Embedded into enterprise systems, artificial intelligence (AI) and machine learning let customers augment and automate repetitive tasks and unlock entirely new kinds of digital innovation by learning from data, rather than programming explicit rules.
Natively integrated into SAP applications, cloud, and business networks, AI ensures that digital intelligence can be easily consumed across the entire business to create better customer service, optimize business operations, improve employee job satisfaction, reimagine existing business processes, and more.
It ranges from well-established product offerings, such as SAP HANA’s Predictive Analysis Library (PAL) and SAP Predictive Analytics, to services offered through SAP AI Business Services.
SAP offers a variety of intelligent applications for AI, each designed to precisely meet a unique business need. We’ll walk through each in the following sections.
SAP AI Business Services Powered by Data Intelligence
This application provides an enterprise-grade platform for machine learning in the cloud. It enables simple consumption and tight integration with SAP’s enterprise software combined with openness toward various machine learning technologies. Developers can benefit from a scalable and secure platform to augment business processes with machine learning technology and to infuse applications with intelligence.
SAP AI Business Services allows you to deploy, publish, and run a machine learning model as a service. The lifecycle of the machine learning model is shown in the figure below.
SAP Conversational AI
Formerly known as Recast.AI, this is the leading AI bot platform for enterprises. With more than 30,000 developers building more than 60,000 bots, SAP offers a world-class technology, an end-to-end bot platform, and off-the-shelf customer support bots to lead the revolution of customer relations around the world and enable the intelligent enterprise. SAP Conversational AI can be used to integrate with multiple platforms, such as Facebook, Twitter, Slack, and so on, to create seamless social connectivity.
This is the first step toward a digital assistant, and it supports daily tasks by offering relevant action options based on the consumer’s role, context, and business situation. For example, it allows the user to search for business information or chat in business context with experts to help find solutions to a current problem. Based on the context of the screen, a user can create, collect, and share artifacts such as notes, objects, messages, screenshots, and quick actions. Features of SAP Conversational AI include the following:
- Digital assistant: Advancements include natural language interaction (NLI).
- Notes and screenshots: Create notes and capture screenshots from apps, and then navigate to the app from the screenshot. Annotations can be added, and areas can be blacked out.
- Recognizing business objects: Business objects within the current application context as well as those referred to in notes or chats are recognized.
- In-context chat: Chat with other users from your business application context, sharing notes, screenshots, and business objects. You can also save the conversations for later use.
SAP Intelligent Robotic Process Automation
SAP acquired Contextor SAS, a European leader in the design and integration of robotic process automation (RPA), to help SAP accelerate the development and expansion of its SAP Intelligent RPA portfolio. RPA is a software robot (also called a bot or digital assistant) that is executed on the end-user’s machines or servers and either in the foreground or background. These bots are mainly used to automate labor-intensive, monotonous, and repetitive tasks to give end users more time to perform higher value tasks.
SAP Intelligent RPA provides traditional RPA capabilities along with seamless integration possibilities with technologies such as SAP Conversational AI, SAP Workflow Management, and various other services (e.g., document processing and machine learning models). It enables expert developers, citizen developers, and business process experts to build bots in the following ways:
- Allocating more effort toward higher value-added activities
- Addressing the problematic issue of staff attrition
- Saving money as RPA software can cost less than employees for comparable workload
- Avoiding errors as RPA tools don’t make keying errors
- Time stamping, tracking, and auditing automated work
- Reducing the need for multilingual capabilities
- Providing agility and resilience to support strategic initiatives, for example, moving to SAP S/4HANA
- Increasing compliance and analysis capabilities
- Designing RPA tools to operate 24/7
The overall architecture of SAP Intelligent RPA contains three components:
This is the development environment for building bots. Initially only a desktop version of the studio was available, but a more improvised cloud version is now available. This is a low code/no code infrastructure and SAP’s long-term investment strategy for the bot building environment. To understand the intent behind having a cloud studio, you need to understand the word “hyperautomation,” which means finding automation everywhere: in every company, every LoB, and every industry; for everything: every technology is a candidate to be automated; and, of course, for everyone: business analysts, citizen developers, or bots expert developers. Everyone will be able to build their own bots.
This is the application to be installed on the end-user’s machine. The bot execution, be it attended mode or unattended mode, happens via this component.
This is the central component in the overall architecture of SAP Intelligent RPA. This is where the bot administrator will configure, schedule, monitor, and orchestrate the bots to the end-user’s machine. The most differentiating factor between SAP Intelligent RPA and other RPA products in the market is the Bot Store, which is also an embedded component within this cloud factory. The Bot Store provides ready-made bots for processes in SAP ERP as well as in SAP S/4HANA. There are multiple standard bots in almost all LoBs. With every passing quarter, the count of these standard bots increases. The activation of these bots is very easy with a complete menu-driven approach of their deployment.
