Recorded once
Renewal variance brief
Turns licence-renewal exports into an executive variance brief with evidence links and a decision list.
5 measured takings · Microsoft Skill Recorder export · illustrative scenario
Aristo's read Grounded in page evidence
Owned agents and reusable skills sit together with provenance, reach and attestation attached.
Agents & skills
The agents working across Contoso, the governed skills people can reuse, what each asset is for and what it is allowed to reach.
Agents built herecounted
4
forged by your own people, from a sentence
Source: Aristo agent register As of: 2026-08-29
People who have built onecounted
fewer than 5
withheld below 5, as every group of people is
Source: Aristo agent register As of: 2026-08-29 Groups below 5 people are shown as a fact without a number.
On the skill shelfcounted
6
reused 9 times — skills, not agents
Source: Aristo skill shelf As of: 2026-08-29
Governed reusecounted
4
4 skills taken by somebody · 2 still shelfware
Source: Aristo skill governance projection As of: 2026-08-29 Counted from the existing skill-governance projection. “Taken” means at least one recorded download or Graph-confirmed push; shelfware exists but has no measured taking.
Who works here that is not a person
Every agent this page can see, counted once, in the section it belongs to.
Where each agent came from
current register agents Counted
2 agents (16.7%)4 agents (33.3%)3 agents (25.0%)3 agents (25.0%)
Evidence & data — Where each agent came from
Comparison
Every agent this page can see, counted once, in the section it belongs to. An agent nobody has registered anywhere is an agent this product cannot see at all — which is the sentence at the top of this page.
Source & method
Aristo's own register, your tenant's agent registry and the Copilot catalogue read.
Basis: Counted
Accessible data
| Origin | Agents | Share of total |
|---|---|---|
| Provided with Aristo | 2 | 16.7% |
| Built here | 4 | 33.3% |
| Registered in your tenant | 3 | 25.0% |
| In the Copilot catalogue | 3 | 25.0% |
Where yours sit in review
Your own workflow, not a run state.
Agents built here, by review stage
current register agents Self-attested
1 25%
1 25%
1 25%
1 25%
Agents built here, at each stage of your own review
Evidence & data — Agents built here, by review stage
Comparison
A stage is a record of YOUR process, not a statement about whether an agent is running — Microsoft exposes no run state for one, and this product never invents it.
Source & method
The agent register's own lifecycle records.
Basis: Self-attested
Accessible data
| Stage | Agents |
|---|---|
| In review | 1 |
| ODAT | 1 |
| Approved | 1 |
| Draft | 1 |
Each stage is recorded by the people who moved the agent through it. Self-attested, and labelled as such wherever it appears.
Skills and Skill Recorder
Three ways in — prompted, adapted, or recorded — and one governed shelf. Inventory and measured taking stay separate.
Prompted here
2
1 ship-ready
Found and adapted
2
2 ship-ready
Recorded once
2
1 ship-ready
Taken by somebody
4
downloaded or Graph-confirmed push
Still shelfware
2
exists, but nobody has taken it
Movement
3 more takings than last month.
2026-08 vs 2026-07
Bring a recorded skill
In a connected tenant, a signed-in leader can paste or upload the recorder's SKILL.md here. This illustrative scenario is read-only, so it shows the governed shelf without offering a write that would pretend to work.
How recording and privacy work
The recording happens on your own machine — the recorder captures your screen locally.
Its Analyze step is the part to know about: that one sends screen frames, window titles and URLs, and your spoken narration to GitHub’s cloud to turn the demonstration into a skill.
So never record credentials, secrets, or a screen with tenant-sensitive data on it. Record the job, not the sign-in.
Aristo never receives or stores a recording. The only thing that reaches us is the finished SKILL.md you choose to bring — that is true by architecture, not by policy.
Browse the governed skill shelf6 skills · quality, provenance and measured taking
91/100 · ship-ready
Found and adapted
QBR storyliner
Adapts a governed presentation workflow into a concise quarterly business-review narrative.
2 measured takings · Approved internal skill library · illustrative scenario
Prompted here
Vendor risk synopsis
Builds a consistent risk synopsis from a procurement questionnaire without inventing missing evidence.
