Workspace Assistant
General read-and-draft help across the visible workspace, including status questions, project context, and safe draft preparation.
AI Workspace Assistant
Ask Avanqi about the work it can see.
The AI Workspace Assistant is a permission-scoped, read-and-draft chat interface for change management teams. It can summarize visible work, suggest next actions, cite source pages, and prepare drafts for review.
Agent
Project Co-pilot
Draft an update for the operating model project.
I can prepare a project update draft using visible status, milestone, and communications context. I will not save it automatically.
Project overviewTimeline
Draft card
Project update ready for review
Open the existing form with the draft or copy it into your current workflow.
What it can do
A practical AI layer for change management.
The assistant is designed for status, summarization, suggestions, and reviewed drafting. It does not replace the project record.
Ask for a status summary across the visible portfolio.
Find projects or milestones that need attention this week.
Get suggested next actions before a governance meeting.
Draft a stakeholder communication for an existing project.
Prepare a project update or milestone draft for review in the existing form.
Agent modes
Select the assistant for the job.
Agent profiles expose smaller tool sets to the model, so each mode stays focused on the work it is meant to perform.
Portfolio Analyst
Summarizes visible portfolio status, health, milestones, and attention areas with source chips back to Avanqi pages.
Project Co-pilot
Focuses on one visible project, checks project-level context, suggests next actions, and drafts milestone or project updates for review.
Comms Drafter
Drafts stakeholder communications from visible project context and never claims messages have been sent.
Governance Reviewer
Reviews visible project and portfolio signals for missing status, milestone, and communication information without inventing findings.
Security first, then convenience.
The assistant follows the same tenant and project permission model as Avanqi. It is not a general API browser.
The model receives allowlisted tool schemas, not arbitrary API access.
Every tool runs server-side as the signed-in Avanqi user through existing workspace and project guards.
The assistant cannot pass or override organization IDs, user IDs, roles, or project-scope IDs.
Tool results are minimized and redacted before being sent to the model.
Draft cards require human review; V2 keeps no automatic writes in the chat flow.
Production AI is gated behind provider configuration and confirmed enterprise data-control settings.
Ready to try it in context?
Sign in, open a workspace or project page, then launch the AI chat button. Your current page context travels with the request.