Skip to main content
Use Collivo AI to draft, summarize, and analyze project work. Before you share an answer or use it to make a decision, check that it matches your sources and the task you asked it to complete.
AI-generated output can be incomplete, outdated, or incorrect, even when it sounds confident. Verify important claims before relying on them.

Check an answer against your sources

  1. Confirm the context. Check the project, working folder, and files selected for the chat. Make sure the agent has the material needed to answer your question. See Chat basics and AI permissions, context, and privacy.
  2. Open the source material. Compare key claims with the original file or an authoritative source. When an answer includes a citation or link, open it and check that it supports the claim.
  3. Verify the details. Check names, dates, figures, calculations, and quoted text. Confirm that the answer uses the current version of the material.
  4. Look for gaps. Check whether the answer leaves out exceptions, conflicting information, or part of your request. Ask the agent to identify assumptions and information it could not verify.
  5. Review the finished work. Open generated files and check their content before sharing or publishing them. For an action that changes project content, inspect the affected item as well as the chat response.
For decisions involving health, law, finance, security, or compliance, have an appropriately qualified person review the work. An AI response does not replace professional advice or your organization’s review process.

Correct an inaccurate response

Point out the specific error, name the authoritative source, and explain what you want checked again. For example:
The summary says the launch is on 12 June, but the Approved launch plan lists 19 June. Reread the timeline section, correct the summary, and identify any other dates that do not match. Do not edit the source file.
If the answer is missing context, add the relevant files and ask a narrower question. If the conversation contains outdated assumptions, start a new chat with the current sources. Review the revised answer against the source again. See Troubleshoot AI answers and actions if the problem continues.

Rate an AI response

Use the feedback controls on an AI reply:
  • Thumbs up — mark a helpful response.
  • Thumbs down — mark an unhelpful response.
Selecting the same rating again removes it. To correct the work, also send a follow-up message explaining what needs to change. Rating a response does not replace checking the result.

Share product feedback or report a problem

For product ideas and general feedback, open the help menu in Collivo and select Send feedback. Choose Feature idea, Love note, or Other. For an incorrect or inappropriate response that needs investigation, or a technical problem, contact support. Include what you expected, what happened, the approximate time, and the agent mode. Share only the details needed to explain the issue; leave out passwords, access tokens, and unrelated private or sensitive information.