Celonis buys MIT spin-out Ikigai Labs to give its AI agents a memory of how work happens

Celonis is acquiring Ikigai Labs, a Boston-area decision-intelligence startup built on research from MIT, and using the deal to launch a new platform layer called the Celonis Context Model. The move, announced in May, hands Europe's best-known process mining company new forecasting and simulation tools and, unusually, makes MIT itself a shareholder in Celonis.

Ikigai Labs has around 40 employees and will join Celonis once the deal closes; both companies have declined to give financial terms. Devavrat Shah, Ikigai's founder and chief executive, is also a professor in MIT's electrical engineering and computer science department. Under the agreement, Shah becomes Celonis' chief scientist of enterprise AI, and Celonis gains exclusive rights to several MIT-licensed patents that underpinned Ikigai's technology.

Why a process mining company is buying a forecasting startup

Celonis made its name mapping how work actually flows through large enterprises, pulling data out of systems such as SAP or Salesforce to show where processes break down. That business has increasingly collided with the rush by big software vendors to deploy AI agents inside corporate workflows. Celonis' argument, echoed by outside analysts, is that those agents are only as useful as the operational picture they are given, and most AI systems today still cannot see that picture in real time.

That is the gap the company says its new Context Model is meant to close. Celonis describes it as a continuously updated, real-time digital twin of a business's operations, built by pulling process data and context from every connected system, application and device. The idea is to give AI agents the live situational awareness to act on their own, rather than just summarise what already happened.

"With Ikigai Labs, we're expanding our platform: from transparency about how companies work today to the ability to understand how processes could work in the future," said Carsten Thoma, president of Celonis.

Ikigai Labs supplies the forward-looking half of that pitch. Rather than relying on large language models, its technology uses a graphical modelling approach tuned for the kind of tabular and time-series data found in spreadsheets, databases and cloud warehouses. Before the acquisition, Ikigai already sold tools to large enterprises for shortening supply-chain planning and forecasting cycles, including a live product called AI Co-Planner that pulls demand, supply chain, finance and sales data into a single, continuously updated plan. Ikigai had raised a $25m Series A round in 2023.

What changes for Celonis customers

Folding Ikigai's simulation and forecasting capability into the Context Model lets Celonis argue it can model not just how a company runs today, but how changes to a process might play out before they happen. Shah has said the combined technology gives Celonis the fullest operational picture of a business yet, letting AI agents draw on its past, present and likely future to adapt continuously.

The Context Model is designed to plug into the AI tools enterprises already use, including Microsoft Copilot, Oracle's OCI Enterprise AI, Amazon Bedrock, IBM's watsonx Orchestrate and Anthropic's Claude, alongside data platforms such as Databricks, Snowflake and Microsoft Fabric. Celonis is also extending a separate partnership with spend-management software maker Coupa, linking Coupa's spend data to its own process intelligence layer so finance and procurement teams can automate tasks such as invoice reconciliation and payment management. Coupa's Navi AI agents will draw on that added context, a capability Celonis says will reach customers through Coupa's app marketplace.

"By connecting Celonis Process Intelligence with the Coupa platform, AI agents gain the operational context they need for effective autonomous work," said Dan Brown, Celonis' chief product and engineering officer.

A bet on owning the context layer

The deal comes without a disclosed price tag and without a new funding round attached, a contrast to the $400m Series D extension Celonis closed in August 2022 behind the Qatar Investment Authority. Celonis has raised roughly $1.7bn since its founding in 2011 by Alexander Rinke, Bastian Nominacher and Martin Klenk at the Technical University of Munich, and now employs close to 3,000 people across its dual headquarters in Munich and New York.

What the Ikigai deal buys is positioning rather than scale: a seat at the centre of what several enterprise-software vendors are now racing to call the context layer for agentic AI. Celonis is betting that owning both the record of how a business actually works and the ability to simulate how it might work tomorrow gives it an edge competitors building purely from language models cannot easily copy. Whether enterprise customers pay a premium for that combination, and how quickly the acquisition clears its remaining closing conditions, are the questions Celonis has yet to answer.