How to View the Bulk Assessment History and Add a Comment to a Run?

The bulk assessment history lists every bulk assessment run of your instance: who started it and when, where the chats came from, how many were assessed, skipped by rules, not assessed by topic and how many ended with an error. The page of a run shows the result of every chat, and you can add a comment to the run for your team.

When It Will Be Useful

• A run has finished, and you need to see which chats were assessed, which were not and why.
• You need to find the assessment of a particular chat from a bulk run.
• You want to label a run for the team, for example "Calibration, part 1/3".

What's Important to Know

• When it is available. The history is available when the AI QA auto-assessment settings are enabled on the instance. Otherwise the page shows "History is available when the auto-assessment settings are on".
• Who sees it. Every user with the right to view quality assessments sees all runs of the instance.
• Chat statuses in a run:
  • Assessed — the assessment was created and the AI finished it;
  • In progress — the assessment is still running;
  • Error: reason — the assessment was not created, or the AI finished it with an error;
  • Skipped: reason — the chat was skipped by a skip rule;
  • Not assessed by topic — "Do not assess" applied to the chat's topic in this run.
• Skip reasons here and in the dialog card differ. The "AI QA skip reason" field of the dialog card is filled only by the automatic assessment when a chat closes. Why a bulk run skipped a chat is shown only in the run history.
• Who edits the comment. The author of the run, a user with full access, or a role with the "Edit run comments" right. A comment holds up to 500 characters; everyone with access to the history sees it together with the author and the time of the change.

Before You Start

• You have the right to view quality assessments.
• The AI QA auto-assessment settings are enabled on the instance.
• You have navigated to the Quality Assurance module through the menu Apps -> Quality Assurance.

Step-by-Step Instructions

1. Open the history

1. In the "Assessments" section of the Quality Assurance module, click "Bulk assessment history".
2. The runs table has the columns Date, Author, Source ("From the chats table", "File or IDs"), Status, Assessed, Skipped by rules, Not assessed by topic, Errors and Comment.
3. Filter the runs by status if needed.

2. Open a run

1. Click the run you need.
2. At the top are the run counters and the mark "Skip rules: on" or "Skip rules: off".
3. The chats table shows for every chat the Chat (the link opens the chat in the operator panel), the Checklist, the Status and the Assessment (a link to the created assessment).

3. Deal with skipped and failed chats

1. Match the reason with the list below.
2. To assess a skipped chat, start a new run without skip rules or assess the chat manually from its card.

4. Add a comment

1. On the run page, type the text into the comment field (up to 500 characters).
2. Save it.

5. If the run was stopped on timeout

If a run made no progress for a long time, it ends with an error, and the run page shows a hint. There is no automatic recovery:

1. Click "Copy IDs".
2. Click "Start from the chats table" and create a new run through "Bulk assessment" → "File or IDs" with the copied IDs.

How to Allow Editing Other Users' Comments

1. Open Settings -> Roles and the role you need.
2. In the quality assurance card of the role, find the "Bulk assessments" block.
3. Turn on "Edit run comments" and save the role.

The block appears when the AI QA auto-assessment settings are enabled on the instance.

How to Make Sure Everything Worked

• For a finished run, Assessed + Skipped by rules + Not assessed by topic + Errors equals the number of chats in the run.
• The saved comment is shown in the Comment column of the history.

Related Materials

• Run a bulk assessment from the chats table or a list of IDs
• View assessment details
• View assessments list