Skip to content
Ethic
NL EN

for reception, admin offices and customer service

Your shared mailbox sorted before the day begins.

AI reads every new email in info@, labels it, routes it to the right colleague and drafts a reply. Your team checks it and sends it themselves, in Outlook or Gmail.

No obligation · 30 minutes · online, with someone who builds, tests and fine-tunes it

Ethic is the AI studio of Zenjoy from Leuven. Since 2011 that team has built and maintained software, platforms and integrations for more than 400 organisations.

See Zenjoy’s work

A man sorts emails into a rack of pigeonholes in the morning; the emails flow in along a teal ribbon.

At a glance

Works with
Outlook and Microsoft 365, Gmail and Google Workspace
Control
AI sorts by itself. A colleague sends every reply.
Start
Pilot of 4 to 6 weeks on one mailbox

Demo

Demo with fictional data

What a mailbox workflow looks like

Three new emails in the shared mailbox of a made-up non-profit. On the left what comes in, on the right who does what.

Shared mailbox · info@nonprofit.example

3 new emails

  • Comes in

    Sofie

    Routine question

    Certificate for the health insurance fund

    “Could I get a certificate for my son’s summer camp?”

    AI does this itself · reversible

    Label: Certificate

    → to the Admin team

    Registration looked up and ready

    A colleague decides · goes outside

    Draft reply with certificate is ready. A colleague reads it over.

  • Comes in

    Jonas

    Info missing

    Any places left?

    “Can my daughter still join next week?”

    AI does this itself · reversible

    No label

    → to the "To review" folder

    Reason: activity not mentioned. AI doesn’t guess.

    A colleague decides · goes outside

    A colleague picks the label and asks Jonas which activity he means.

  • Comes in

    Mira

    Wrong label

    Paid twice?

    “I think I paid the registration fee twice.”

    AI does this itself · reversible

    Registration → Payment

    ↺ Reverted: a colleague moved the email to Accounts.

    That correction is logged.

    A colleague decides · goes outside

    Accounts checks the payment and adjusts the draft reply.

Reversible and internal: AI does it itself. Going outside: a colleague first.
We agree on that in the AI fact sheet

Sound familiar?

Every morning starts with clearing the mailbox

Reception opens info@ and first reads everything that came in since yesterday. Only then does the actual work begin.

  • The start of the day

    Today

    Someone reads every email from top to bottom to see what it’s about.

    With this AI workflow

    Every email already has a label when you log in: registration, invoice, appointment, complaint or question.

  • Forwarding

    Today

    Emails are forwarded to a colleague one by one, asking "is this for you?".

    With this AI workflow

    The email is already with the right colleague. Doubtful cases are gathered in one "To review" folder.

  • The same questions

    Today

    The answer about opening hours, certificates or a payment is typed out again every week.

    With this AI workflow

    A draft reply is ready, based on your own standard answers. Your colleague adjusts it and sends it.

  • When someone is away

    Today

    Emails sit there and nobody really knows what is still open.

    With this AI workflow

    The labels show straight away what still needs following up, even when the usual colleague is away.

AI fact sheet

What this workflow may read is on one sheet

Every workflow we build comes with an AI fact sheet. Your DPO and your IT lead can see at a glance what the AI does, what it sees and who can switch it off.

See how an AI fact sheet works

Example

Sorting assistant for info@

Version 1.0 · October 2026

  • What it does

    Reads new emails in info@, picks one of 8 labels, routes the email to the right colleague and drafts a reply for routine questions. Sends nothing itself.

  • Which data

    Sender, subject and text of new emails in info@. No attachments, no personal mailboxes, no calendar.

  • Model and location

    A language model that runs in the EU. Agreement with the provider: your emails are not used to train models.

  • Who checks

    The AI labels and routes by itself. A colleague can move any email back. A colleague reads every reply and sends it themselves. The coordinator checks 20 random labels every month.

  • What it doesn’t do

    • Sending replies itself to clients, patients, parents or customers
    • Deciding on a request, a complaint or a case
    • Reading colleagues’ personal mailboxes
    • Reading attachments without a separate agreement
  • What we log

    Per email: time, label, whether a colleague moved the email and whether the reply was changed. The text of the email stays in your mailbox. Kept for 90 days.