Alternatively, with the intention to automate the simplest of the tasks and to enable the real end users to start building bots for themselves, SAP came up with the cloud studio, which is released as a low code/no code environment. Cloud studio is a browser-based environment and tight coupled with your cloud factory component. The bot-building experience is substantially enhanced with the following:
- No low/low code approach for building automations
- Seamless collaborative bot building
- Enhanced and simplified testing tool user experience
- Auto detection of the user interface (UI) technology during screen capture
- Improved screen capture and workflow designer UX
- Deeper understanding of filter criteria via validations within the document object model (DOM) structure
- Completely automated package generation and deployment because the cloud studio is the orchestrator for cloud factory
Now, let’s walk through some key capabilities of SAP Intelligent RPA bots:
Manage Payment Advice
This is an intelligent automation utilizing the standard machine learning model delivered as part of SAP Cash Application.
It basically automates the process of importing the payment advice files to the system for different company codes and then triggers the notifications to the user with the payment advice number and status in a Microsoft Excel file to ask the user to confirm the payment advice.
The bot is a thorough example of how intelligence can be derived by making the bot work with daily-use applications, such as Outlook, Microsoft Excel, Microsoft Windows file browser, and PDF documents.
Create Supplier Invoices from Spreadsheets
The bot can automatically fetch a variety of source files and convert the source files to destination files with a unified format that the bot can recognize. Afterwards, the bot will create supplier invoices via the Supplier Invoices – Create, Read, Release, Reverse API and can also upload additional attachments via the Attachments API. This can avoid manual work and reduce human errors.
Automated Upload of General Ledger Entries
Automated Upload of General Ledger Entries can automate the upload of manual general journal entry vouchers into SAP S/4HANA. It can substantially reduce the time required and provide file tracking to ensure that all files are both uploaded and posted.
The RPA bot, in step 1, will scan through your inbox for a specific subject line variant, pick up the attached files, and save them in a specific folder (root folder) on your local machine. As part of step 2, the bot will pick those files and process them into SAP S/4HANA. Upon completion, the bot will then create appropriate success and results folders and create the files based on whether the processing was successful or erroneous.
This way of working, where the bots create folders for each of the steps, is also a systematic approach so that you can investigate at the folder level whether something isn’t working as expected.
Automatic Creation of Sales Orders from Excel
This API-based bot simplifies the task for internal sales representatives immensely. Most of the time, the internal sales representatives have to create sales orders manually and individually in the system based on the received spreadsheets. This is a very time-consuming and error-prone process, which adds unnecessary manual effort and poses potential risks to the business. This bot automates the task and relieves the internal sales representatives of the manual work.
This bot imports the spreadsheet from a predefined folder to the SAP S/4HANA system and sales orders are created automatically. When sales orders are created, relevant sales employees and their customers will be notified automatically through email.
Artificial Intelligence in SAP S/4HANA
How do the SAP AI Business Services capabilities fit into the backend processes of SAP S/4HANA? SAP enables its customers to reimagine the backend processes more efficiently, more securely, and more transparently as a new solution or as a middleware intelligence product. Singular business processes that can have a heavy impact in each business segment are coupled with machine learning to make them more intelligent and autonomous. Some of the examples are as follows:
Goods and services classification, product category normalization, invoice payment block, and image-based detection of invoice.
Sales and Marketing
Brand impact analysis, quotation conversion probability rate, customer retention, sales forecast, selling recommender, and more.
SAP Predictive Asset Insights, quality inspection through image processing, and stock substitution.
SAP Cash Application, accounts payable, remittance advices, cash and liquidity management, dispute proposal, and cash collection reminder. SAP Cash Application automates the process of matching incoming payments to open receivables, reducing the time required.
Conversation commerce, service ticketing, customer support, and solution recommender. Service Ticket Intelligence builds a model based on successful past ticket completion, using it to automatically categorize service tickets and provide recommended solutions to service agents.
Learning recommender, career path recommender, simultaneous training content translation, job standardization, and resume matching.
Semantic search, text analysis on master data, business rule mining, and deduplication
One important piece of an intelligent enterprise is utilizing artificial intelligence to streamline processes and free up human labor for other tasks. This blog post introduced you to the artificial intelligence capabilities for those running SAP.
Editor’s note: This post has been adapted from a section of the book SAP S/4HANA: An Introduction by Devraj Bardhan, Axel Baumgartl, Nga-Sze Choi, Mark Dudgeon, Piotr Górecki, Asidhara Lahiri, Bert Meijerink, and Andrew Worsley-Tonks.