Accountable AI decisions.
When AI helps make a decision, people need to know where each fact came from, which rules applied and who approved what. Decidient records all of it, for every recommendation.
How we keep decisions accountable
- Rules set by peopleYour organisation decides which sources may be used, which AI models, what each step may spend and who must approve. These rules are written down and applied to every step.
- Every step checkedBefore and after each step, software checks the work against those rules. A step that breaks a rule is stopped and recorded.
- Every fact tracedEach fact keeps the exact words, the page and the date it came from. Code confirms the words are there. Missing information is recorded as unknown, never as no.
- Disagreements shownWhen sources disagree, both are kept. The person sees both when it matters. Work stops only when the disagreement would change the top recommendation.
- People approve what mattersA new tool goes live only after the people your rules name have approved it.
- A signed recordEach recommendation comes with a record of the sources, rules, steps, AI models and approvals behind it. Decidient signs the record and publishes the key to check it, so anyone can confirm it has not been changed.
Privacy and data
Your data stays yours
Each organisation sees only its own orders, tools and records.
Personal information protected
Personal information is sent only to AI models hosted in Canada or the United States, and only when the work needs it.
No copied wording
Tools explain official information in their own words and link to the source.
No personal traits asked
Tools ask about needs, such as step-free access, never about who a person is.
Built on established methods
Our approach draws on decision analysis, on how lawyers test document reviews, and on recognised frameworks for managing AI risk, such as the NIST AI Risk Management Framework and ISO/IEC 42001. We use them to guide our design. They are not certifications.
Contact us