Process mining company Celonis has acquired MIT-linked decision intelligence supplier Ikigai Labs to flesh out its offer to eliminate artificial intelligence (AI) “blind spots” from business IT. Linked to the acquisition is the launch of what the supplier calls a “context model” that functions as a real-time digital twin of business operations for customers.
In a press statement, the company said: “As organisations … attempt to deploy enterprise AI, they face a critical challenge of ensuring AI does not have blind spots in understanding how their businesses operate. Without this understanding, AI agents cannot make a real impact, so companies struggle to see meaningful returns on their enterprise AI investments.”
The supplier claimed its Celonis Context Model (CCM) “fixes this by providing a dynamic, real-time digital twin of operations, which translates the business into a language AI understands”.
It added: “Built on process data and business knowledge from every system, application, device and interaction across the business, it gives enterprise AI the operational clarity it needs to reason correctly, act reliably and deliver results at scale.”
Dan Brown, chief product officer and executive vice-president of engineering at Celonis, stated in a blogpost: “AI models don’t know about how your specific invoices are related to your shipping records because that data is proprietary, private and fragmented across internal systems, applications and devices. And without that deterministic foundation – the ground truth of your operational reality – no AI agent can be trusted to make reliable real-time decisions and take actions that effectively drive your business outcomes.”
Enabling better business decisions
Celonis said its acquisition of Ikigai would “bring state-of-the-art enterprise decision intelligence and cutting-edge AI innovation – which includes planning, simulation and forecasting capabilities – to the CCM”.
Ikigai was co-founded in 2019 by Devavrat Shah (CTO and CEO) and Kamal Ahluwalia. Shah holds a professorial chair of AI at MIT. By its own account, on its website, Ikigai Labs specialises in processing and analysing structured enterprise data, such as tabular and time series data.
It offers a generative AI (GenAI) platform built on proprietary large graphical models (LGMs), which are said to be designed for structured enterprise data, in contrast with large language models (LLMs) that focus on unstructured text.
As part of the acquisition agreement, Celonis will gain exclusive rights to MIT-owned patents, which Ikigai had licensed from MIT, and MIT will become a shareholder in Celonis.
In an interview with Computer Weekly, Celonis president Carsten Thoma said: “If you look at the noise level out there, CIOs are completely confused.
“Two years ago, we decided that we wanted to have something like a holistic business graph that can serve as the brains of an operation for a company. We knew from our own insights that application landscapes are super-fragmented, that data lakes are competing, that AI is on the horizon, but that some of the AI is very hard to deploy in an efficient manner.
“AI is only as good as the context it has. Every organisation needs to give its enterprise AI a holistic, living model of how a business truly operates. This has never been possible until now”
Carsten Thoma, Celonis
“We took a step back and thought about what customers need to master those challenges and navigate in a way that they don’t blow the budget, that they don’t start 10 pilots in AI where eight yield no results, where they can really take a position of strength, courage and intelligence to execute on the future.
“We had a few pieces in place already with our process intelligence platform, and then there were a few pieces missing. The last piece is the large graphical model we are acquiring now with Ikigai.”
Thoma said that it now offers a “robust and future-proof approach to provide a control tower and platform for operational context and intelligence”, adding: “It is important to understand we are domain agnostic and system agnostic from the operational context model [point of view] because other vendors talk about specific domains.”
In the press statement, Thoma said: “AI is only as good as the context it has. Every organisation needs to give its enterprise AI a holistic, living model of how a business truly operates. This has never been possible until now. And with Ikigai, we’re making our platform even stronger, extending its intelligence beyond how your business runs today to how it should – and could – run tomorrow.”
Devavrat Shah, now chief scientist for enterprise AI at Celonis, said: “Ikigai was built on a simple but firm conviction: better enterprise decisions require AI that works with enterprise data. Ikigai has proven foundation model technology for structured data at scale; Celonis has encoded enterprise processes. Together, we provide the fullest operational representation of business reality.”
Context essential for trusted AI
Jerome Revish, chief technology officer for digital and technology services at Celonis customer Cardinal Health, a US health services company, said, in support of Celonis’s contextual message, that in the healthcare industry, “precision is paramount” and “AI that’s only right most of the time” is simply not acceptable.
“We use AI as a tool to accelerate operational insight – process context enables agents to support our team in acting with precision. Defining guardrails then gives us the confidence to act,” added Revish. “Ultimately, context is what makes the difference between AI that’s impressive in a demo and AI that’s trusted and safe to deploy.”
Context is what makes the difference between AI that’s impressive in a demo and AI that’s trusted and safe to deploy Jerome Revish, Cardinal Health
Rafael Domene, CIO at Consentino, a Spanish company that makes and distributes surfaces for architecture and design, and whose brands include Silestone, Dekton and Sensa, was also aligned with the contextual message.
“Our goal is to build a digital workforce of AI agents that can run and improve our business operations at scale. What we’ve learned is that an agent is only as good as the context you give it,” he said.
“When you provide AI with a real understanding of your processes – the data, the business rules, the decision logic – it stops being a tool you experiment with and becomes one you trust to act. That’s what makes the difference between an agent that makes a recommendation and one that runs a process,” added Domene.
Chicago-based snacks maker Mondelez International, whose brands include Cadbury, Oreo and Toblerone, is another Celonis customer that says the supplier is helping with its agentic AI programme of work.
“We’re in the middle of one of the most consequential technology transformations in our history while simultaneously building the foundation for agentic AI,” said Filippo Catalano, chief information and digital officer at Mondalez.
“We’ve learned you cannot sustainably deploy and run trusted AI agents across a landscape as complex and varied as ours, unless those agents understand and act based on the reality of how your processes run across every market, system and function – not just how they were designed in theory,” he added.
‘Context graph thesis made real’
Celonis says its platform brings data together with zero-copy integrations to sources like Amazon Web Services, Databricks and Microsoft Fabric, as well as pre-built connectors to Oracle and other enterprise resource planning (ERP) and customer relationship management (CRM) platforms.
It says it has also built integrations with agentic platforms, including Amazon Bedrock, Anthropic’s Claude Cowork, IBM Watsonx Orchestrate, Microsoft Copilot and Agent365, and Oracle Cloud Infrastructure Enterprise AI.
For one of Ikigai Labs’ investors, Foundation Capital, general partner Ashu Garg added: “This is our context graph thesis made real. Celonis has built the deepest operational understanding of how enterprises actually function – as a live, process-native model of how work happens, why it breaks and what should happen next.
“With the acquisition of Ikigai Labs, they’ve added the decision intelligence and simulation capabilities that make it truly effective.”
