Know what energises and drains your team — every quarter, in 15 minutes
Each employee has one short chat or voice check-in with an AI interviewer. You get a clear six-part report, commit to at most three actions, and next quarter your team tells you whether they worked.
- For teams up to about 20 people
- Anonymous by cryptography
- Processed in the EU
- No cookies on this site
How often do you have enough time to finish your work?
Thanks. Think of one concrete moment this quarter that gave you energy. What happened?
When we finally fixed the planning board together. No more double work.
One conversation a quarter. One loop that closes itself.
No consultants, no 80-question surveys, no spreadsheets. Set it up yourself in minutes.
-
15 minutes · per person
Check in
Everyone gets one conversation in the Sainsei app, by chat or voice: about 10 minutes of validated survey questions, read word for word, then about 5 minutes of AI follow-up on what energises and drains them.
-
Week 3
Read the report
The same six-part report every quarter, so trends are easy to read. In a small team you get distributions and themes, honestly labelled as indicative — no fake statistics.
-
Week 3–4
Commit to at most 3 actions
Choose up to three actions, each with an owner and a deadline. Results and actions are shared with the whole team in the app.
-
Next quarter
Loop back
The next check-in starts by asking everyone whether last quarter's actions changed anything: better, same or worse. You see what worked — and what didn't.
The six-part report
Same structure every quarter. Readable in ten minutes, by a founder, not a data scientist.
- Headline — response rate, three scale scores, three key messages
- Report-back — how last quarter's actions landed
- Outcomes — engagement, energy and overall satisfaction
- Drivers — the resources and demands behind them
- Voices — top energisers, drainers and change requests, paraphrased
- Proposed actions — at most three, for you to accept or change
Anonymous by cryptography, not by policy
The login proves who may answer — never who did. Not the employer, not our staff, and not us under pressure can link an answer to a person.
- Blind-signed tokens. Your phone gets a one-time participation token that our server signs without ever seeing it (RSA blind signatures, RFC 9474, audited library).
- A separate response service. Answers go to a different service with its own database, stored under a random ID. It never sees an account, an email address or a device.
- No side channels. IP addresses stripped, dates kept per day only, answers released in batches with a random delay. Reminders go to people who haven't started — never to "people who haven't answered".
- No audio stored, no emotion recognition. Voice is transcribed and discarded during the session. We never analyse tone, emotion or voiceprint.
- Your own scores stay on your phone. Employees can see their personal results; they are never sent to the employer.
Built on the Job Demands–Resources model
JD-R is one of the most widely researched models in work psychology. Engagement and wellbeing come from the balance between what energises people at work (resources) and what drains them (demands). Sainsei measures both, the same way, every quarter.
Energisers · job resources
- Autonomy and say in how work gets done
- Support from colleagues and the manager
- Clear roles and goals
- Feedback and room to grow
- Psychological safety
Drains · job demands
- Workload and pace
- Emotional demands
- Clunky tools, processes and interruptions
- Unclear or conflicting expectations
- Work spilling into private life
Validated items, word for word
Survey questions are shown or read exactly as written, never paraphrased by the AI, so scores stay comparable over time.
Honest statistics for small teams
Statistical tests switch on only when there are enough respondents. Below that, you see distributions and themes, clearly labelled as indicative.
Neutral, concrete follow-up
The AI asks about one concrete moment, with at most two follow-up questions. No leading questions, no advice or therapy, no questions about health or named colleagues.
Designed for European rules from day one
Sainsei measures the organisation, never the individual. That principle shapes every design choice below.
EU-only processing
All app data is processed and stored in the EU. The AI runs on an EU endpoint — no US language-model endpoint.
GDPR: you control, we process
You are the controller, we are your processor: click-through DPA, published subprocessor list, and DPIA and legitimate-interest templates. Participation is always voluntary.
EU AI Act: no individual evaluation
Aggregate-only output, no per-person scores for the employer, and a contractual ban on individual decisions. No emotion recognition. The AI interviewer is disclosed upfront and visibly non-human.
Belgian psychosocial risk analysis
Belgian employers must analyse psychosocial risks. The quarterly report can feed into that analysis, and onboarding prompts you to involve your prevention advisor and, where it exists, the CPPT/CPBW committee.
This is a description of how Sainsei is designed, not legal advice. Our compliance documentation is reviewed by counsel.
Pay only for completed check-ins
One simple unit: one employee completing one quarterly check-in. Invites that are never answered are never billed.
- Text and voice check-ins are priced separately, because voice costs more to run.
- The price is reviewed every quarter. When our costs fall, your price falls too.
- Start with a free pilot and decide after your first report.
Your price is derived from what it actually costs to run your check-ins — AI, voice, hosting and any survey-scale licences — with one fixed rule on top that we show you in the app. No seats, no hidden tiers.
Start free pilotThis website uses no cookies, no analytics and no tracking scripts. We care about how your team feels, not about your clicks.
Privacy noticeQuestions people ask
Can my employer see my answers?
No. Your employer only sees results for groups of at least 5 people, and open-text themes for groups of at least 8. Quotes are paraphrased and stripped of names, roles, clients, dates and places. Your own scores are visible to you, on your phone only.
Can Sainsei see who said what?
No. The service that knows who you are only hands out blind-signed participation tokens; the service that stores answers never sees an account, email address, device or IP address. Nobody holds both halves of the link.
We're only 8 people. Does this work?
Yes — Sainsei is built for teams of up to about 20. With 5 or more respondents you get the full report; team splits appear only if every team meets the threshold. With fewer than 5 respondents company-wide the report is not released, and people are told this upfront.
Is the interviewer a person?
No. The interviewer is an AI, says so at the start, and has a clearly non-human avatar and voice. Nobody at your company or at Sainsei reads raw answers.
Do you record my voice?
No. Speech is transcribed and discarded during the session. No audio is stored, and we never analyse tone, emotion or voiceprint — only the words you choose to say count.
What if someone raises harassment or a safety issue?
The AI shows a pointer to confidential channels, such as your prevention advisor or an external service. It does not forward the issue to the employer.
Does it replace our psychosocial risk analysis?
No, but it can feed into it. Sainsei gives you a regular, structured picture of resources and demands; your prevention advisor stays in charge of the analysis itself.
Which languages are supported?
Dutch, French, English and German — for both the check-in and the report.
How do we start?
Create an account at app.sainsei.com, add your team and open your first quarterly cycle. Employees install the Sainsei app and get a sign-in link by email. The pilot is free.
Your first check-in can run this quarter
Set up in minutes. Fifteen minutes per person. Three actions, at most.