An assistant that answers with company policy, not generalities


From a specific question to the policy applied1. The doubt: A buyer, a site manager or a third party facing a real case. 2. The question: In natural language, from the same place where they already work. 3. The answer: Citing the relevant internal document. 4. The diagnosis: What gets asked reveals where the policy is ambiguous.FROM A SPECIFIC QUESTION TO THE POLICY APPLIED1The doubtA buyer, a site manageror a third party with a real case.2The questionIn natural language, fromthe place where they already work.3The answerCiting the relevantinternal document.4The diagnosisWhat gets asked reveals wherethe policy is ambiguous.
Trained on the organization's own documents: code of conduct, anti-corruption policy, gift policy and any others the company chooses to upload.

Integrity policies fail at the moment of a specific doubt. A buyer who isn't sure whether they can accept an invitation from a supplier, a site manager who doesn't know what to do about an irregular request, or a third party unsure whether its situation amounts to a conflict of interest rarely opens a sixty-page manual: they ask a colleague or decide on their own judgment.

Verdana includes an AI assistant trained on the organization's own documents —code of conduct, code of ethics, anti-corruption policy, gifts and hospitality policy, procedure manuals and any others the company chooses to upload— that answers questions in natural language, citing the relevant internal source. It is available both to internal users and to third parties through their portal, in the same place where they already work.

Its secondary value is diagnostic. The questions people ask reveal where the policy is ambiguous, which areas concentrate doubts and which situations keep recurring, which is exactly the input an integrity program needs to be updated based on evidence rather than intuition.