1 measured taking · Created in Aristo · illustrative scenario
Recorded once
Month-end close coach
Records the close checklist and escalation path; two quality checks still need revision before release.
No measured taking yet — shelfware, not “used” · Microsoft Skill Recorder export · illustrative scenario
Found and adapted
Adoption office-hours planner
Adapts an enablement playbook into role-based sessions, prompts and follow-up measures.
1 measured taking · Approved internal skill library · illustrative scenario
Prompted here
Contract redline triage
Draft skill held for review because its escalation and source-boundary instructions are incomplete.
No measured taking yet — shelfware, not “used” · Created in Aristo · illustrative scenario
There are no per-person rows here. Maker cohorts remain subject to the 5-person privacy floor, and a skill is never called “used” merely because it exists.
What Microsoft does not measure per agent
Microsoft exposes no per-agent usage and no per-agent value, so this page counts what exists and refuses to model what does not: there is no "runs this week" figure here, and no dollar attached to anybody's agent.
Explore every agent inventoryProvided, built here, tenant registry and Copilot catalogue
Provided with Aristo
Ready to upload today. These ship with the product — nobody in your organisation has to maintain them.
Aristo — Executive Intelligence
Run state not exposed
Provided with Aristo
Microsoft 365 Copilot agent, provided with Aristo
Board-ready answers on your organization’s Frontier Firm score, strategy, maturity, department chargeback, and license position, grounded in the same counted signals as Aristo’s reports.
Owner: Aristo
the description inside the package itself
Copilot chat
The Executive Score
Run state not exposed
Provided with Aristo
Microsoft 365 Copilot agent, provided with Aristo
A personal working-habits assessment for executives: it examines the evidence available in your Microsoft 365 permission boundary and applies Aristo’s Executive Score rubric, with coverage and confidence shown.
Owner: Aristo
the description inside the package itself
Copilot chat
Built here
Every agent Aristo has built for one of your people, newest first. 4 in the register.
Renewal brief
Run state not exposed
Built here
Microsoft 365 Copilot agent
Pulls the seat position, the idle block and the renewal date into one page a week before any supplier call.
Owner: Ada L.
the purpose its maker wrote when they asked for it
Microsoft 365 Copilot
Can reach
your OneDrive and SharePoint
attested
Risk tier
Low risk
attested
Governance stage
In review
History (2)
- Built from its maker's description · 2026-08-23
- Moved from Draft to In review · 2026-08-26
Not shown, and why: How often this agent is used, and by how many people. Microsoft exposes no per-agent usage signal, and Aristo will not model one. The package download. The illustrative organisation hands out nothing — in your own tenant this is a one-click sealed link.
Bid librarian
Run state not exposed
Built here
Microsoft 365 Copilot agent
Finds the closest past proposal to a new one and says which parts of it were reused and won.
Owner: Made here — name withheld by consent
the purpose its maker wrote when they asked for it
Microsoft 365 Copilot
Can reach
web search · your OneDrive and SharePoint
attested
Risk tier
Medium risk
attested
Governance stage
ODAT
History (3)
- Built from its maker's description · 2026-08-10
- Moved from Draft to In review · 2026-08-13
- Moved from In review to ODAT · 2026-08-18
Not shown, and why: How often this agent is used, and by how many people. Microsoft exposes no per-agent usage signal, and Aristo will not model one. The package download. The illustrative organisation hands out nothing — in your own tenant this is a one-click sealed link.
Month-end checklist
Run state not exposed
Built here
Microsoft 365 Copilot agent
Walks the finance close in order and says which step is waiting on whom.
Owner: Made here — name withheld by consent
the purpose its maker wrote when they asked for it
Microsoft 365 Copilot
Can reach
your OneDrive and SharePoint
attested
Risk tier
Low risk
attested
Governance stage
Approved
History (2)
- Built from its maker's description, on the deterministic template · 2026-07-26
- Moved from In review to Approved · 2026-07-31
Not shown, and why: How often this agent is used, and by how many people. Microsoft exposes no per-agent usage signal, and Aristo will not model one. The package download. The illustrative organisation hands out nothing — in your own tenant this is a one-click sealed link.
Onboarding buddy
Run state not exposed
Built here
Microsoft 365 Copilot agent
Answers a new starter the way their own team would, from the documents that team actually keeps.
