Odoo
Requests go into the Odoo app your team uses to follow them up, for example CRM.
for non-profits, federations, schools and care organisations
In intakes, registrations and grant applications, the most important information is often in free text. An AI workflow suggests a value for each field in Odoo, your CRM or your own system, shows the original sentence next to it and flags what is missing.
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.
At a glance
How it works
A request comes in through your website, Microsoft Forms or email.
AI suggests a value for each field, such as a preference, category or question, with the original sentence alongside.
The workflow checks required details and existing contacts. It does not fill in missing information itself.
The request goes into Odoo, your CRM or your own system as a draft. A colleague must approve it before it becomes final.
Your colleague checks and adjusts. Your team decides who gets a place or a grant and which email the applicant receives.
Always checked: the draft includes the source sentences and the status “To review”. If the AI is unsure, the field stays empty.
Sound familiar?
This is how a new volunteer’s application, an intake or a grant application goes, today and with this AI workflow.
“Preferably at weekends, something with children” sits in a text box and ends up nowhere.
The free text becomes a preference and a category, with the original sentence next to it.
Only after a week do you notice a document or detail is missing. Then the emailing starts.
What’s missing is flagged straight away, with a suggested email to ask for it. Your colleague sends it themselves.
The same person is in your system three times, each time with a different email address.
The workflow looks up existing contacts and suggests merging them. Your colleague decides.
Requests go into the Odoo app your team uses to follow them up, for example CRM.
Requests become contacts and leads, with existing contacts recognised.
Responses from Microsoft Forms or a shared mailbox in Outlook.
Working with a different system or custom software? Then we look together at what access is available.
AI fact sheet
Every workflow comes with an AI fact sheet: what goes into it and what to watch for. This is what it looks like for a volunteer form.

Example
Requests to Odoo
Version 1.0 · October 2026
What it does
Reads new submissions of the “Become a volunteer” form, puts them into Odoo as a draft with the status “To review” and flags what is missing. Sends no emails itself.
Which data
Only the answers on the form, plus name, email address and phone number in Odoo to find duplicates. No national register number, no health data, no attachments.
Model and location
A language model that runs in the EU. Your data is not used to train models; that is in the contract with the provider.
Who checks
The workflow fills in the fields itself. A colleague from the volunteer team checks every request, decides and sends every email. When in doubt, a field stays empty.
What it doesn’t do
What we log
Every suggestion with its source sentence, every change and who handled the request. Kept for 90 days, visible only to the owner.
Owner and stopping
Owner: the volunteer coordinator. You can stop at any time: we withdraw the access in Odoo and the form keeps working as before.
Want a blank AI fact sheet for your own AI tools? Request one.
How you start
30 minutes, no obligation. You show us which form gives your team the most work. We’ll tell you honestly whether AI helps.
We test with real requests from the past, in your own environment. We measure time and corrections.
Roll out, adjust or stop. Based on figures, not gut feeling.
If your form or CRM 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
After a first conversation you get a proposal with a clear scope and a fixed price.
What’s in the proposal
Which form and which channels are included · which data the model sees, agreed with your DPO · what we measure during the pilot
After a 4 to 6-week pilot
Frequently asked questions
A data processing agreement and security overview are available on request. Question not listed? Just ask us.
Not always. Does your form mostly ask for fixed details, such as a name, a date or a choice from a list? Then improve the form, with drop-down lists and required fields. Are the form and the CRM in the same software package, and does that package have its own AI features? Try those first. We check with you whether it’s enough for your review step and your data.
An AI workflow helps most when requests come in by website, email and paper and need to go into more than one system. Is there sensitive data in it, such as about health, minors or someone’s financial situation? Then the workflow only reads what is really needed, with the smallest model that works. We discuss that with your DPO beforehand.
After a 30-minute introductory call, we test for 4 to 6 weeks with real requests from the past, in your own environment. Then you have a working workflow around one form, figures on time and corrections, and our advice: roll out, adjust or stop. Want to know first whether it makes sense? Then start with a 1-hour Pitstop.
Only the fields we agree on together and that are in the AI fact sheet. In Odoo or your CRM the workflow gets its own user with only the permissions it needs. The model runs in the EU. We choose the model and hosting based on how sensitive your data is. Your data is never used to train models.
That can happen. That is why the workflow only puts a request into your system as a draft, with the status “To review”. It only becomes final once a colleague approves it. If the workflow is unsure, the field stays empty. In the pilot we measure how often your colleague has to correct things, and we agree in advance what error margin is acceptable. Everything is logged, so you can see afterwards what happened.
The workflow has checks and error alerts, so a change gets noticed quickly. Adapting it is part of the maintenance agreement. Zenjoy has been maintaining platforms and integrations for years after launch, with Salesforce, Odoo and Airtable among others.
This workflow reads requests and prepares them. It doesn’t decide about people, and that is stated in the AI fact sheet. If a system decided by itself who gets admitted to a course, or on behalf of a public authority who is entitled to a benefit or to care, it would count as a high-risk AI system (Annex III of the AI Act; those rules apply from 2 December 2027). We don’t build systems like that. The AI Act also asks your organisation to take measures to support the AI literacy of people who work with AI (Article 4). We agree together in advance who is the provider and who is the deployer. This is not legal advice: have it checked by your DPO or lawyer.
Often, yes. The workflow can read emails the same way as a web form. For paper you need a legible scan. In the pilot we test with real examples, including the messy ones. If a channel turns out to be too unreliable, we say so, and that channel stays manual.
Related
Integration
CRMSector
Non-profitsWorkflow
Shared mailboxBack
All agents and workflowsShow us an example in the introductory call, without personal data. We’ll tell you honestly whether an AI workflow helps, or whether a better form will do.
No obligation · 30 minutes · online, with someone who builds, tests and fine-tunes it