Appearance
Quickstart
Five minutes. Connect your HRIS, see your headcount, install the Agent, see your AI spend. That's it — you're using Flowstate.
This is the happy path. If you hit something gnarly, the per-integration guides have the details.
1. Connect your HRIS
Headcount data flows in from your HR system, so do that first. Go to Settings → Data Integrations and pick your provider.
| If you use… | Pick |
|---|---|
| Workday, Rippling, BambooHR, ADP, Gusto, HiBob, SAP SuccessFactors | The matching tile under HR Systems |
| Anything else | Custom Integrations — write a small JS hook |
Authorise. Pick what to sync. Click Connect & Test. Initial sync usually finishes in a couple of minutes for orgs under 1,000 employees.
2. See your headcount
Go to Plan → Main → Team → People. You should see every employee, contractor and vacancy on one roster, grouped by team. Salaries, FTE allocations and start dates are all populated from the HRIS.
If something looks off — wrong currency, missing team, unmatched manager — fix the mapping in the integration settings rather than editing records by hand. The HRIS is the source of truth and the next sync will overwrite you.
3. Install the Flowstate Agent
The Agent is the daemon that captures AI traffic at the request line. In Telemetry mode it captures usage metadata; in Enterprise mode it can capture prompts and responses too. In both modes, traffic is attributed to the engineer who made it.
- Provision your organisation's cloud-proxy key at
https://app.flowstate.inc/settings/ai/cloud-proxy. The same key embeds in every device. - On a single machine, run the one-shot installer:bashPer-platform deep dives: Install on macOS · Windows · Linux.
sudo flowstate install --key="<your-org-key>" --user="<your-email>" - The Agent generates the per-machine CA, starts the daemon, and configures macOS PAC / Windows WinDivert / Linux systemd. No further config needed for personal use.
Roll it out to the team via MDM — the macOS install guide walks the Kandji + Custom App path, and the Windows guide covers Intune. See Enforcement for the policy side once the fleet is reporting.
4. Hook up provider invoices (optional but recommended)
Agent telemetry tells you who used what. Provider invoices tell you what it cost. Reconcile both and the numbers line up to the cent.
In Settings → Data Integrations add the AI providers you pay for: Claude, ChatGPT, Gemini, GitHub Copilot, Cursor. Each one needs a read-only admin API key.
5. See your AI spend
Go to Plan → Main → AI → Insights → Spend. You should see today's burn, this month's run-rate and per-person breakdowns. Open Attribution to see how each dollar maps to a project and cost centre.
That's the loop — payroll on the same ledger as AI spend, both attributed to the same projects.
Where to go next
- AI Governance — policy, alerts, agent enforcement.
- Workforce — projects, allocations, scenarios.
- Finance — CapEx, R&D Tax, variance.
- Domain model — how the entities relate.