Enterprise Automation: Deep Dive into the new Generative AI capabilities!

[EvgeniyGorbunov]

After the introduction to the topic by Thomas Volmering, I would be happy to support you with the deep dive into the AI capabilities of Enterprise Automation in this blog post.

1. SAP Signavio which provide visibility into business processes, potential bottlenecks, root-cause analysis, and suggestions for improvement.
2. SAP Build Process Automation which automates workflow processes and tasks with drag-and-drop simplicity using pre-built workflows, bots and connectors for SAP and non-SAP systems.
3. SAP Integration Suite which helps to seamlessly integrate diverse capabilities available in complex IT landscapes with proven pre-built integrations, APIs, connectors, and best practices.

Generative AI in SAP Signavio

In order to leverage Large Language Models (LLMs) technologies for business processes, we introduced the concept of Large Process Models (LPM). This new concept was presented last year by Dr. Andreas Gerber, VP Head of Engineering SAP Signavio Process AI.  The idea is to go far beyond the traditional LLMs  by finetuning them and augmenting them to focus on their application in the area of business processes.  

For that, Large Process Models are combining generic LLM AI technologies with the know-how of decades of process experience around SAP from thousands of experts. They are trained with anonymized customer data from various customers and refined with domain- or industry-specific fine-tuning, and customer-specific fine-tuning. This broad and diverse training set helps ensure that our process AI capabilities are uniquely equipped to tackle real-world business and process challenges with precision, drawing on deep industry knowledge and process understanding. 

Our LPM-powered process AI capabilities have the exceptional ability to answer your specific process questions and help you uncover issues you were not aware of. They act as your trusted advisor, providing you with ready-to-consume AI-powered recommendations you can apply right away or use as a foundation to drive process enhancements. 

Upcoming capabilities in SAP Signavio powered by GenAI  leveraging Large Process Models will be:

Process Recommender 

Our AI-assisted process recommender ​will provide process and domain owners with the ability to obtain preconfigured process models that are a best match to their needs, based on a database of 5,000 best practices. ​

With this capability, you will narrow down thousands of options to those process models that best suit the needs of your organization. ​

Recommendations are grounded in years of SAP experience working side by side with industry leaders and thousands of projects delivered. ​

Planned for Q3 2024, Process Recommender will bring key benefits such as: ​

• Instant recommendations based on best practices ​
• Accelerated process model design ​
• No need for consulting-heavy services to define the first process draft to start the modeling activity

 

Performance Indicators Recommender

Our AI-assisted Performance Indicators Recommender ​capability will provide process owners and process analysts instant recommendations on the process performance indicators to be monitored.

Users can move from problem identification to a measurement approach with no effort by leveraging a repository of thousands KPIs and PPIs

Key benefits of Performance Indicators Recommender, to be available in Q3 2024, include: ​

• Instant recommendations based on best practices ​
• Simple connection between business problem, affected process and relevant metric to be measured
•  Self-service approach to define initial process monitoring framework

SAP Signavio Process Analyzer

 Our AI-assisted Process Analyzer, empowers users with any skillset to ask questions in natural language and receive immediate insights that can inform business decisions.

Key Benefits ​of this new module, scheduled for Q4 2024, include:

• Correlation and anomaly detection easily consumable by business users ​
• Democratized process mining ​
• Accuracy and reliability and relevance of the analysis output based on business user prompts  ​

 

Generative AI in SAP Build Process Automation, with Executable Process AI

Our aim for Process Automation is to go further, and we now want to provide Executable Process AI and more.

Indeed, there are challenges when automating processes:

• No active assistance to business experts while starting automation of critical processes
• Slow pace of incrementing large process automation​
• Difficulty to comprehend already modelled artifacts when lacking adequate documentation

To tackle these challenges, we decided that SAP Build Process Automation should include

• Conversational AI capabilities within the studio

• Semantic data access and knowledge graph access within automation and process design time
• Auto-generation of automations and process segments based on embedded intelligence within the studio

6 months ago, at SAP TechEd 2023, we unveiled a first demo of Joule copilot creating the workflow for an approval process. You can see in this video a preview of how generative AI will help creating a new process workflow automation, with a business rule based on a decision table for approval.

Our solution, SAP Build Process Automation, will soon bring embedded generative AI capabilities to generate processes, automations, and decisions, including:

• Interactive Model Generation and Modifications
• Summarization of Complex Artifacts in natural language

With Artifacts Generation and Modification, you will be able to describe your process in natural language, interact and iterate to finalize desired process flow and finally generate the process that fits your requirement. It will generate processes, automations, forms, and decisions (business policies) using the inputs provided by business expert in natural language. And of course, generated artifacts which can be reviewed and modified as per need.

Summarization will create a short or comprehensive description of complex historical or newly generated complex artifacts, for instance the summary of a complex decision table. You will be able to share or save the description along with the artifact.

In a few words, SAP Build Process Automation will bring tremendous benefits to our customers, by

• offering a conversational experience for citizen developers
• accelerating time to value for authoring automations and processes
• standardizing SAP application access for automations through a semantic access layer fostering reuse and fail-safe connectivity

These evolutions are planned for Q3 2024, and you can discover all the improvements planned in Process Automation by browsing SAP Build Process Automation public roadmap.