Owner: No maker recorded
the purpose its maker wrote when they asked for it
Microsoft 365 Copilot
Can reach
your OneDrive and SharePoint
attested
Risk tier
Low risk
attested
Governance stage
Draft
History (1)
- Built from its maker's description · 2026-07-02
Not shown, and why: How often this agent is used, and by how many people. Microsoft exposes no per-agent usage signal, and Aristo will not model one. The package download. The illustrative organisation hands out nothing — in your own tenant this is a one-click sealed link.
In your tenant registry
Agents Microsoft reports for your tenant. Aristo surfaces them; your IT governs what actually runs.
Contoso Service Desk
Run state not exposed
Found in your tenant
Agent in your tenant registry — works when asked
Answers a question from the knowledge base and drafts a ticket update for a person to send.
Owner: No maker recorded
what Microsoft's own registry reports it does
Sees
The service-desk knowledge base and the open-ticket queue.
Not shown, and why: Whether your IT has switched this agent on. There is no app-only read of per-agent enablement, so Aristo never claims to have verified it.
Supplier onboarding
Run state not exposed
Found in your tenant
Agent in your tenant registry — works when asked
Checks a new supplier pack against the register and lists what is missing.
Owner: No maker recorded
what Microsoft's own registry reports it does
Sees
The procurement SharePoint library and the supplier register.
Not shown, and why: Whether your IT has switched this agent on. There is no app-only read of per-agent enablement, so Aristo never claims to have verified it.
Nightly invoice sweep
Run state not exposed
Found in your tenant
Agent in your tenant registry — works on its own
Files matched invoices and raises the unmatched ones to a person each morning.
Owner: No maker recorded
what Microsoft's own registry reports it does
Sees
The finance mailbox and the invoice folder.
Not shown, and why: Whether your IT has switched this agent on. There is no app-only read of per-agent enablement, so Aristo never claims to have verified it.
Found in your Copilot catalogue
Read from Microsoft's Copilot agent catalogue for your tenant.
Contoso Travel Desk
Run state not exposed
Found in your tenant
Copilot agent, published by Contoso IT
No purpose was recorded when this agent was made. Aristo shows the gap rather than writing a sentence on its maker's behalf.
Owner: No maker recorded
nothing was written to quote
Usage
not yet exposed by Microsoft
Benefits Answers
Run state not exposed
Found in your tenant
Copilot agent, published by Contoso People Team
No purpose was recorded when this agent was made. Aristo shows the gap rather than writing a sentence on its maker's behalf.
Owner: No maker recorded
nothing was written to quote
Usage
not yet exposed by Microsoft
Warehouse Stock Check
Run state not exposed
Found in your tenant
Copilot agent, published by Contoso Operations
No purpose was recorded when this agent was made. Aristo shows the gap rather than writing a sentence on its maker's behalf.
Owner: No maker recorded
nothing was written to quote
Usage
not yet exposed by Microsoft
How this register is counted, and what is not measured
Every agent under Built here is a real package in the register represented by this view. The purpose on each card is the sentence its maker recorded — Aristo quotes it and never rewrites it. Security answers are the maker's own attestation, captured before the forge would build, and they are labelled attested rather than measured everywhere they appear.
A maker is named only where your organisation has turned on named recognition and that person has personally agreed to it. Both are required. The count of how many people have built something is a count of people, so it follows the same 5-person floor as every other group in this product.
- Microsoft exposes no per-agent usage and no per-agent value, so this page counts what exists and refuses to model what does not: there is no "runs this week" figure here, and no dollar attached to anybody's agent.
- Nothing here says an agent is running. There is no app-only read of per-agent enablement or run state in Microsoft Graph, so every card reads "run state not exposed" rather than "live".
- Sharing and retiring an agent have no endpoint in this build. They are named here as absences rather than rendered as controls that would do nothing.
- How many people have built something is a count of people, so it is withheld below 5 — the fact that agents were built is never withheld, only the size of the group that built them.
Source: Aristo agent register + skill shelf As of: 2026-08-29 Counted from your own tenant's records. The Copilot catalogue and tenant registry are read from Microsoft where consent allows, and stay dark where it does not.