Problem being solved
For text-based content, if content is being matched by a text-matching signal and routed to review, it can be hard to find what precisely actually triggered the routing. I feel like additional context would help a reviewer more quickly evaluate a report and take an action.
Proposed solution
When reviewing a job, we could highlight text-based matches in text content (similar to a Ctrl+F highlight in the browser) to make the report easier to scan.
Alternatives considered
Text-based highlighting would not work with signals that do things like sentiment analysis, image hashing, or otherwise analyze the content as a whole versus doing simpler string/regex matching—so maybe this isn't a perfect solution.
Instead, maybe having a field where signals can return context back to be surfaced on the job would be a better design? I.e. simple matching signals could list the strings that matched, but more complex integrations could return more free-form responses if supported by that integration.
Anything else
I'd love to understand what other review tools do and hear from adopters/moderators what would be the most helpful here in practice.
Screenshots
No response
Checklist
Problem being solved
For text-based content, if content is being matched by a text-matching signal and routed to review, it can be hard to find what precisely actually triggered the routing. I feel like additional context would help a reviewer more quickly evaluate a report and take an action.
Proposed solution
When reviewing a job, we could highlight text-based matches in text content (similar to a Ctrl+F highlight in the browser) to make the report easier to scan.
Alternatives considered
Text-based highlighting would not work with signals that do things like sentiment analysis, image hashing, or otherwise analyze the content as a whole versus doing simpler string/regex matching—so maybe this isn't a perfect solution.
Instead, maybe having a field where signals can return context back to be surfaced on the job would be a better design? I.e. simple matching signals could list the strings that matched, but more complex integrations could return more free-form responses if supported by that integration.
Anything else
I'd love to understand what other review tools do and hear from adopters/moderators what would be the most helpful here in practice.
Screenshots
No response
Checklist