Celonis AI Lab Chicago: Leading companies build Enterprise AI solutions

In March, Celonis commenced its first AI Lab of the calendar year in Chicago—our ongoing series of one-day AI hackathons designed to turn ambition into architecture. Throughout the day, participants worked side-by-side with Celonis solution experts to build high-impact prototypes tailored to their most pressing use cases.

Using Celonis Platform functionality, like Annotation Builder, Orchestration Engine, and Prediction Builder, teams transformed their own unique process data and business context into tangible AI solutions.

At the end of a busy day, each team had five minutes to showcase their solution. During their presentation, they explained the business opportunity they addressed, the potential impact of their solution, and most importantly, how they used AI to achieve it. This hackathon was more than problem solving–it was about collaboration, competition, and innovation.

Here’s a round-up of what our customers built in Chicago.

Global Tier 1 Automotive Supplier: Product lifecycle management prediction

Opportunity: Effective product lifecycle management (PLM) enables manufacturers to reduce costs, improve customer satisfaction, ensure product quality, and mitigate regulatory risk. Change requests are a critical part of PLM. However, the change request process can be bogged down by overly long cycle times, blockages at feedback points, target dates that aren’t properly prioritized, and traditional BI tools that provide limited visibility into ongoing and block requests.

Solution: The team built a solution on the Celonis Platform to improve both cost realization in PLM and customer satisfaction. The solution uses machine learning to calculate an expected cycle time for each change request and predict which change requests are likely to get stuck during feedback stages. The team then built a Process Copilot that gives change managers customizable access to the current state of change requests so they can identify stuck requests that need proactive outreach to be moved forward.

Global Food and Pet Care Company: Order risk assessment assistant

Opportunity: Organizations often face a lack of end-to-end visibility into financial risk as data remains siloed across contracts, purchase orders, goods receipts, and invoices. This decentralization makes it difficult to maintain consistent control compliance, such as three-way matches or contract adherence, leading to increased risk exposure. With a massive volume of high-value transactions moving through the S2P (Source-to-Pay) cycle, manual and reactive exception handling becomes an inefficient bottleneck, preventing teams from effectively flagging and triaging the most critical financial discrepancies.

Solution: The team built an operational application for a controls framework that ensures consistent validation across the entire S2P lifecycle. Leveraging Celonis Annotation Builder, the solution identifies exceptions and assigns them a dynamic risk score with actionable insights, allowing analysts to prioritize the highest-impact issues and accelerate triage. This model is designed to evolve into an autonomous agent that intelligently routes exceptions to the appropriate analysts, streamlining resolution and eliminating manual overhead.

USG: Order-to-Cash

Opportunity: Across many industries, the transition from physical logistics to digital billing is too often a time-consuming manual process, which is prone to errors. Consider the movement of raw materials by semi-trailer truck in the United States. Trucks are loaded with material and move from point to point within the supply chain. As they travel American roads, the trucks are inspected and weighed to ensure they comply with highway regulations. At many companies, when trucks are weighed, physical weigh tickets and bills of lading must be scanned and manually transcribed into Excel by a centralized team. This "analog-to-digital" bottleneck is prone to entry errors that cascade downstream, resulting in data quality issues and manual rework. Furthermore, this manual process is difficult to scale as companies expand and the workload increases.

Solution: Using a Large Language Model (LLM), the USG team developed an automated ingestion engine using Celonis to read, digitize, and extract structured data directly from images of physical tickets and Bills of Lading. This data is then mapped directly into the O2C (Order-to-Cash) data model, ensuring consistency across systems. To bridge the gap to current workflows, the solution automatically formats validated data into standardized Excel invoices. To ensure 100% accuracy, the system includes a streamlined error-handling and approvals workflow that flags any low-confidence extractions for human review, significantly reducing manual effort while maintaining data integrity.

In addition to these innovations, a global materials science company leveraged Prediction Builder and Annotation Builder to tackle 3-way match processes, generating proactive predictions to identify potential mismatches before they occur, providing automated handling recommendations that prevent payment delays. Meanwhile, an American commercial vehicles manufacturer focused on supply chain resilience using Prediction Builder and Process Copilot. This team created an intelligent assistant for inventory planning, allowing their teams to optimize stock levels and respond dynamically to shifting logistics timelines.

Discover more from AI Lab at Celonis

We are planning a full slate of AI Lab events, bringing innovation to more cities and even more customers for collaboration and breakthrough AI solutions. And you can catch up on the groundbreaking AI solutions developed during previous AI Lab events in the following articles: