Switch between Monitors while reviewing a conversation | Community
Skip to main content
Submitted

Switch between Monitors while reviewing a conversation

Related products:Fin & AI
  • Aug 21, 2026 7:38 AM UTC
  • 7 replies
  • 40 views

Forum|alt.badge.img

Currently, the same conversation can appear in multiple Monitors. Sometimes this is expected because a conversation is relevant to several evaluation criteria, while in other cases it may appear in a Monitor but not actually meet its criteria.

When reviewing a conversation in one Monitor, I’d like to be able to switch to another Monitor and evaluate the same conversation against its criteria, without having to leave the conversation, open another Monitor, and search for the conversation again.

A “Switch Monitor” dropdown/button would make this much easier. It could show the Monitors where the conversation is available and allow us to immediately switch the evaluation context.

This would be especially useful for teams that use multiple Monitors to evaluate different aspects of the same conversations.

7 replies

Forum|alt.badge.img+3
  • Active User
  • Oct 27, 2026 2:56 PM UTC

Are you using different Monitors for separate QA and risk criteria? I'd be interested to understand which parts of reviewing flagged conversations take the most time for your team.


Forum|alt.badge.img
  • Author
  • Connector
  • Oct 27, 2026 4:02 PM UTC

Are you using different Monitors for separate QA and risk criteria? I'd be interested to understand which parts of reviewing flagged conversations take the most time for your team.

Yes, we use different Monitors for billing, presale, and technical support conversations, since we evaluate them using different criteria.

The problem is that Monitors don't always identify the conversation type correctly. This happens most often with billing and presale chats.

Sometimes I open a conversation, read through it, and realize it doesn't belong in that Monitor at all. So I've already spent time on it, and if I want to evaluate it using the right Monitor, I have to go back, open another one, and find the same conversation again (if it's even there).

It would be really helpful to simply switch to another Monitor while reviewing a conversation, even if that conversation wasn't originally flagged by it.


Forum|alt.badge.img+3
  • Active User
  • Oct 28, 2026 7:37 PM UTC

Thanks, that clarifies it! Having to leave a conversation, switch Monitors, and search for the same conversation again sounds like unnecessary work, especially when billing and presales conversations need different evaluation criteria.

 

I'm building ChatShield to flag potentially risky support conversations for human review, alongside existing QA workflows.

 

I'm curious: when a billing or presales conversation contains an unsupported promise, incorrect policy information, or a potential privacy issue, do your existing Monitors reliably catch those risks, or can they slip through your current review process?


Forum|alt.badge.img
  • Author
  • Connector
  • Oct 29, 2026 7:30 AM UTC

Thanks, that clarifies it! Having to leave a conversation, switch Monitors, and search for the same conversation again sounds like unnecessary work, especially when billing and presales conversations need different evaluation criteria.

 

I'm building ChatShield to flag potentially risky support conversations for human review, alongside existing QA workflows.

 

I'm curious: when a billing or presales conversation contains an unsupported promise, incorrect policy information, or a potential privacy issue, do your existing Monitors reliably catch those risks, or can they slip through your current review process?

We try to review as many conversations as possible, but we prioritize those with negative CSAT ratings and low CX Scores. So far, we haven't had any major issues with risks slipping through our review process.


Forum|alt.badge.img+3
  • Active User
  • Oct 29, 2026 8:01 AM UTC

That makes sense, and it's good to hear your review process is catching the major issues so far.

 

One thing I'm curious about: since you prioritize negative CSAT and low CX scores, how do you handle a conversation where the customer seems happy but the agent or AI made an incorrect policy promise or shared information they shouldn't have? Would that still be something your team would want to flag, or is it already covered by your current review criteria?

 

I'm asking because I'm exploring whether risk-specific monitoring can complement existing QA without creating more work for the team.


Forum|alt.badge.img
  • Author
  • Connector
  • Oct 29, 2026 8:14 AM UTC

That makes sense, and it's good to hear your review process is catching the major issues so far.

 

One thing I'm curious about: since you prioritize negative CSAT and low CX scores, how do you handle a conversation where the customer seems happy but the agent or AI made an incorrect policy promise or shared information they shouldn't have? Would that still be something your team would want to flag, or is it already covered by your current review criteria?

 

I'm asking because I'm exploring whether risk-specific monitoring can complement existing QA without creating more work for the team.

Monitors don't catch every issue perfectly, which is why we try to review most conversations ourselves. For billing and presales, we actually review every single chat manually, since there may be issues that Monitors miss, even when the customer seems happy.

We've also set up separate Monitors for different types of conversations: billing, presales, and technical support, each with its own evaluation criteria. This helps us review conversations across the entire support team and catch potential issues.


Forum|alt.badge.img+3
  • Active User
  • Oct 29, 2026 8:45 AM UTC

Thanks, that gives me a much clearer picture! It sounds like your challenge isn't a lack of QA processes, but the manual effort needed to make sure important issues don't slip through even when Monitors miss them.

 

That's actually the problem space I'm exploring with ChatShield. It flags potentially risky conversations involving both human agents and AI for human review, alongside existing QA workflows rather than replacing them.

 

Since you're already reviewing every billing and presales conversation manually, I'm curious: would having a separate layer that flags potentially risky conversations for closer review be useful alongside your existing Monitors, or would that duplicate what your current workflow already does?

 

I'd genuinely value your perspective as someone who's dealing with this problem in practice.