Generative AI in SAP Integration Suite

Integration, the foundation of an intelligent enterprise, is becoming the number one strategic priority for organizations. Only by connecting customer, product, employee, and brand experiences with business operations, continuously and in real time, can organizations exploit the full potential of their system landscape and make the right decision at the right time. With SAP Integration Suite, you can integrate solutions inside and outside your data center. You can accelerate connectivity with prebuilt integration flows, connectors to hundreds of systems, thousands of APIs, and other capabilities that simplify integrations with SAP and third-party applications. As shown in the figure, connecting people in real time becomes a reality with event- driven architectures and intelligent technology that helps you automate comprehensive business processes to gain efficiency and visibility.

There are a few AI-based capabilities in SAP Integration Suite already, such as a conversational assistant in SAP Business Accelerator Hub (beta), flow step recommendations, and recommendations for interfaces and mappings.

Now we are working to further integrate AI for 2 purposes:

• Generation of Integration
• Anomaly Detection

Generation of Integration

Our customers are facing complex integration challenges with the diverse applications and systems in their landscape. Integrators must have the right domain & technical expertise and invest a lot of time and effort to design new integrations & APIs.

Our solution is to leverage AI to generate Integrations: SAP Integration Site will be able to create new integration flows, based on the specific integration scenarios, different connecting applications, and business partners. It will leveraging Unified Customer Landscape – UCL - to discover the customer landscape and offer seamless integrations between those systems, and provide pre-configured connectors and scheduler.

Anomaly Detection

It is important to monitor deviations from the normal usage or performance in the API traffic, as it indicates security threats, system malfunctions or other issues that hinder developer & user experience.

Empowering our customers to maintain security, reliability & performance of their APIs, we will leverage AI for anomaly detection, identifying unusual or abnormal behaviors by continuously monitoring the API traffic. This AI-based mechanism will immediately notify API owners upon detection of anomalies in API Call Volume, API Error count, API Response time, while providing insights to aid investigate the cause & analyze impact and recommend remedial actions.

Watch the demo of Integration Suite new AI capabilities during Juergen Mueller's Innovation Keynote at SAP Sapphire 2024:

Click on the image to watch the video.

Thus, SAP Integration Suite will enable organizations to deliver better services & experience and hence achieve their business goals.

Demos of advanced Enterprise Automation at SAP Sapphire 2024

On June 4th at SAP Sapphire 2024, SAP CTO Juergen Mueller showcased an end-to-end demo of Enterprise Automation, in the context of extending SAP ERP in compliance with our Clean Core strategy:

Click on the image to watch the video.

First, new SAP Signavio Benchmarking Analytics presents a comprehensive dashboard of the business processes, called Process Performance Overview, which shows that the Lead to Cash process does not perform as well as it could. In particular, the analysis shows that many sales orders involve a manual step because delivery block is not removed after a customer pays an overdue bill.

Second, by using the new Build Extension Wizard embedded in SAP S/4HANA Cloud, one creates a new extension project:  a Business Event, occurring when a Journal Entry is created, triggers an automation created with SAP Build Process Automation which automatically removes the delivery block after the overdue bill is finally paid by the customer. That’s it!

Now, let us look into the other business scenario in more details: how can enterprise automation support us in intelligent processing of blocked sales orders.

This demo features the incremental value of all three components within the enterprise automation portfolio and highlights the latest advancements and innovative AI-based features that were added to SAP Build Process Automation, SAP Signavio, and SAP Integration Suite.

We start this use case with SAP Signavio Process Insights, where we can evaluate the entire process landscape and identify inefficiencies in the lead-to-cash line of business. We continue with a deep dive to the processes to identify the highest improvement potential. This analysis is supported by SAP Signavio Process AI, which is being used for intelligent language prompts, allowing us to find the right BPMN model and search for KPIs.

Next, we receive the recommendation that we have a very low automation rate for processing sales order documents that have a billing block attached to it. To solve the problem and automate the process, we jump to the SAP S/4HANA Extension Wizard, allowing us to create needed extensions, in this case to define and initiate a process in SAP Build.

Event Mesh service from the integration suite is being used to facilitate the event-based communications.

In SAP Build Process Automation, we can leverage executable process AI capability, allowing us to design the needed approval flow for the sales order. In this scenario, we will define a new business rule to process sales orders differently, which are exceeding a certain order amount. We can also use the executable process AI functionality to ask for a process summary.

With the help of this, we will be able to remove the billing block and then unblock the sales order.

Next, we can also use actions to trigger integration flows in SAP Integration Suite. The extensive library of connectors and adapters allows us to replicate sales order information into any SAP or third-party application, that is required. In this case, we use the embedded AI functionality to describe the integration that we need, which is going to be generated fully automatically. In this case, this integration flow will be used to replicate the sales order document via a SOAP adapter.

On top of that, with the AI-based anomaly detection in API management, we can also detect failures and analyze the entire API traffic to ensure operational excellence.

Finally, advanced event mesh takes care of high-volume event handling in multiple regions.

So, as you can see, Enterprise Automation suite with SAP Signavio, SAP Build Process Automation and SAP Integration Suite is now more powerful than ever. Up to you now to harness this power to improve the efficiency of your business!