Manifesto
What nobody says in the meeting is what decides.
Seven theses on why people keep quiet at work, and on what it takes to really listen. Each one comes with the source behind it. Dialoga was built on them, and every guide and blog article links back to them.
- 1
Silence is not agreement
When people keep quiet at work, it is rarely for lack of an opinion. It is because they have learned that speaking up costs more than staying silent. Research calls this organizational silence, and shows it grows out of shared beliefs about the risk of speaking, not out of indifference.
- 2
People who speak up need safety, not courage
Asking for courage shifts onto the person a risk that belongs to the environment. Teams where mistakes and disagreement are not punished learn more, and it is that safety, not individual bravery, that makes information flow upward.
- 3
Anonymity is a condition, not a feature
Without real anonymity, listening measures what people think is safe to say. That is why it is not an option anyone can switch on or off: no group with fewer than three answers appears in a report, and the lock lives in the database.
Sources
- 4
A form measures what you already knew to ask
A 1-to-5 scale only captures what whoever wrote the question imagined. Long questionnaires produce autopilot answers. An open conversation brings out the cause. Studies show that people share more when they believe no person on the other side is judging them, and give more informative answers in a conversation with AI than in a form.
- 5
Listening once a year is photographing the past
The annual survey arrives after the person who had something to say has already resigned. Listening that works is continuous: a short, frequent conversation about one topic at a time, with the cycle interview when you need to go deep.
- 6
Asking creates a promise
People who answer, point out a problem and see nothing change learn that answering is pointless. That is why the cycle closes on the topic: people see what changed because of what was said, without ever being identified.
- 7
A method is something you can check
A claim without proof is a poster on the wall. Every commitment of the method has a safeguard in the code, every public claim has dated evidence, and whatever doesn't exist yet shows up as "we don't have this yet".
Brazil's NR-1 as a consequence
Since May 26, 2026, Brazil's NR-1 regulation expressly includes psychosocial risk factors in occupational risk management, with worker consultation. The regulation did not invent the problem: it started requiring companies to listen to what was always there. Continuous listening done well already produces the evidence the regulation asks for.
The cost of not listening
Depression and anxiety cost 12 billion working days lost every year worldwide, according to the WHO and the ILO.
Sources
What we don't claim
- We don't promise to cut turnover by any percentage: we have no study to back that up.
- We don't say AI replaces the manager. It listens at scale; people decide and act.
- We don't make individual health diagnoses. The reading is always at the group level.
See the theses at work
Talk to Laura for a few minutes. It's the most direct way to find out whether what's written here holds up.