AI in Flowstate
Flowstate shows what your company spends on AI, who uses it, and what it’s used for — alongside the people, teams and projects you already plan in Flowstate. Start with the bills, then add the Cloud Proxy to see which work the money went on.
The AI modules
Your Flowstate contact switches these on.
| Module | What it gives you | Needs |
|---|---|---|
| AI spend | Every AI provider’s bill in one place, by person, team and provider. The Agent insights Dashboard, Breakdown and Adoption tabs. | Your AI providers connected |
| AI attribution | Each AI conversation on your company Macs as a session, tied to a person and a project. The Agent sessions tab, and spend by project, initiative and task. | AI spend, and the Cloud Proxy on your Macs |
| Agents & copilots | The AI tools and agents your organisation uses, My AI for everyone, and monthly AI caps. | Budgets, for AI caps |
When a module is off, its screens say ”… is not enabled”.
Every AI screen
Insights → Agent insights
| Tab | What it’s for | Guide |
|---|---|---|
| Dashboard | AI spend for the period: what providers billed, how much of it Flowstate can tie to sessions, and which projects, initiatives and tasks it went on. | Dashboard |
| Breakdown | AI spend split by provider, model, person, team, project, initiative or objective. | Breakdown |
| Adoption | How widely and how deeply people use AI, by function and by person. | Adoption |
| Agent sessions | Every captured AI session: who had it, with which tool, for which project, and what it cost. | Agent sessions |
| Alerts | Alerts for spend spikes and dips, sensitive data sent to AI, unapproved AI use and new model releases. | Alerts |
Resourcing → Agents & copilots
| Tab | What it’s for | Guide |
|---|---|---|
| Agents | The AI agents you run, each with an owner, planned alongside your people. | Agents |
| Copilots | The AI tools your organisation actually uses. | Copilots |
| Models | The AI models agents can use, with their prices. | Models |
| Rules | Your AI usage rules, each written as one plain sentence. | Rules |
How agent and copilot costs are counted: The AI cost model.
My AI
Everyone’s own page, listed as My AI agents in the sidebar.
| Tab | What it’s for | Guide |
|---|---|---|
| Overview | Your own AI use at a glance. | My AI |
| My sessions | Your AI sessions. Pick the project a session was for. | My sessions |
| Copilots I use | The AI tools Flowstate has seen you use. | Copilots I use |
| Request centre | Ask for access to an AI tool. | Request centre |
Finance → Budgets → AI caps
A monthly AI spending cap for the organisation and for each team. See AI caps.
Settings → AI
| Page | What it’s for | Guide |
|---|---|---|
| Overview | Links to the AI settings pages. | — |
| Agent Policy | Who receives AI policy alerts, and the Flowstate agent’s settings. | Agent policy |
| Cloud Proxy | The profile your IT team deploys to Macs, and your private AI endpoints. | Cloud Proxy |
| AI Service Accounts | An identity and key for each of your own services that call AI providers. | AI service accounts |
AI provider billing connections are under Settings → Integrations. See Connect AI providers.
Insights → Engineering insights → AI impact
How much of what your engineers ship is AI-assisted, and what AI costs per merged pull request. See AI impact.
How it fits together
flowchart LR
B["Provider billing<br>Anthropic · OpenAI · Cursor · …"]
P["Cloud Proxy<br>on company Macs"]
S["AI service accounts<br>your own services"]
RS["Provider spend"]
SE["Sessions"]
B --> RS
P --> SE
S --> SE
RS --> D["Dashboard · Breakdown<br>Adoption · AI caps · Alerts"]
SE --> A["Agent sessions · My AI<br>spend by project and task"]
SE --> AI["AI impact"]
SE --> AG["Agents' actual spend"]- Provider bills say what you spent, and on whom where the provider reports it.
- Sessions say what it was spent on: the Cloud Proxy captures conversations on company Macs, and your own services report through AI service accounts.
- The Dashboard puts the two side by side, so you can see how much of the bill is explained by sessions.
- Agents you plan in Resourcing are compared with what their AI service accounts actually spend.
More detail: How AI data arrives in Flowstate.
Set up
- Track AI adoption — see who uses AI, how deeply, and which paid seats nobody opens.
- Track AI in your workflows — tie AI sessions to projects and kinds of work, and plan the agents you run.
- Attribute AI spend — connect your providers and see the bills. Splitting spend by project needs the Cloud Proxy.
What you can see and change depends on your role — ask your Flowstate admin.