How Rigyx works

What it actually is

A service business — a web studio, a marketing agency, a consultancy — lives or dies on follow-up. Proposals that never go out. Cold emails never written. Invoices chased a month late. The founder is heads-down on delivery, and the growth work quietly slips.

Rigyx fills that gap with a standing team of AI agents. Each agent owns a domain — marketing, sales, operations, project management — and works the way a competent hire would: it knows your business, it drafts real work, it waits for your sign-off on anything that leaves the building, and it gets better at your business over time.

The mental model: a team, not a tool

Most software is a set of buttons you press. Rigyx is closer to colleagues you delegate to. You describe an outcome — “find ten guest houses on the coast and write each one a cold email” — and the Marketing agent decides the steps, does the research, writes the copy in your voice, and hands you a batch to approve.

Because they are colleagues, the agents also talk to each other. When the Sales agent closes a deal it hands the new client to Operations. When a campaign finishes, Marketing hands the numbers to the Project Manager for a retro. You do not orchestrate this — it happens between them.

How work actually gets done

Every agent runs the same loop, whether you triggered the task or the system noticed something on its own:

1. 1
   Observe

   The agent picks up a task you gave it — or a signal it caught on its own, like a competitor changing their pricing page or an invoice slipping past due.
2. 2
   Gather context

   Before writing a word it loads your business context: who you are, your brand voice, your archetype playbook, your knowledge base, past results. Output is never generic.
3. 3
   Draft

   The agent produces the real artifact — the email batch, the proposal, the campaign plan, the invoice — not a vague suggestion.
4. 4
   Wait for approval

   Anything that leaves the building or cannot be undone — sending mail, launching an ad campaign, placing a call — pauses for your explicit yes. This gate lives in code, not just in a prompt.
5. 5
   Execute & log

   On approval the agent runs the task through its tools and records exactly what it did, so there is always a trail.
6. 6
   Learn

   Background services read the outcomes — opens, replies, conversions — extract what worked, and feed it back so the next draft is sharper.

Where you work from

The same agents and the same business brain sit behind three surfaces. Use whichever fits the moment.

The rules it never breaks

An AI that runs real business communication has to be trustworthy by construction, not by hope. Three rules are wired into the platform: