How do I refine an existing automated line item

Once you've built an automated line item, you don't need to start from scratch when you want to change it. Open it in the Line Item Builder, iterate with the AI on any part of the definition, and save the result as a new version — either as a draft you can compare against your current one, or as an active version that takes over scoring straight away.

Required Feature Flags

The following feature flags and permissions are required to use this feature:

Feature Flag	Technical Name	Description
SmartScore v2	`feature_smartscore_v2`	Enables building, testing and managing custom automated line items

Required Permissions:

Manage Smartscore (evaluagent-cx.smartscore-v2.manage) — to build, test and manage custom Automated Line Items

Getting Started

Go to Analytics > Line Item Builder and click Improve an existing line item.

The Improve an existing line item picker opens. It shows every GenAI-based automated line item on your contract that you can iterate on:

Draft and Active line items are listed. Archived line items are excluded.
Line items provided by evaluagent (out-of-the-box topics) are excluded — you can only refine line items your team has built.
The search box filters the list by name.

Pick the line item you want to refine. The Line Item Builder loads its current definition — name, description, theme, prompt, and scoring template — into the conversation so you can see what you're starting from.

Iterating with the AI

Once a line item is loaded, the refine bar sits above the conversation with the actions you'll use most:

Chip	What it does
Suggest improvements	Asks the AI to review the current definition and propose targeted edits — for example, tightening the prompt, adding a Pass example, or clarifying an N/A condition.
Apply these changes	Applies the AI's proposal to the working copy so you can test it.
Review all changes	Shows a side-by-side diff of what you're about to save against the source version.
Choose a different line item	Reopens the picker if you want to switch to another line item.

You can also type your own instructions into the message input at any point — for example, "add a Pass example for authenticated callers" or "the Fail explanation should mention refunds explicitly". The AI edits the definition in place and the changes flow through to the test panel.

Testing changes against real conversations

Use the test panel to see how the working version scores against real contacts before you save.

For each conversation in the panel:

View opens the full detail for that result — the AI outcome, the AI's reasoning, extraction cards that highlight and scroll to the matching utterance in the transcript, and prev/next paging so you can walk through the sample.
Agree / Disagree feed back into the same signals used across the platform, so your judgement on borderline cases informs future prompt tuning.

The test panel scores conversations the same way the live scorer does, including the same handling of system-generated messages — synthetic or bot-only contacts that used to be rejected as "no transcript" now score correctly.

Saving a new version

When you're happy with the working copy, open the confirm panel to save:

Save as Draft — the new version is stored as a Draft alongside your current Active version. Existing scorecards keep scoring against the Active version. Use this when you want to review the change with your team, or run more validation before you switch over.
Save & Activate — the new version becomes Active immediately. Scorecards using the line item pick it up on the next scoring pass. The previous Active version is retained in the version history — you can reactivate it later if you need to roll back.

Every version records who saved it, when, and a short commit message describing the change. Open the topic in Manage topics to see the full version history for the line item, including which version was Active at any point in time.