Cylonis, the $13 billion German mining startup, takes a diversified approach to artificial intelligence. Like many enterprise software companies, it has been using machine learning models for a number of years, but with the advent of generative AI, it is adding the Copilot feature, which was announced today at Celosphere, the company’s customer conference taking place this week.
The company helps clients understand how work flows through different processes in a company, using software to find inefficiencies in the flow, something traditionally done by higher-priced consultants. Last year, it introduced a new feature that allows customers to view multiple operations by displaying them on a subway-shaped map.
This year, with the advent of generative AI, they added Celonis Copilot, a feature that sits alongside the subway map and allows users to ask questions about what they see.
The company built Copilot on top of the OpenAI API, says CEO and co-founder Alexander Reinke. This has been something they’ve thought about for some time, dating back to GPT-2, but when it really took off this year, they decided to integrate it into the product.
“It took a lot of coordination, proper input, prompting, and vector database work. It wasn’t very easy, but we definitely accelerated our investments in this area because we really saw the potential,” Reinke told TechCrunch.
Unlike Copilot, the company is trying to help customers make the data in Celonis available to a large language model within their companies, while also making it easier for partners or other third-party customers to build applications on top of process data stored on the Celonis platform. Rather than trying to introduce an LLM, they focus on how to handle the variety of data types tracked within Celonis, which can be a challenge for a large language model, especially when each client has different ways of describing elements of the same type. Of the process.
It’s a complex problem, so the company decided to take a multi-pronged approach to solving it. For starters, they offer their customers a standardized, structured way to process data within Celonis, which they call the process data model.
“We are launching this process data model, so customers can consolidate all their processes and scenarios into one view, so that it is naturally connected,” he said. This makes it easier for large language models to understand this data because it is combined into a single entity.
The second entity involves defining the different elements of the process, such as what you mean by on time or by late billing, for example. “We take all the knowledge that we’ve gathered over the many years that we’ve been doing this to define these business definitions,” he said.
Then there’s a whole API layer on it to expose it to the Celonis ecosystem to build applications on top of that, or expose data to LLMs.
However, the power of the ecosystem comes when you combine a process data model with a dictionary of process definitions to build what’s called a process intelligence graph, which reveals connections between different types of data.
“We do this in order to organize business intelligence across systems and departments into one connected product,” he said.
“It basically gives you a common language to describe processes across the company, and it’s completely independent of the systems underneath. This really provides much more value to customers. Faster time to value assessment, and it positions us to build a network and a platform.”
It also makes it easier for customers or external partners to use the data in their own large language model applications.
These products are mostly in special edition at the moment, being designed and tested with customers, but should be released sometime next year.
Celonis has raised $2.4 billion, Per CrunchbaseIt was valued at $13 billion when it raised $1 billion in October 2022.