Back to home
Function · Customer Support
Artellis builds a tool that drafts the reply from your own docs in seconds; you check it, it sends.
The pain
Why this one is worth fixing
The same six questions answered from scratch all day. The answer is already in your policy. It just isn’t where the ticket is. Your best people spend their afternoons on refunds, resets and shipping updates instead of anything harder.
What Artellis builds
A reply-drafting layer that reads the ticket, looks up your policy and the account, and puts a specific answer in front of the agent. Agent checks, one click sends.
How it works
- 01Reads the ticket and the accountPulls the customer’s history, plan, past tickets and any relevant order or invoice.
- 02Drafts from your policy, not the internetGrounded in your own docs: quotes the policy line and links the source, so nothing is invented.
- 03Agent edits, sends and closesOne click sends, tags the ticket, and files the case with the resolution reason.
Example artefact
What the output actually looks like.
Ticket #4821 · Refund
Draft reply · pulled from Refund Policy v4
Hi Aisling,
You’re right. We billed you on the 3rd and the 17th. I have refunded the second charge (€49.00) just now; it will appear on your card in 3–5 working days.
The double charge came from a retry on our side after a failed authorisation. I have flagged the account so it does not repeat.
Anything else, hit reply. Support
You edit and send.
Where it plugs in
Zendesk / Intercom / FreshdeskNotion / ConfluenceYour Stripe / billingSlack for hand-offs
Guardrails
- Human-in-the-loop: nothing sends, signs or spends without you approving it.
- Your data stays yours, hosted in your tenant or a dedicated Artellis workspace, never used to train public models.
- Week-one build: something small and useful is live in the first week, before we scope anything bigger.
FAQ
How long until we see something working?
A first working version of the customer support tool inside week one. Refinement and rollout happen after you have seen it work against your real inputs.
What do you need from us to start?
About 45 minutes to walk through the current flow, sample inputs (emails, files, screenshots, whatever you actually work from) and one person on your side who owns the outcome.
What happens if the AI gets it wrong?
Every draft goes to a human before it leaves the building. When you correct it, the tool learns your standard so the next one lands closer to right.
Bring us this workflow
One workflow. One week. Something working.

