The Context Model: The missing piece of the Enterprise AI tech stack

We have spent the last few years captivated by the potential of Enterprise AI. We’ve seen the demos, felt the excitement, and talked the talk. But for the C-suite, the period of experimentation is rapidly reaching its expiration date. Boards are no longer asking what AI can do; they are asking what AI has done for the bottom line. It’s time for AI to walk the walk.

Yet for Enterprise AI to transition from delivering simple productivity gains to autonomous decision-making and action that drives transformational business outcomes, it needs more than data. It needs to understand how your business actually runs, and how to improve it. Without this understanding, AI agents cannot make a real impact, so companies struggle to see meaningful returns on their Enterprise AI investments.

That’s why we’ve launched the Celonis Context Model to eliminate Enterprise AI’s operational blind spots. And, we’ve entered into a definitive agreement to acquire Ikigai Labs, which brings state-of-the-art enterprise Decision Intelligence and cutting-edge AI innovation to the Context Model.

Agents are only as smart as the context they have

AI models are brilliant reasoning engines, but they’re probabilistic. They predict based on patterns — and in the case of LLMs, those patterns are the ones they’ve learned from the public internet. They know generally what an invoice is and how a supply chain works. But when it comes to your business, they have major operational blind spots.

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.

Operational context is no longer a luxury; it is a fundamental requirement for Enterprise AI.

Context Model vs. Context Graph vs. World Model

Jaya Gupta and Ashu Garg from Foundation Capital called out context’s critical importance to AI in December, describing a “context graph.” This is a structured record of how decisions actually get made inside a specific enterprise, including the exceptions, overrides, and precedents that currently live in Slack threads and people’s heads.

At the same time, “world models” — simulators that understand the real world (either physical or social) — have gained significant attention within business and AI circles. These models build on the context graph concept by modeling the dynamics of a system. A world model learns how a business behaves under different conditions through prediction and simulation. It allows an AI system to estimate what is likely to happen next and evaluate alternative scenarios.

The context graph is grounded in recorded history. World models generalize beyond observed data to reason about possible futures. Each serves a distinct and complimentary role, but to move from describing how a business operates to actively supporting how it should operate, you need to combine them both into a “context model”.

The Celonis Context Model is the critical layer for Enterprise AI

The Context Model brings together the ground truth and memory provided by a context graph and the inference and foresight provided by a world model.

https://delivery-p141552-e1488202.adobeaemcloud.com/adobe/assets/urn:aaid:aem:c22a3c11-d849-46e7-9ab7-d23aa09ad70c/original/as/Celonis_Context_Model_Slide.png

It is the new heart of the Celonis Platform — an evolution of our Process Intelligence Graph. And with the acquisition of Ikigai Labs, we’re adding even more advanced simulation and prediction capabilities into the Context Model. So people and agents can understand not just how your business runs today, but now, how you can predict and decide how it runs tomorrow. The Context Model combines hindsight, insight, and foresight to give Enterprise AI the operational context it needs to succeed.

The Celonis Context Model gives Enterprise AI complete operational clarity by being:

• Process-centric: It uniquely understands the sequence of events—the steps, decisions, and exceptions—that make up how work gets done in your organization. It means your agents know what’s happening, how and why it happened, and what should happen next.
• System-agnostic: It allows your AI of choice to reason and act across your entire digital landscape instead of being trapped in functional or vendor-specific silos.
• Dynamic: It is continuously evolving, mining new processes as they happen and learning from how your people and agents interact in real-time, constantly evolving with your business.
• Open: It is open by design, allowing you to assimilate and share context across your entire ecosystem or integrate new technologies through APIs and MCPs, ensuring you’re never locked into a single vendor.
• Trusted: By providing AI with full operational clarity, Celonis ensures your agents reason correctly, decide sensibly, and act reliably—so you can trust AI in business-critical areas.

With the Context Model, organizations get the tangible benefits of:

• AI costs that your CFO can budget against: With the Context Model, agents no longer need to reason from scratch, anchoring every AI interaction to pre-structured business logic. By eliminating wasted tokens and silent retries, it established a predictable cost-per-outcome that allows CIOs to move from uncapped experimentation to a defensible line item.
• Context that compounds over time: The Context Model continuously learns from your agents and your people, so your context is constantly compounding as it is fed back into the model, making every new agent cheaper and more effective than the previous one. This is the only AI infrastructure on your balance sheet that gains value with every deployment.
• Context that remains even when your vendors change: By decoupling business logic from the underlying tech stack, the Context Model makes sure your AI outputs are deterministic, auditable and grounded in your operational reality. Its open-standard architecture means you can swap in models, data sources and more as the technologies evolve while your context endures.

Celonis is the trusted platform to industrialize Enterprise AI

With the Context Model at its heart, the Celonis Platform and our ecosystem provide end-to-end capabilities to analyze, design, and operate AI-driven processes and drive business transformation. The Platform enables customers to not just give AI the context it needs, but also to identify the best opportunities to deploy AI strategically, and to orchestrate agents, humans, and systems to work together.

We’ve partnered with the leaders in both the underlying data layer and the agentic execution layer to build this new context layer that bridges the two. The Platform brings data together from across the enterprise with zero-copy integrations to sources such as Microsoft Fabric, Databricks, AWS, and Snowflake, as well as pre-built connectors to systems of record like Oracle and other leading ERP and CRM platforms. We’ve also built deep integrations with the leading agentic platforms, including Microsoft Copilot and Agent365, Oracle OCI Enterprise AI, Amazon Bedrock, IBM watsonx Orchestrate, and Anthropic’s Claude Cowork. However our customers are building agents, the Context Model is accessible and consumable by them.

The context revolution is really an evolution for us at Celonis. We’ve been helping global industry leaders understand and reinvent their operations for the last 15 years. Now, we’re combining that deep expertise, unique technology, and innovation from Ikigai Labs and our partners to provide the missing piece of the trillion-dollar AI puzzle.

Learn more about the Celonis Context Model and Decision Intelligence at Celonis:Next on May 19 or at an upcoming Process Intelligence Day.