AI-first, honest pricing, and a tool surface you can actually reason about.
Make great customer support accessible to teams who aren't willing to choose between paying per seat, wrestling a configuration tree, or trusting a chatbot they can't audit. The tools the AI uses are the tools you use — nothing hidden, nothing locked away.
Support software got complicated because it priced complexity. We think the next wave is the opposite: fewer surfaces, clearer billing, an agent that does real work, and a vendor you can trust to ship the boring stuff — tenant isolation, audit logs, data export — without an asterisk.
Four commitments we don't bend on — they shape every design review and every line of code.
One tool surface for the UI and the AI. If it's hidden from the agent, it's hidden from you — and the other way around.
Enforced at the database with row-level security, not hoped for in application code. No cross-tenant access paths, ever.
Every mutating action — human, AI, or automation — is logged with actor, subject, and cause. You always know what changed and why.
Pay when the product delivers — a solved ticket. No seat tax, no surprise overage, no punishment for growing a team.
A small, senior team in Odense with a bias for shipping. We've built support tools, SaaS billing systems, and multi-tenant platforms before — we just hadn't built one we'd happily pay for. Now we're building it.
We run on AWS in the eu-central-1 region (Frankfurt). Your tickets, messages, attachments, and embeddings never leave the European Union. No US-region failover, no "emergency" replication outside the EU.
We're a Danish company, so GDPR isn't a checklist for us — it's the floor. Every tenant has hard isolation at the database level (row-level security in Postgres, not just app-level filters). You can export everything, delete everything, and you'll get a real DPA with a real signature, not a click-through.
We list every subprocessor we use (AWS, Stripe, the LLM provider for AI replies) on the DPA, with what data they see and why. If we add one, you're notified before they touch your data.
AI calls run through a provider with zero-retention enabled — your tickets are not used to train anyone's model. Enterprise plans can swap to a self-hosted model on request.