Customer support agents
Customer support automation, in your voice.
The questions that repeat, on WhatsApp, email and chat, answered in your voice the moment they land, around the clock. The hard ones go to a real person on your team, by name.
What it handles
The questions that repeat
Not the delicate calls. The same questions, every day, on every channel, answered the way you would.
The same questions, answered
Order status, opening hours, what is in stock, how to book. The twenty questions that arrive every day, answered in your voice the moment they land, on whichever channel the customer used.
Open when you're closed
The message that lands at midnight is answered at midnight, not at nine the next morning. A reply in a minute keeps the sale; one the next day loses it.
A person, by name
The moment a question has no documented answer, or a customer asks for a human, the agent says so and passes the thread to a named person on your team. It never traps anyone in a loop.
How it stays yours
An agent that sounds like you.
The voice is written and approved before the build, and it ships to you as prompts you can edit. It answers only what your business has documented, and it tells customers it is an agent from the first message.
How it greets, what it will and will not say, where it hands over: written into the spec you approve, and yours to change as plain text.
It answers from your help docs and records, not the open internet, so it does not invent policy. No documented answer means a handover, not a guess.
It says it is an agent in its first message and never pretends to be a person, and it is still unmistakably your brand.
Not listed?
Looking for something else?
These are the shapes support work usually takes, not a menu. If what your team needs is something else, tell us. Twenty minutes, no pitch.
FAQ
About support agents
What people ask before they let an agent answer their customers.
Yes, by design, and it is the first thing we build. Before anyone opens an editor we write how the agent greets, what it will and will not say, and where it hands over, and you approve it. That voice ships to you as plain text prompts you can read and change. A generic corporate persona is the first thing we remove, never the thing you are left with.
The ones your customers already use. WhatsApp on the WhatsApp Business API, email, web chat, and the social inboxes where questions actually arrive. The agent answers on the channel the customer chose, in the same voice across all of them, and the specification names the exact channels once we have seen your accounts.
It answers only from what your business has documented, your help pages, policies and records, so it does not invent an answer. When a question has no documented answer, or the customer asks for a human, it stops, says a person is taking over, and passes the thread to a named member of your team. It is built to hand over rather than guess.
No, and it is built not to. The agent takes the repeat volume, the same questions answered a hundred times a week, so your people spend their time on the conversations that actually need a person: the upset customer, the unusual case, the judgment call. You decide where that line sits before anything ships.
A chatbot follows a fixed script and breaks the moment a customer phrases things their own way. An agent understands the question, answers from your documented knowledge, acts where it is allowed to (looking up an order, checking availability, booking a slot), and escalates to a person when it should. It holds a real conversation rather than walking a decision tree.
The agent runs in your own accounts, on a server you choose, and only your people have access; we hold no copy, and our access comes off at handover. For an EU or Swiss business it is built to the GDPR and the FADP, with personal data minimised and kept in a region you choose. If you operate under another country's law, we build to that law instead, and the rules that apply to you are written into the spec first.
Almost always, and usually as a configuration question rather than a rebuild. The agent reaches your tools over MCP and n8n's connectors, so a helpdesk or CRM like Zendesk, Intercom, Freshdesk or HubSpot is a connection we name in your specification once we have seen the account, not a claim we make before.
Weeks for a first agent, and no. It runs on your own accounts and credentials and does not check in with us to keep working. There is no standing fee holding it alive. Training is included so your team can operate it, and if you want a change later you ask, and we price that change on its own.