Assistant brings the outside world into your copilot.
Your teams ask about a stakeholder or an issue, in Genie or in their own copilot over the Model Context Protocol. Each answer draws on resolved identities, your engagement history and delivered intelligence, with its sources attached.
Capabilities
Ask in your copilot. Draw on the field and your own record. Act where you asked.
- 01 · Ask in your copilot
Ask about a stakeholder, an issue or a meeting without leaving your copilot.
Nobody learns a new tool. A request in your copilot or your own agent becomes defined calls into Genie, and the answer comes back where the question was asked.
- Defined tools
- Stakeholder, signal and engagement search, briefs and delivered intelligence. The agent calls each by name.
- Any MCP client
- The enterprise copilots and agent frameworks your teams run.
- In the flow
- A pre-read before a meeting, a profile mid-draft, a watch set up from a chat.
- 02 · The outside world, consolidated
Bring the outside world in: who they are, what they have said, and what has moved.
The external context your AI cannot assemble for itself: one profile per stakeholder with roles current and former, their positions by issue, the relationships around them, and the signals that moved this week, classified to your issues.
- Resolved identities
- One profile however the name appears, with time-valid roles. The agent never briefs yesterday’s title.
- Positions and relationships
- Where they stand on each issue and who they report to, advise, fund and sit with, as data.
- What moved
- The signals on the issue or the stakeholder, classified and ranked, not a web search.
- 03 · Your own context beside it
Add your own context: engagement history and delivered intelligence.
The agent builds on what your organisation has said and promised to the stakeholder, and on the briefs and reads Lens has produced. Each user is answered from what that user can see in Genie.
- Meetings, calls and commitments
- Against the person, from the record your teams keep.
- Delivered intelligence
- The briefs and reads Lens has produced, in reach of the agent.
- Your permissions, exactly
- By team and market, the same as in Genie.
- 04 · Act from your copilot
Shape what Genie tracks and delivers for you, from the conversation.
A saved filter with its alert, or a newsletter, drafted in your copilot, shown as a draft, and committed to the workspace only once you approve it. Every answer carries its sources and every write is logged to the user.
- Drafted, then approved
- Nothing saves on the agent’s say-so.
- Lands in Genie
- The approved item appears in the workspace your team uses.
- Sourced and logged
- Every answer cites its items; every call and write is logged with its inputs and outputs.
How it works
Governed context in, sourced answers out.
- 01
Natural requests
Users stay in their copilot. Genie is called when stakeholder context is needed.
- 02
Structured tools
The copilot calls defined tools for stakeholders, signals, engagement and delivered intelligence, and gets structured, sourced results back.
- 03
Governed access
Authentication runs through your Genie account and its permissions. Your AI sees what the user can see.
- 04
Approved write-back
A filter, an alert or a newsletter drafted in your copilot is shown as a draft and saved to Genie only once approved. Nothing enters the asset unseen.
Reads from the platform
- Stakeholders
- Signals
- Engagement
- Lens outputs
Writes back to the platform
- Saved filters
- Alerts
- Newsletters
Specifications
Your copilot, your workflow, and the external context it was missing.
What an IT or AI platform team asks first. Scope and pricing are set per engagement.
- Protocol
- Model Context Protocol.
- Context
- Resolved identities with time-valid roles, your engagement history, delivered intelligence, and a source behind every statement.
- Tools
- Stakeholder, signal and engagement search, briefs, delivered intelligence, saved filters with alerts, newsletters.
- Clients
- Any copilot or agent framework that speaks MCP.
- Authentication
- Your Genie account and its permissions, by team and market.
- Write-back
- Saved filters, alerts and newsletters, approved before they save. Every write logged to the user.
- Data protection
- Keys specific to your organisation. Zero-retention model endpoints. Nothing of yours trains a model.
- Setup
- Connected with your IT or AI platform team.
Questions
Frequently asked questions
Which AI clients can connect?
Any copilot or agent framework that speaks the Model Context Protocol. Your IT team connects it with ours.
How is access controlled?
Through your Genie account and its permissions. A user’s AI sees exactly what that user can see in Genie.
Does Genie replace our AI?
Genie supplies the governed, sourced context. Your AI supplies the reasoning. How we compare sets out where each class of tool is strongest.
What does this give our AI that a news or web connector does not?
Resolved identities with current and former roles, your own engagement history under your permissions, the intelligence delivered in Lens, and a source behind every statement. A connector gives your AI text to sort out for itself.
Give your AI the context it’s missing.
Tell us which copilots and agents your teams run. We will propose the connection, the data scope and the setup with your IT team.