IT Services Management with SAP Intelligent RPA and SAP Conversational AI (part I) (Challenge Submission)
In my previous blog, I explained how to interface SAP Conversational AI and SAP Focused Insights for SAP Solution Manager to build powerful digital assistants transforming customer and employee experiences for your IT services.
In this serie of tutorials, we will explore the different possibilities to interface SAP Intelligent RPA, SAP Conversational AI and Focused Insights for SAP Solution Manager.
Chat-bots and Automation bots are a natural fit: Chatbots provides the natural language interface while automation bots execute attended or unattended workflows.
In this picture, Focused Insights for SAP Solution Manager is acting as a unified analytics engine with generic APIs to consume, expose and publish metrics for the bots to push communication to customers.
SAP departments have now the opportunities to build and deliver new capabilities for IT services and SAP Application Management.
SAP Intelligent Robotic Process Automation (SAP Intelligent RPA)
SAP Intelligent RPA automates repetitive manual processes by creating, scheduling, managing, and monitoring bots.
In IT function specifically, SAP Intelligent RPA can improve service desk operations, optimise systems monitoring, detect security attack, perform backups, request authorisations, trigger testing, deliver virtual machines, …
SAP Intelligent RPA relies on 3 components:
- Desktop Agents executing process automation for attended bot (digital assistant) and unattended bot (digital worker).
- SAP Cloud Platform orchestrating & monitoring process automation by managing agent hierarchies, environment, configuring and distributing packages.
- Desktop Studio offering an integrated development to define, build and test process automation.
SAP Intelligent RPA manages the following artefacts:
- Applications implementing the automation logic. Applications are built with the SAP Intelligent RPA desktop studio and imported in the SAP Cloud Platform factory.
- Environments containing hierarchies of agents (Machines or logged-in user accounts).
- Deployments mapping Applications to Environments. It maps the workflows content to the execution environments.
- Trigger defining the execution schedule via jobs.
SAP Intelligent RPA interfaces
SAP Intelligent RPA manages four types of APIs:
- API Notifier: A REST call executed when a job execution status reaches a specified state (START, SUCCESS, FAILURE, CANCEL).
- API Trigger: Start the execution of a scenario or a process in a specified deployment from an external application via a HTTP POST call.
- CAI Notifier: Send a message or update a SAP Conversational AI conversation when a job execution status reaches a specified state (START, SUCCESS, FAILURE, CANCEL).
- WEB Services CALL: A Web Service call triggered from an SAP Intelligent RPA process or scenario.
Based on those APIs, we can support the following architectural model:
- SAP Intelligent RPA provides the automation layer but also captures data.
- SAP Conversational AI is able to trigger processes based on contextual data while providing access of dynamic insights derived from the end-user conversation.
1) API Trigger
1.1) Create the API Trigger
To start an execution of a scenario or a process in a specified deployment from an external application via a HTTP POST call, we need to declare an entry point by adding a new API trigger entry to the target deployment.
The API trigger provides an URL and a token for external calls.
1.2) Invocation of the API Trigger
The SAP Intelligent RPA trigger is accessed from the SAP Conversational AI skill is defined in a Webhook:
The header should contain the SAP Intelligent RPA trigger token to identify the correct trigger.
Reference the SAP Conversational AI conversation
The correlation ID of the bot conversion is added to the body of the post message.
2) CAI Notifier
2.1) Create the CAI notifier
The CAI notifier is created by adding a new entry to an existing trigger API.
The main attributes of the CAI notifier are the conversion ID to map the chatbot instance and the type of reaction: a message send or a memory conversation update.
In addition, the notifier can be accessed on different states of the scenario: Start, Success, Failure.
3) BOT execution
During the execution of the Chat-bot, when the WEBhook is triggered, the execution of the SAP Intelligent RPA scenario will return a message displayed in the conversion flow:
3.1) Trigger SAP Intelligent RPA scenario
3.2) Update SAP Conversational AI chatbot with SAP Intelligent RPA scenario execution status
Source of information
Trial blog: https://blogs.sap.com/2019/12/26/sap-intelligent-rpa-trial-get-ready-to-build-bots…/
Open SAP Training: https://open.sap.com/courses/rpa2
software Download: https://tools.hana.ondemand.com/#cloud
SAP HELP: https://help.sap.com/viewer/c836fab4182e45548b6c6c6d0d0a9146/Cloud/en-US/f2e3d2d3e84d42eabe14913f7ee9269c.html
Great blog post Xavier DUPEYRAT!
Definitely, articulating SAP Intelligent RPA with SAP Conversational AI brings Digital Assistants to an unprecedented level of efficiency.
Robotic Process Automation + AI-based chatbots = ??
excellent post, I want to make a prototype with your guide.
thanks for this great blog ! It is working great for me, but i just can't get to have the answer from the iRPA when it starts the bot. I get the following message:
I think it is related to the token i need to insert into the notifier right ? Wich one is the right token ?
Hello Xavier DUPEYRAT
thank you for your post. I tried to make a prototype using your tutorial.
However, I always get a following error - message:
"error": "Error requesting oAuth access token (error-reference: d788c7d2-c8cf-427b-8b78-d09b067eb8b4)"
I think I have a mismatch in tockens. Am I right?
Thank you in advance!
Hi Xavier, thanks for the blog.
When we execute the chat bot,the automation bot is being called in Cloud factory but it did not executed completely, it is in ready status. I have added CAI notifier as well.
Do we need to perform any other steps to auto run?
we are facing the same problem. We are waiting for around 15mins till the job runs.
Hi Yiwei & Vijay
Were you able to resolve this issue?
No, not yet
Hi Vijay Vegesana
You have to make sure your bot is in Unattended mode. only if it is in the unattended mode it can execute the automated task from CAI.
You can find your Agent status in the Agent Tab of Cloud Factory
HI Xavier DUPEYRAT,
I create a notifier and my bot executed all the automation processes. In CAI Notifier it shows error.
I have provided proper OAuth Token and my Automation is completed. Error in reply to the chatbot.
IS this issue resolved? I am facing similar issue. Can you please provide the solution that worked for you!
I have the same problem. The status is reported really poorly from SAPs side. Why don't you let us see some logs so we can actually solve the error?
Someone has a solution for this?