Hello ​@TIBLEOFORCE ,

Interesting issue! 

Just to double-check — could you please confirm if the required permission set is assigned to the user you’re testing with?
Also, it would be great if you could share a screenshot of what you see on the screen, so we can better understand the behavior.

Normally, even without all the advanced configurations, if you’re on IFS Cloud 25R1 and the FNDGPT module is enabled, the Copilot chatbot should appear.

Please verify the following:

• The permission set FNDGPT_RUNTIME is granted to the user.

• To allow a user to administer and publish prompts, the permission set FNDGPT_ADMIN must also be assigned.

If the user encounters an error while using IFS.ai Copilot, navigate to:
Solution Manager → Copilot Management → Copilot Error Logs
This page provides detailed error messages that can help identify the root cause quickly.

Thanks,
Hardik

Hi there 

we fixed all the permissions and this might have done the trick because it is showing up now. Thank you.

However we are getting a pretty generic error when interacting with it: 

“ I apologize, an error has occurred. Please, try again later.”

In the log we get an error “500 Internal Server Error” / “Error when getting recipe”.

any idea or clue on how to solve it? 

Hello ​@TIBLEOFORCE ,

If you encounter an error while using IFS.ai Copilot, navigate to:
Solution Manager → Copilot Management → Copilot Error Logs
This page provides detailed error messages that can help identify the root cause quickly.

can you share what do you in see in logs?

Thanks,
Hardik
 

Hi there

I see this error:

500 Internal Server Error” / “Error when getting recipe”.

Hello ​@TIBLEOFORCE ,

can you check if the region is correct in nxsuri? -- i see “eu”
But your IFS.AI tenant seems to be in the US region (Region: US).
this mismatch (EU vs US) is one of the most common causes of 500 “recipe” errors.

can you try changing the URLs in your YAML to US-region endpoints and redeploy?

Thanks,
Hardik

Hello

that is a good catch.

I did not find in the documents the URL for the endpoints in the US.

I will assume this is something like:

“https://plt-us.nxs-stg.ifs.cloud”

would be that correct? 
 

Tks

Leo

Hi ​@TIBLEOFORCE 

Unfortunately there has been an issue with the deployment of the Coplot recipe files for customers activating IFS.ai. The issue will be solved in 25R1 service update 5 (and 24R2 SU11) and there is also a fix that can be deployed without taking he service update. Please refer to this knowledge base article for the fix: IFS Product - IFS.ai Copilot chat gives error "I apologize, an error has occurred. Please try again later."
 

Sincere apologies for the inconvenience and the issues you have experienced!

/Kristina, product manager for IFS.ai Copilot

 

Hi, Kristina.

Thank you for replying to this. 
I took a look at the KB you sent and we deployed the files for the 25R1 zip file from there. 

I also changed the tenant to be in EU instead of US, changed the customvalues.yaml and we run the installation again.

But we are still getting errors but the error has changed to Nexus.

“500 Internal Server Error
Error fetching Nexus access token.”

Anything I am doing wrongly? The tenant shows status CONFIGURED, so the NONCE worked etc.

Also, now the CHAT shows the prompts which were not there before.

But still getting this 
 

Any ideas?

best regards,
Leonardo Nobrega

For some reason the connection to the Nexus platform is not working now. That is unfortunately not within my competence, but let me send this post to someone else who might be able to answer! ​@TIBLEOFORCE 

Is there a support task for this? 

No, we are trying to figure out ourselves first since we are in production yet. I have a platform engineering team so I'm first trying to reach a place where either it gets working or there is a real issue.

We could advance it since my last message. We fixed Nexus URLs, recreated the tenant to use EU instead of US, deployed the fix from the KB, etc.

Specially after the doc to connect IFS.AI was updated with the correct URLs (It was missing it before): https://docs.ifs.com/techdocs/25r1/070_remote_deploy/600_connecting_to_ifsai/#configuration_steps_-_connecting_to_ifs_ai

We are still getting the 500 Internal Server Error, however the error message is now different and now we are getting this error of 401 unauthorized. 

ML service response error.ML service error: <html><head><title>401 Unauthorized</title></head><body style='text-align:center;'><h1>401</h1><h2>Unauthorized</h2></body></html>

Any idea?