI Control Center. Each skill card shows its activation status (Not started, Out-of-box, or Custom) and an Activate/Deactivate action, so an admin can see and manage the organization’s AI skill footprint from one place.