About

Every organisation has people whose day is mostly routing. The applicant asks whether the role is still open, the customer asks where the parcel is, a colleague asks what the policy says — and someone opens three systems, finds the answer, and types it back. Then does it again.

We started Wetel to take that work off their desks. Not with another chatbot that answers from a document uploaded last quarter, but with AI employees: each one hired for a defined role, following your procedures, working inside the systems you already use, and checking with a person before anything that costs you something.

What we believe

• Sell the work, not the talking. An AI employee is measured by what it got done — the candidate filed, the status found, the record updated — not by how fluent it sounded.
• Supervised, not autonomous. Small things it just does. Consequential things — money, promises, someone’s record, anything irreversible — go to a person. Every run is on the record.
• Your systems, your data. It works inside the tools you already run, with your permissions, and every organisation’s data is isolated at every layer.
• No lock-in to one AI model. Each role can run on the model that fits its cost, quality and data needs, with automatic fallback if a provider fails.
• Honest about tense. If we say something is live, it’s live, and you can go and check. Everything else is labelled as being built.

What’s live, and what isn’t

We’re careful about this distinction here more than anywhere else, because an About page that oversells is worse than a homepage that does.

Live in production today:

• AI employees in recruitment, logistics operations and higher education — screening applicants and filing them into an HR system, tracking shipments (including reading a tracking code from a photo or a voice note), and answering from an institution’s own documents with bookings and renewals.
• A visual workflow engine: each role’s procedure is a graph you can read and inspect run by run, not a prompt you hope holds.
• 170+ certified operations against real business systems, plus custom APIs and MCP connectors, with credentials that never enter the model’s context.
• Chat, voice with genuine mid-sentence interruption, voice notes, a real-time 3D avatar, web embedding and Telegram.
• Per-organisation knowledge and isolation, with a dedicated-database option.
• A supervising agent that delegates background work to worker agents.
• The same infrastructure running, unbranded, inside partner products in customer conversations, education and recruitment.

Being built now:

• Work that starts on a schedule or a trigger, not only when someone writes in.
• A supervision view for each AI employee: what it handled, escalated and missed.
• Approval routing across a team, and swipe-to-approve on mobile.
• Phone calls, running the same procedures.

Where we’re headed, not built yet:

• Ready-to-hire role templates.
• Memory that carries across conversations.
• AI employees in different departments sharing one picture of the business — teams where people and AI employees genuinely work side by side.

The roadmap and changelog track both sides of that line as it moves.

Two ways to work with us

Hire an AI employee. Start with one role on your own systems, set up with our team, measured on your own numbers — then add the next role on the same foundation.

Build on the platform. If you’re a software company adding AI to your own product, Wetel is the runtime, isolation and governance layer underneath — under your brand, for your customers. We don’t build inbox or contact-centre products, and we won’t compete with you for your customers.

Get in touch

Reach out directly: kai@wetel.dev. Or leave your details here if you’d rather we come to you — either way, tell us which work you’d hand over first.