cancel
Showing results for 
Search instead for 
Did you mean: 

Chatbots and licensing limitations

Hemachanderrao
Level 4
Hei, We have managed to integrate a chatbot with blueprism robot successfully during our POC. However, we see that we will get limitations with the number of licenses we have for blueprism that will limit rolling out chatbots to produciton.   The challenge we see is that we have 'x' number of licenses which support 'x' concurrent robots to run. If we implement chatbot, we are sure the concurrent sessions that may be initiated through chat will be over 'x' . Is there a special licensing for such scenarios or did someone manage to overcome this challenge. Please share ideas / experiences.  Thank you in advance
2 REPLIES 2

PSSupport
Staff
Staff
Your digital/robotic workforce solution designs should always be asyncronous - you are simply handing off work to a back office team.  So a valid Chatbox request might be to take a new address from a customer via the chatbot - the final response from the chatbot should then be along the lines of, I have passed your change of address request to the correct team and your accounts will be updated within x minutes/hours (depending upon the SLA for that process).   Even if you had a one to one relationship between the maximum number of clients using chatbots and the number of robots (which even if BP licensing was different would probably be unaffordable in infastructure) - a system update via the UI might take minutes to perform if it is a slow system, are you expecting the customer to just sit there waiting for the back office update to be performed?

Hemachanderrao
Level 4
Thank you for your response. The plan is not to have the user waiting for the bot to complete its task but , as you mentioned, it should return a message saying the request has been accepted and he/she will soon receive a response/outcome.    However, the issue we face is that even a process that gets requests from users and writes to a work-queue can also face concurrent sessions challenge. Do you recommend that its handled by load balancing or just dedicating one resource for handling requests and run it at a schedule ?   Thanks!