  • Owner and stopping

    Owner: the reception coordinator. You can stop at any time: your IT lead or we withdraw access to the mailbox. Your labels and folders stay in place.

From conversation to pilot

One mailbox first, then you decide with figures

What we need from your team

  • One colleague who knows the mailbox: agree on labels, select example emails and report incorrect labels or replies.
  • Your IT lead grants access.
  1. 01

    Discuss your workflow

    30 minutes, no obligation. You tell us which mailbox and which emails it’s about. We tell you whether rules are enough or AI makes sense.

  2. 02

    Pilot of 4 to 6 weeks

    We first test with past emails under agreements approved by your DPO, then in your own mailbox. We measure daily sorting time and how often your team needs to make changes.

  3. 03

    You decide

    Roll out, adjust or stop. Based on the figures from the pilot, not gut feeling.

  4. 04

    After the pilot

    If Outlook or Gmail changes, we adapt the workflow. Updates, monitoring and small changes are covered by a monthly agreement. Code, prompts and access are in your name. Read what happens after the pilot

What you get after the conversation

After a first conversation you get a proposal with a clear scope and a fixed price.

What’s in the proposal

  • which mailbox and which types of email are included
  • which data the model sees, agreed with your DPO
  • what we measure during the pilot and how much time it takes from your team
Discuss your workflow

Frequently asked questions

What you want to know before you start

Question not listed? Just ask us.

A data processing agreement and security overview are available on request.

Do we actually need AI for this?

Not always. If you get few emails, mostly from the same senders or with a fixed subject, rules in Outlook or filters in Gmail are enough. If you mainly want to reply faster and sorting works fine, first see what your software can already do: Microsoft Copilot Chat in Outlook summarises emails and suggests a reply, in a shared mailbox too (Microsoft, checked on 1 October 2026).

An AI workflow pays off with lots of emails, several colleagues and sensitive data from clients, parents or patients. In a 1-hour Pitstop we work out together what suits you.

What if an email ends up with the wrong colleague?

Then your colleague moves the email or changes the label, just like now. The workflow logs that correction. At the review we see which labels often go wrong and adjust the list or the instructions. If the AI is unsure, it picks no label and the email goes to "To review", with the reason attached.

Where are our emails processed, and does the model see everything?

Your mailbox stays where it is now, in Microsoft 365 or Google Workspace. The model only gets the sender, subject and text of new emails in the agreed mailbox. As standard, the model runs in the EU. Which provider and which country is in the AI fact sheet. We only work with a provider that doesn’t use your emails to train models, and we put that in writing.

How much time does this take from our team?

Less than you’d think. The introductory call takes 30 minutes. During the pilot we ask for one colleague who knows the mailbox. First we go through example emails and labels together, then that colleague gives short feedback when a label or reply isn’t right. Your IT lead and DPO approve the access and the agreements beforehand.

What if Microsoft or Google changes something in Outlook or Gmail?

That happens regularly. Every workflow has checks, so we notice when something goes wrong, and the fix is part of the maintenance. If the AI built into your software becomes good enough for your mailbox, we’ll tell you. Then you can stop and take everything with you.

What does the AI Act mean for this workflow?

Your organisation uses the workflow and is therefore usually the deployer. Who the provider is depends on whose name the workflow runs under. We record both roles in advance in the AI fact sheet. Since February 2025 the AI Act has required your organisation to work on the AI literacy of people who work with AI (Article 4). That is why your team gets an explanation and a short training during the pilot. The workflow doesn’t decide about people and doesn’t talk to the sender itself: a colleague reads every reply and sends it. This is not legal advice. Have it checked by your DPO or lawyer.

Does the AI also read attachments, such as a doctor’s note or a report?

Not by default. The label comes from the subject and the text of the email. Reading attachments is only possible with a separate agreement with your DPO, and then as limited as possible: for example only recognising whether a certificate is attached, without storing its contents.

How much time goes into your shared mailbox every week?

Tell us which mailbox and which emails it’s about. We’ll tell you honestly whether rules are enough or an AI workflow makes sense, even if that means you don’t need us.

No obligation · 30 minutes · online, with someone who builds, tests and fine-tunes it