Every internal audit team is getting the same demo these months. A tool that reads documents, summarises, drafts and promises to give hours back. Almost all of them look similar on screen and look far less similar when the work has to survive a review. These eight criteria help separate them before signing.
What should you require from the tool?
- Every conclusion with its source. Correct text is not enough. Ask to see one specific conclusion and follow the trail to the document, the row and the date it came from. The difference between an opinion and a finding is that the finding can be defended.
- The decision stays yours. Check what the tool does when sources disagree. If it drafts anyway, closing the conclusion on its own, the problem is not text quality, it is who carries the responsibility.
- Traceability of who did what. Who asked, what ran, who accepted and when. That record is what you show afterwards, and it has to exist without anybody assembling it by hand.
- The whole population, not a convenient sample. Ask how far the analysis reaches. Many tools work well with twenty documents and fall over with the full export, which is exactly where the real work is.
- Your methodology, not theirs. Your frameworks, templates and criteria have to go into the tool. If the work starts from an empty chat, the firm's judgement is lost on every engagement.
- Where the data sits and what is done with it. Where it is processed, whether it trains models and what happens when the contract ends. All three answers, in writing.
- Integration with what already exists. Your systems hold the records. A tool that requires moving everything into a new store adds a data governance problem instead of solving an audit one.
- An exit as clear as the entry. The file belongs to the organisation. Before signing, confirm the format it comes back in if the relationship ends and the deletion window the vendor accepts.
The difference between an opinion and a finding is that the finding can be defended.
How do you test it in the demo?
There is a quick test that orders the comparison. Ask for the demo to run on one of your cases, not theirs, and with one source contradicting another. What the tool does at that moment, whether it stops and flags it or drafts a clean conclusion, tells you more than any product sheet.
Why compare now?
The comparison is happening now because the cycle's plan brings engagements the team cannot cover, which is what the capability gap in the profession describes. A tool is judged by how much of that work it leaves ready to review, not by how much text it writes.