SAI in-house build

How we run our own business on agents

We use what we sell, every day.

Enquiries

Team alert#enquiries

HermesApp

New enquiry triaged: a question about a workflow.

Waiting for a personWaiting for a person

Draft reply

Drafted by
Claude
Sends
After a person approves

Illustrative sample data

  1. The problem

    We ask clients to trust agents, so our own work should show it.

  2. What we built

    Claude Code builds, Hermes coordinates, Claude triages enquiries and drafts replies.

  3. Where it stands

    Agents do the repeatable steps. A person makes the decisions.

The full write-up

The problem

We ask clients to trust agents with real work, so our own operations should show what that looks like day to day: agents doing the repeatable work and a person making the calls.

What we built

Build work happens in Claude Code. Hermes, the open-source agent runtime from Nous Research, coordinates the workflow around it, and it is the same runtime we deploy and configure for each client. Enquiries are triaged and replies drafted with Claude, with the providers involved listed on our security page.

Where it stands

The approach we sell is the one we use: agents handle the repeatable steps, and decisions stay with a person.

How it fits together

Parts: Hermes (Agent runtime, coordinates); Claude Code (Builds the work); Enquiry (Comes in by email or form); Claude (Triages and drafts replies); A person (Makes every decision). Connections: Hermes to Claude Code (coordinates); Enquiry to Claude (triage); Claude Code to A person; Claude to A person (drafts).

Want a build like this for your business?

Tell us the process you would hand over. We will say plainly whether it fits.

Book a call (opens in a new tab)

or email contact@strategicaiinnovations.com