How it works
AgentWorks Studio: hire an AI employee for your business. Built for your industry, works in your tools. Build it yourself in the Studio, have us build it, or build it with your team.
Step 01
Pick your first AI employee
Start from one built for your industry, then teach it your business: your services, your hours, your answers and the things it must never say.
Step 02
Test it before anyone sees it
Chat with it in the Studio. Ask it what your customers ask. Fix an answer, test again, until you would hand it the job.
Step 03
Put it to work
Choose how much it does on its own, connect it to your tools and let it take the job. Want it on a real phone line? That is a Front Desk plan.
It does what you allow, and nothing else
Every AI employee has an approval level you set and can change at any time.
Your rules decide
The decisions in a job are written as rules a person can read, in a fixed order. When a message comes in, the rules decide what it is and where it goes; the AI writes the words inside that decision.
It works from what you taught it
Your services, hours, policies and answers live in one library your AI employee reads before it acts. Each document carries a last-checked date, so it works from what is current. No other business can see yours.
Approval levels, task by task
Off, suggest, draft or autonomous, granted by you per task. Some work never goes past draft, like an outbound text or a deletion. Drafts wait in your approval queue, and it cannot raise its own level.
Everything on the record
Every action it takes is logged with what it did, where and when. You can pause it at any time, and a paused AI employee takes no new work until you say so.
The short version is on the facts page, and the controls on security.
The log is the scoreboard.
The same log the daily report is built from feeds these stats. Each is defined once, with its formula and sources; pick one to see how it is counted.
How many times the workspace's agents called a tool in the window, as the tool-call ledger recorded it.
- Formula
SUM(calls) over GET /agents/tools/usage; calls = COUNT(*) FROM agent_tool_calls WHERE tenant_id = $tid AND created_at > now() - <days> GROUP BY tool_name (top 400 tools)- Sources
- agent_tool_callsGET /api/v1/agents/tools/usage
- League analogue
- Touches
- Value here
- — · computed live inside your workspace; this page carries no customer numbers
- Window
- The board's date selector
agent_tool_calls (one row per call, with its outcome) — not vertex_usage.tool_calls, which fleet.tool_calls sums; the two can differ by calls made outside a model run.
Start with one job
Anyone can start building in the Studio. No subscription, no credit card required.
Build it yourself
Owners, office managers and web people who want to try it today.
Open the Studio, pick an AI employee for your industry and teach it your business. No subscription, no credit card required.
Start building in the StudioHave us build it
Busy owners who would rather not build anything.
Send us your ten most-asked customer questions. We build your AI employee and put it on your website.
Request a Quick BuildBuild it with your team
Teams of 5 or more, and developers building agents into their own systems.
Get AI employees working across the whole business, with your people running them. Accelerator for teams of 5 to 19; Blue Heron CTO for 20 or more.
Book a call about your team