Celonis offers process transparency across company boundaries

With Celonis Business Collaboration Networks, the Munich-based company is pursuing the goal of expanding process intelligence by providing comprehensive insights into cross-company processes. This shared, actionable, information base is intended to enable companies to identify and eliminate weaknesses in processes.

The first users include the electronics retailer Conrad Electronic and the two suppliers, Schukat electronic and TD SYNNEX. The trio had previously used EDI technology, but repeatedly had problems with orders, for example due to outdated prices or delivery dates. With the help of Celonis Networks, they now exchange 5,000 updates on orders every day automatically and in real time.

The shared data includes order information such as the status of the order process, the planned delivery date and shipping details. If problems such as delayed delivery occur, Conrad recognizes this early on and can take action. On the other hand, it is important for suppliers like Schukat to know that a larger order will come in the next quarter.

One man’s purchase is another man’s sale

With regard to process intelligence, the expansion was a logical step, said Manuel Haug, Field CTO of Celonis. “If you take the example of Conrad Electronic and Schukat electronic, the sales process of Schukat is more or less the purchasing process of Conrad.”

According to Haug, Celonis Networks is an extension of the Process Intelligence Graph it introduced a year ago to map the overall processes in the company. Now, individual processes such as purchasing and warehousing or the production lines are also linked with each other.

Integrating processes instead of data

In his keynote speech  to the Celonis customer event, co-CEO Alexander Rinke summed up how it differs from API integration as follows: “API is system first. We, on the other hand, integrate processes with each other, exactly at the information level that is at stake.”

This also benefits security, Haug said, because as a partner you can query the data, but it is not exported. In addition, you can go into the system and specify that you no longer want to share a certain type of information, which then also applies to historical information.

Toolkit for process-intelligent AI agents

Of course, Celonis could not avoid the topic of artificial intelligence and genAI in particular at its company event. It unveiled AgentC, a suite of AI tools, integrations, and partnerships that allow users to build AI agents themselves or leverage AI agents pre-configured by partners. Celonis’ process intelligence provides the data and relevant business context to improve processes across systems, departments, and organizations.

The first platform integrations include Microsoft Copilot Studio, IBM watsonx Orchestrate, Amazon Bedrock Agents and open-source development environments such as CrewAI. For example, Celonis’ cooperation with Microsoft Copilot Studio enables customers to develop AI agents for specific use cases, with which users can then interact via Microsoft programs they use every day, such as Microsoft Teams.

“AI agents can help automate tasks, but they lack the understanding of the company to do so efficiently,” explains co-CEO Rinke. With AgentC, Celonis ensures that every AI agent that customers develop has the necessary knowledge to perform the right actions and achieve the desired results.

AI agent checks credit freezes

Celonis outlined what this can look like in practice using an AI assistant for checking and releasing credit blocks. The Spanish manufacturer Cosentino developed this with the help of Celonis: The agent analyzes blocked orders and comparable process flows based on data from the company’s existing business systems and the digital process twin in Celonis. Based on this, it makes a recommendation for the appropriate measures to be taken by the credit managers.

At the same time, it provides detailed information on why it makes the respective recommendation. With all of this information available in one place, managers can quickly accept or reject the agent’s decision with a single click and provide a justification for it.