Next article
Celonis + IBM watsonx Orchestrate: AI agents and automation enabled by Process Intelligence
As intelligent agents augment human decision-making and transform how work flows through the enterprise, companies that can scale the value of their data with AI will gain a marked advantage.
Earlier this year, Celonis and IBM showed how GenAI is helping companies unlock the value of their unstructured data through the combination of IBM watsonx and Celonis Process Intelligence. Today, the companies have extended their partnership with the integration of the Celonis Process Intelligence Graph and IBM watsonx Orchestrate. This combination lets customers seamlessly identify automation opportunities for AI agents and then build, run and monitor the AI agents.
“AI’s true potential is its ability to transform processes, and to do that it needs process intelligence,” said Marc Kinast, VP of Corporate Development at Celonis. “Process Intelligence provides AI the context it needs to make the right decisions. Bringing together IBM watsonx and the Celonis Platform gives our customers an intelligent automation solution infused with process insights that enables end-to-end business transformation and massive value creation.”
“IBM watsonx Orchestrate empowers teams to focus on work that brings true value to the business and not time-consuming tasks,” said Parul Mishra, VP of watsonx Products for Digital Labor at IBM. “Celonis Process Intelligence gives watsonx Orchestrate AI agents the context to understand how processes run across each organization and lets customers intelligently automate complex tasks within their core processes, like Accounts Payable, Customer Service and Procurement.”
Celonis Process Intelligence + IBM watsonx Orchestrate
At Celosphere 2024, Celonis launched AgentC— a suite of AI agent tools, integrations, and partnerships that enables our community to develop AI agents in the leading AI agent platforms, including IBM watsonx Orchestrate.
IBM watsonx Orchestrate is an AI and automation solution that empowers enterprises to build AI agents and assistants across any application, integrate into existing workflows, and manage and govern them in a responsible way. The new two-way integration between IBM watsonx Orchestrate and the Celonis Process Intelligence platform will allow information and insight to flow between the two. For example, an agent built with watsonx Orchestrate can extract unstructured data from documents, such as emails, call transcripts and sales contracts, and feed it into Celonis. Conversely, Celonis can feed process insights (e.g., recommend best payment terms based on PO, invoice, master data and contracts) into watsonx Orchestrate so AI agents can automatically execute specific tasks, such as scheduling the payment date of a PO.
Celonis provides work instructions and triggers, from which the watsonx Orchestrate AI agents extract the required context and actions. First, the Celonis Process Intelligence Graph ingests and transforms process data into an Object-Centric Data Model, creating a company’s process digital twin. Celonis monitors process performance for deviations and identifies value opportunities (i.e., what to automate and when to trigger AI agents). Using detailed handling instructions, Celonis then describes how to fetch process context and automate process actions. All these process insights are fed into watsonx Orchestrate where the AI agent is configured using the required tools (based on Celonis Intelligence APIs) for context and actions extracted from the handling instructions.
Joint Celonis and IBM customers are already innovating with diverse use cases for watsonx Orchestrate AI Agents fed with Celonis Process Intelligence:
Multinational CPG company - Payment terms: Mismatches between the payment terms in a company’s ERP master data, purchase orders, and invoices can lead to business slowdowns, additional costs and damaged supplier relationships. To resolve these mismatches, teams need a single source of truth for payment terms. Unfortunately, for many organizations that source of truth doesn’t exist within their ERP system, nor in the order or invoice documents; instead the correct payment terms are “buried” in contracts as unstructured data stored in a spend management system. Manually digging through this data can be like trying to find a needle in a haystack, but with AI and process intelligence that search is streamlined and the results more accurate. A multinational CPG company is using Celonis and watsonx Orchestrate to tackle this exact problem—an opportunity to improve its working capital by ~$9M. Here’s how it works. The watsonx Orchestrate AI agent extracts the true payment terms from vendor contracts and passess the information to Celonis. The Celonis platform matches the contract payment terms with the payment terms in the ERP vendor master data, purchase orders and invoices. Finally, an AI agent compares these four payment terms and delivers a recommended fix for the mismatches to the Accounts Payable (AP) team members.
Global construction materials company - Recategorization of AP and Procurement incidents: Celonis and IBM Consulting collaborated to create a reusable, AI agent for the Celonis ecosystem that was leveraged by a global construction materials company. The AI agent allows Celonis to integrate service incidents (such as payment requests or invoice inquiries) with the broader ERP process. Specifically, the AI agent both classifies and extracts needed data points from free-text incidents so that Celonis could tie them back to their POs and invoices. Additionally, this allows users to take direct action from Celonis. For example, the user can have the AI agent generate responses to those incidents and send out an email response to the ticket with a few clicks. With this solution, the company was able to reduce the time required to review credit blocks, credit managers cut through a backlog of blocked orders, and the team can now process 5X as many blocks each day, without additional risk.
To learn more about how Celonis enables companies to create effective AI agents and orchestrate them once they’re up and running, read Manuel Haug’s latest article, How Process Intelligence helps companies build effective AI agents.
