Skip to main content
Answered

How to prevent agents from being assigned conversations in languages they do not speak?

  • 10 June 2024
  • 3 replies
  • 38 views

Hello!

Our team uses multiple teams to organize our workload based on tasks and communication channels. Adding language-based separation would create an excessive number of additional teams, making it difficult to manage shifts efficiently.

We've found that:

  1. We cannot add "skills" to agents to designate the languages they should handle.
  2. Workflows do not support the "do not assign to" option.

Is there an alternative way to prevent agents from being assigned conversations in languages they do not speak, other than creating separate teams?

3 replies

Userlevel 5
Badge +4

@Tereza Zaloni This is a very interesting question, I don't have experience yet with this issue but would be interested to see how to solve this as well. 

 

My initial thought if there's no existing solutions out of the box, maybe we could understand if you have a number of agents that speak target languages and if you could use workflows and audience details to direct conversations of a specific language to those agents vs the alternative of trying to not assign something to an agent who doesn't speak that language. 

 

But I need to look at what's possible with languages and assignment as in my work with consulting on Intercom we haven't come across this use case yet. 

 

Happy to look into this more if the Intercom team doesn't have an out of the box solution, we may need to get creative but I've got to think they have run into this issue before and solved for it..

 

It seems the help center definitely has a lot of support for delivering the right language content, I'm now curious to dig into the support for language in the assignment process because I agree that dividing by department/team and then language teams would add a lot more complexity, you would probably need a duplicate team for each language you serve and then workflows would need to keep this in mind it would be a real challenge although might be doable if you had generalist support leaders for each language but it would depend on volume of requests etc as well. 

If you don't get a clearer answer from someone with experience solving this issue and want to dig into this together please feel free to reach out. 

 

Userlevel 5
Badge +4

One thing we did previously with a client on the topic of language was install some 3rd party translation apps in the Intercom interface to automatically translate the users messages and send responses in the target language. I'm wondering if the new Copilot beta might also be able to help with this as well now where it can understand the language and enable any agent to better handle the responses even in unfamiliar languages? That could be a very interesting use case.  Even the translation tools at the time were quite good this was before AI implementations like Fin and Copilot, it helped a lot. 

Hello @Nathan Sudds! 

Thank you for your thoughtful response and suggestions.

Our team primarily handles workload in Greek and English. While all agents support English and some are multilingual, there are a few agents who do not speak Greek.

Indeed, we can direct conversations based on language, but the volume of requests in languages other than Greek and English is not substantial enough to justify creating separate teams or implementing language-based assignments.

Additionally, all our Greek-speaking agents also cover English. Thus, creating workflows to assign agents to both language cases would require adding each agent separately, which could be challenging due to our team's size.

Most importantly, if we proceed with language-based assignments, it raises concerns about potentially conflicting with our existing assignment rules based on task teams.

Given these considerations, the most straightforward solution for us remains to exclude these specific agents from handling conversations in Greek.

I would be happy to discuss any further insights or suggestions you may have regarding this issue.

Reply