AI Office and Hub
The AI Office where AI employees work under named human owners
Wetel gives each AI employee a role, a board and an owner. AI drafts, people decide.
What changes when AI employees have owners
Faster first answers
An AI employee answers from your own documents and drafts the next step while the person is busy.
Work that stays in your hands
Money, promises and records go to a named owner for approval before anything happens.
A record of every step
Each run is traced step by step, so you can see what the AI employee did and why.
// AI Office Pilot
See the work, not just the chat
The AI Office shows your AI employees, their tasks and who owns them.
Board view. Tasks are grouped by status and each one shows its AI employee and owner. Demo workspace, preview feature.
Office, level 1. Each AI employee has a desk and shows whether it is working or waiting. Demo workspace, preview feature.
Office, level 2. A larger floor with more roles. Demo workspace, preview feature on a test environment.
AI employees
Start from a role template
Nine of the 24 published role templates. Each is a starting point you configure for your own documents and rules.
Support desk
Answers customer questions from your knowledge base and drafts replies.
HR onboarding buddy
Guides new hires through policies, forms and first-week questions.
Student advisor
Answers course, calendar and campus questions for students.
Hotel guest assistant
Handles guest questions about stays, facilities and local tips.
IT helpdesk
Triages common IT requests and hands the hard ones to a person.
Sales enablement coach
Helps sales staff find product facts and prepare for calls.
Lead qualifier
Asks the first questions and passes qualified leads to your team.
Customer onboarding guide
Walks new customers through setup, step by step.
Knowledge officer
Keeps answers tied to your documents and flags gaps.
Live Templates are drafts you review. A person approves anything consequential. See all 24 role templates
// Actions
AI employees that do things, not just chat
An AI employee's workflow can reach your systems, in the same conversation. Every step that changes something outside can wait for an owner's approval, and every run is logged.
- 1 Customer asks: where is my order?
- 2 AI employee calls the order system
- 3 Replies with the live status
- 4 Refund request? Owner approves first
-
Call your systems
Custom Actions call your REST or GraphQL endpoints with a stored credential. Create or update records in connected systems.
Live -
Certified operations
Pick from a catalogue of certified operations (190 at the last docs update) for connected partner platforms.
Live -
MCP connectors
Connect an MCP server and let the workflow use its tools.
Live -
Approval before action
Put any step that changes something outside behind an owner approval. Runs can also start on a schedule.
Live
- Webhooks for live lookups: Live
- Generated files: voice-over, jingle, explainer video: Pilot
- PDF export: Live
The ready-made role templates answer from your documents. Actions are added per customer: we connect your CRM, order system or calendar and set the approval rules.
Control and governance
Small things it just does. Big things it asks.
Every AI employee has a named human owner. Drafts come first and people decide.
Owner approvals
A step can wait for a workspace owner to approve or reject it from an inbox. If nobody decides in time, it takes its expired path.
Audit log
Runs are traced step by step with a change history. Spend is capped and models can be limited per workspace or key.
Drafts only by default
Actions come from a reviewed catalogue. Anything consequential stops and hands over to a person.
Owner and member roles
Workspace owners manage approvals and hiring. Members work inside what the owner allows.
Approval routing across a team is in development. Approving from a phone is planned after it. See the autonomy ladder How we keep data apart, and what is still open
// Channels and integrations
It works where your people already are
One AI employee, several ways to reach it.
-
Web chat Live
Embed the web SDK on your site.
-
Telegram Live
Connect a Telegram bot to an AI employee.
-
WhatsApp Live
Through partner platforms that run the WhatsApp channel.
-
API and SDK Live
Build your own interface on the GraphQL API.
-
Voice calls Pilot
Browser voice calls run on staging, on request.
WhatsApp groups and phone-network calls are not available today.
// For developers
Build on the same API
The GraphQL API and the web SDK run the same agents and workflows you configure in the AI Office.
// Proof
Honest results only
We publish numbers when we have measured them.
Technical case studies
- A university library
Grounded Q&A, room booking and loan renewal through a white-labelled gateway.
- Insurance claims on WhatsApp Pilot
A pilot of a multi-party conversation state machine.
// Pricing
Pricing is in design
We have not published prices yet. The plan limits that exist today are listed in the docs. Talk to us about your use case and we will tell you what fits and what is still in development.
// Where things stand
What is live today, what is in pilot, what is planned
Every part of Wetel carries one of three labels. The labels are dated and kept in one list that this page reads.
- Live
- Workflow engine
- Knowledge bases
- Telegram
- GraphQL API and web chat SDK
- Owner approvals and member drafts
- Role templates
- Pilot
- AI Office
- Generated files: voice-over, music, explainer video
- Voice calls
- Smart group messaging
- Planned
- Long-term memory
- Native WhatsApp and Slack connectors
- Hub on your own domain
- Auto-approve under a threshold
- Self-serve billing and pricing
Frequently asked questions
What is Wetel?
Does AI replace my staff?
What can an AI employee do on its own?
Which channels are supported?
Is the AI Office available in production?
How do you keep our data separate?
Can developers build on Wetel?
How much does it cost?
What is live today and what is not?
Is Wetel secure?
Start building your AI Office
Tell us the first role you want to hire. We will show you how it works with your documents.
// Contact
Talk to us
Tell us which work you'd hand over first. Whether you're hiring your first AI employee, building AI into your own product, or standardising across an enterprise — we'll reply with a real answer, not a drip sequence.