Skip to content

Agent-first, explained

METAVSHN is the agent-first BSS: the system your customers, contracts and invoices live in, built so the AI assistants your staff already use can do real work in it. Safely, and with a record of who approved what.

What an agent is

An agent is an AI assistant, such as ChatGPT, Copilot or Claude, that does more than chat. It can look things up and prepare work in your systems on someone's behalf.

Most telecom systems were built for people clicking through screens. An agent-first BSS is built so that an assistant can read the same data and follow the same rules as your staff, and hand every change back to a person for a yes.

One credit request, start to finish

  1. CustomerWrites in: last month's invoice charged a service that was cancelled.
  2. AgentFinds the customer, the contract, the cancellation date and the invoice line. Prepares a credit note for the difference.
  3. Your teamChecks the prepared credit and approves it. METAVSHN records who approved it and when.
  4. ResultThe credit appears on the next invoice. Nobody copied numbers between systems.

What agent-first means for a BSS

An assistant is only as useful as the system behind it. Three things have to be true.

One record per customer
Customers, contracts, products, services and invoices sit in one connected model, so the assistant gets the full picture instead of fragments.
Rules per product
Prices, taxes, discounts and provisioning steps are defined once, so whatever the assistant prepares follows the same rules as your team.
Approval and audit trail
Anything that changes data or reaches a customer waits for a named person, and every step is logged.

What it means for each role

The same shift, read from four desks.

CEO
The same team can handle more customers and more products. Launching an offer becomes a request, not a project.
CFO
Billing questions, credits and month-end checks are answered from one source of truth. Every change an assistant prepares is approved by a named person and logged.
CTO and CIO
You choose the assistant: ChatGPT, Copilot, Claude or your own. It works through METAVSHN's access checks. No new data silo, no vendor-locked chatbot.
Head of operations
Routine look-ups and plan changes are prepared by the assistant, and people check and approve them. Fewer steps copying data between systems.

What stays in your hands

Who approves
Changes and anything sent to a customer need a person's yes. The assistant prepares, your team decides.
What is logged
METAVSHN records each request and each approval, so you can see who did what and when.
Which assistant
You use the assistant your company has already approved. METAVSHN does not bring its own.

Questions we hear from leadership

No. Your people stay responsible for every decision. The assistant takes over the look-ups and preparation work, so the team spends its time on customers and exceptions.

That depends on the assistant your IT approves and how it is set up. METAVSHN only answers what the signed-in user is allowed to see, and it checks access on every request.

Nothing that changes data or reaches a customer happens without a person's approval. Product rules in METAVSHN also reject requests that do not fit, and every step is logged.

The METAVSHN CLI is the first way an assistant connects to METAVSHN, and it is in early access. Your team keeps working in the web platform as before. More connection options follow.
For technical readers: the METAVSHN CLI

See what agent-first looks like on your data

We run our own ISP on METAVSHN. Tell us how your team works today, and we show you where an assistant fits.