How we keep the AI honest.
Symphiq proposes work and explains itself. The receipts below are the standing rules the AI runs under, stated plainly, so you can hold it to them. Symphiq proposes. You decide.
The receipts
Five standing rules. None of them is a setting you have to find. They are how the product behaves everywhere.
- It shows its sourcesAsk any recommendation where it came from and it shows you: the evidence, the reasoning, and the metric it is meant to move.
- It does not invent numbersIt is instructed never to fabricate a figure. When it cannot see something, it says so instead of guessing.
- It records honest verdictsCompleted work gets a verdict against the metric it was meant to move, and not met is a real outcome it keeps rather than hides.
- It proposes, you decideNothing activates itself. Coverage and capacity reads advise; they never block a person's decision.
- The assistant is read-onlyThe built-in assistant answers questions about your plan. It never creates, edits, or deletes anything.
Symphoniq, the read-only assistant
Ask it about your goals, objectives, recommendations, and tasks. It answers in plain language and points you to the right place in your plan.
- Read-only: it can never create, edit, or delete anything in your plan.
- Scoped to your shop only: no cross-shop access, no outside data mixed in.
- No fabrication: it never invents an ID or a metric. If it does not have the data, it says so.
Answers
No. It proposes and explains. A person accepts, partly accepts, or rejects, and only then does anything change.
It says so. The assistant is instructed to answer that it cannot see something rather than produce a guess.
No. Subprocessors may not train on your data, and no personally identifiable information goes to AI providers. The details live on the Security & data page.
Rules you can check beat promises you have to trust.
Get Instant Insights