Integration
Microsoft 365
Outlook, Teams and SharePoint: mailbox, procedures, reports.
Fewer emails and less paperwork, more time for care.
Ethic designs AI agents and builds workflows that help care and welfare teams with managing the mailbox, intake requests and putting procedures into practice. We start with a pilot of 4 to 6 weeks. The care professional always decides on care questions, and the AI only gets the data it needs.
No obligation · 30 minutes · online, with someone who puts AI to work every day
Ethic is the AI studio of Zenjoy, the digital product studio from Leuven that has built and maintained software, platforms and integrations for more than 400 organisations since 2011.
At a glance
- Who it’s for
- Hospitals, primary care, practices and welfare services
- Where we start
- One task, such as the shared mailbox or intake requests
- Our limit
- No AI that diagnoses, triages or decides on treatment
Typical AI workflows in care
Four tasks that eat up care time today
Every team has a test case. We start with one task, measure in the pilot how much time it takes and how often your team has to correct something, and then decide together whether to continue.
-
Shared mailbox
Go to Shared mailboxTodayThe general mailbox fills up with questions from clients, families and referrers, often with client data in them. Someone reads everything, sorts it and forwards it.
With an AI workflowEvery email gets a label and lands with the right colleague: that stays internal and can be undone, so the AI may do it itself. For standard questions, a draft reply is ready. Your team does the sending.
-
Forms and requests
Go to Forms and requestsTodayIntake requests come in by email, phone and web form, each in a different shape. Retyping and calling back take a lot of time.
With an AI workflowFree text becomes an intake request with fixed fields. If something is missing, you see it straight away. The AI doesn’t set priority or urgency: a staff member assesses the request and decides what happens next.
-
Knowledge assistant
Go to Knowledge assistantTodayWhere was that procedure again? Dozens of documents on SharePoint, three versions of the same protocol.
With an AI workflowYour team asks a question and gets an answer from your own procedures, with sources. If there isn’t enough to go on, the assistant doesn’t answer and refers the question on. We test that in advance.
-
Reporting
Go to ReportingTodayThe annual report and the returns for the government or the funder take up a lot of time every year.
With an AI workflowThe figures come from your own systems. The AI writes a first commentary, with the source next to every number. You read it and approve it.
Integrations
You don’t need to change software
We build in the systems your team already works in, with no more access than needed. Your IT lead or IT partner has a say in rights and access.
Integration
Odoo
Invoices, requests and reporting if your organisation runs on Odoo.
Do you work entirely in Microsoft 365, and is it about one mailbox or a few documents? Then first check what your own package can already do.
We only touch your patient or client records with your supplier and your DPO at the table, and through an official integration.
For your DPO
Sensitive data calls for strict agreements
01
Health data
Health data is special category personal data and needs extra protection. That’s why we map the data flow first, and only then the workflow. The AI only gets what it needs, pseudonymised where possible.
02
Professional secrecy
Your care professionals are bound by professional secrecy, and a tool must not undermine that. Together with them and your DPO, we set out which data a workflow may see, and which it never may. That’s in black and white in the AI fact sheet.
03
AI Act
If your organisation uses AI, your staff need to know enough about it. Applications such as triage of emergency patients count as high-risk, and we don’t build those.
This is not legal advice. Your DPO or lawyer has the final say.
Ground rules
The care professional always decides about care.
Diagnoses, triage and treatment are always decided by a person, not by AI.
Human oversight
AI may handle reversible internal tasks, such as labelling emails. A colleague first reviews anything sent outside the organisation and any decisions about people.
Only the data needed
We use the smallest model that works and only the data needed. The AI runs on European hosting as standard. Where needed and possible, even locally.
A named owner
The AI fact sheet records what AI may do itself, when approval is needed and when the workflow stops. One named owner on your team can stop the workflow.
Who builds it
A new brand, built on fifteen years of craftsmanship
Ethic is the AI studio of Zenjoy, the digital product studio from Leuven. Since 2011, Zenjoy has designed, built and maintained software, platforms and integrations for more than 400 organisations. You can see it in how we build AI: with tests, documentation and maintenance, so it still works three years from now.
We have used AI in our own work for years. With Ethic, we now put that experience to work where it matters, including in care.
15 years
building and maintaining software
400+
organisations
4 years
using AI daily in our own work
You talk to the people who build
No middleman: the people who design and build your workflow sit at the table themselves.
Code, documentation and access in your name
Code, prompts, documentation and access are in your name, not ours.
We stay on after launch
If a system or a model changes, the same team adapts the workflow.
Getting started
Start small, with one question or one task
1 hour or 2 to 3 weeks
AI advice
A 1-hour Pitstop to sharpen your question, or a 2-to-3-week AI Scan: where AI helps in your organisation and where it doesn’t.
You get concrete advice or an action plan.
4 to 6 weeks
Pilot around one workflow
In your own environment. We measure the time per task and how often your team has to make corrections.
Then you decide: roll out, adjust or stop.
Per group
AI literacy training
Are colleagues already pasting client data into a chat tool? Training for your whole team or a core group makes clear what is and isn’t allowed.
A first step towards the AI literacy the AI Act requires.
Proposal
After a first conversation, you get a proposal with a clear scope and a fixed price.
After the pilot
After the pilot, everything is in your name: code, prompts, documentation and access. Maintenance runs through a fixed monthly arrangement: updates, monitoring and small changes. New workflows are planned separately.
Frequently asked questions
Questions care organisations ask us
You can get a data processing agreement and a security overview on request.
Question not listed? Just ask us.
Is this allowed under professional secrecy?
That depends on what data the workflow sees and who has access. A knowledge assistant on your procedures often needs no client data. A general mailbox does: it contains names and care questions. Your care professionals and your DPO help decide which data the AI may then see. Whatever isn’t needed stays out.
Where is our data, and does the AI train on it?
As standard, the model runs in the EU. If that isn’t possible, the AI fact sheet says so and you decide with us. The AI fact sheet states which model runs, with which provider and in which country. The model is never trained on your data. It only reads the sources we agree on together.
What if the AI gets it wrong? Who checks?
A member of your team. The AI prepares, a person approves. We keep track of what the AI suggests and what your team changes. In the pilot we measure how often that happens, so you decide on figures, not on gut feeling.
How much time does this ask of our team?
Less than you fear, but not nothing. Before we start, we talk to the people who do the task today, because they know where it pinches. During the pilot we ask for a fixed contact person who reviews the AI’s suggestions and gives feedback. Your DPO and IT join at the moments that concern data and access.
What if our software supplier changes something?
As part of maintenance we monitor the integrations, so we spot a change as early as possible. We adapt the workflow without rebuilding everything, and if a better or smaller model becomes available, we switch. That’s covered by the fixed maintenance arrangement. If you stop, everything is in your name.
What does the AI Act ask of us?
If your organisation uses AI, it must take measures that support its staff’s AI literacy. Training is a good first step. For each workflow we record who is the provider and who is the deployer. That depends on what we build and under whose name it runs. We don’t build high-risk applications in care, such as triage of emergency patients.
Which task keeps your team furthest from care?
Tell us where it pinches. We’ll tell you honestly whether AI helps and what kind of help fits. Want to work through one question in depth straight away? Request a 1-hour Pitstop.
No obligation · 30 minutes · online, with someone who puts AI to work every day