CAI and RPA bots for SAP

The importance of intelligent virtual assistants in the corporate context is growing. Experts see them as a key element for working in the future, as they offer great potential for generating competitive advantages. Gartner assumes that as early as 2021, around 25 percent of digital workers will use a virtual assistant on a daily basis. In 2019, only two percent did so.

Automate simple tasks

To date, there are no central assistants in the enterprise environment that are comparable to Alexa or Siri. However, an assistance function can also be achieved here by combining two types of bots: namely specialized Conversational Artificial Intelligence (CAI) and Robotic Process Automation (RPA) bots. Both types of bots are already used in the business environment to automate simple and frequently recurring tasks and queries, but they differ in their function.

Thus, CAI is used to communicate with the user in natural language and offers a simplified user interface for central access to a wide variety of applications. In addition to FAQ bots with pure question-and-answer functions, CAI bots are also used as information bots. In purchasing organizations, for example, they can be used to query the processing status of an order.

With RPA bots, on the other hand, the focus is on the largely autonomous execution of tasks. Work that was previously done manually by humans is thus partially or completely automated. Examples include master data uploads to SAP Ariba, the creation of documents from external applications such as Excel or Outlook, and the replacement of interfaces to third-party systems.

The use of intelligent bots offers numerous business advantages, first and foremost significant time gains due to a higher processing speed. The bots can also be used around the clock and increase data and process quality, as they do not make any mistakes and do not forget any tasks or steps. If necessary, they can be flexibly scaled and copied. In addition, employees can focus on activities with higher added value, which increases their motivation on the one hand and provides the company with a direct financial benefit on the other.

Potential for purchasing processes

In SAP-based purchasing, simpler bots are already used in the area of operational procurement for frequently repeated processes, such as the supplier search or approval processes. However, if RPA and CAI bots are linked in an Enterprise Assistant, their potential can extend to the complete core purchasing processes, as shown in the example of the sourcing process.

PR Processing Bot (RPA): This bot processes and checks incoming purchase requisitions for material group, accounting and source of supply. If no suitable source of supply is found, the bot automatically creates an RFP and sends it to selected suppliers.

Negotiation Bot (CAI): This is a chatbot that guides the supplier through the proposal creation process. The negotiation component is achieved by implementing different negotiation strategies.

Awarding Bot (RPA): This bot compares the received offers, awards the contract according to defined rules and creates a follow-up document.

Combined into an EA, these three bots can run the sourcing process completely autonomously. In each case, a process step is taken over by a bot, which is followed by a subsequent bot after completion. Since the boundaries between the bots are fluid, the result is end-to-end automation. In addition, various machine learning (ML) algorithms are used as "fuel" for the bots.

For example, if the PR processing bot has to determine a material group in free-text purchase orders, it calls up a suitable algorithm. This algorithm suggests a suitable material group based on the texts in historical purchase requisitions. Since the ML algorithm learns with each new purchase requisition, it can support the PR Processing Bot more and more effectively over time.

Apsolut is involved in the development of Enterprise Assistants (EA) specifically for SAP-based purchasing and can draw on an extensive use case library for RPA and CAI bots for this purpose. Although the existing use cases can be implemented with all common technologies, it is recommended for customers to also rely on the corresponding SAP CAI and SAP RPA products due to the native SAP integration.

Apsolut has already implemented solutions for typical standard use cases, such as master data uploads, using these two techniques. For this purpose, SAP's existing customers have access to smaller prefabricated solution modules that can be integrated directly into the SAP purchasing environment. However, if purchasing organizations want to achieve higher efficiency gains and use the bots to simplify complex core processes, solutions must be developed in the form of customized Enterprise Assistants. This requires structured implementation projects that are accompanied and supported with a holistic consulting approach.

Project kick-off with process analysis

Since RPA, CAI and AI/ML implementation projects are not so much about technology but about well-defined purchasing processes, a process analysis is the first step. The goal is to identify optimization potential in the processes and suitable starting points for the use of bots. The requirements for a bot can be recorded in a user story and backlogs can be built up. At the same time, alternatives are examined: A bot should not be implemented just to have a bot implemented.

For rapid scaling, it makes sense to create the bots in an agile approach - for example, Scrum. Experience has shown that bots can be developed quite quickly, so they are also ideal development objects in a sprint. In the case of CAI bots, care should also be taken to ensure that they each have a uniform personality - that is, that they exhibit identical speech, behavior and reaction patterns in communication. In this way, the impression of a complete assistance function can be reinforced.

Then it's on to the design and development phase. To achieve a real assistance function, a certain critical mass of bots must be combined, which are coordinated with each other and mesh seamlessly. Depending on the requirements, the prefabricated solution modules from the use case library can be used, but they usually have to be customized. In other cases, new bots have to be developed.

With RPA and CAI bots, recurring routine tasks in the SAP purchasing environment can be automated - quickly and easily, without changing the existing applications. Combined into innovative Enterprise Assistants, the bots form an important building block for the digitization of procurement, keyword: Purchasing 4.0, which is considered a central lever for future corporate success.