Celonis: ‘There Can Be no Artificial Intelligence Without Process Intelligence’.

According to the software manufacturer, artificial intelligence must go hand in hand with process intelligence (PI) to understand how businesses work and how to make them work better.

Last week, Celonis’ Celosphere 2024 conference took place in Munich, where the process intelligence and process mining company showcased to more than 3,000 attendees the latest additions to its software platform, with Celonis AgentC as the key player, a set of AI agent tools, integrations and partnerships based on the Celonis Process Intelligence platform.

The excitement around artificial intelligence agents is undeniable. While the concept of agents collaborating to solve business problems autonomously is not new, Large Language Models (LLMs) that drive generative AI have brought this idea to life. These agents are expected to define the next big wave of AI innovation, revolutionising the way both society and business make decisions.

3,000 professionals gathered in Munich to learn what’s new at Celonis and engage in discussions about its technology.

But what exactly are AI agents and what are the barriers to their widespread adoption? And most importantly, how can organisations effectively leverage them in their operations? Manuel Haug, CTO of Celonis, explained:

What is an AI agent?

An AI agent is a software program that interacts with its environment, collects data and uses it to perform autonomous tasks that meet predefined goals.

Generative AI agents are entities capable of orchestrating complex workflows, coordinating activities between multiple agents, applying logic and evaluating responses. In business terms, this means that agents can automate processes, increasing productivity, reducing costs and improving the customer experience.

Challenges in implementing AI agents

Despite advances in technology, there are several challenges that businesses must overcome to effectively deploy AI agents:

1. Problem identification: AI agents work best when the problem is clearly defined and limited in scope. Enterprises must be able to identify the right problem to maximise the value of the agent.

2. Writing effective instructions: Although LLMs are improving, creating work instructions for agents is still a time-consuming and iterative process. The quality of instructions and configurations is key to agent performance.

3. Monitoring actions and decisions: Unlike other automation systems, AI agents act similarly to humans, making them susceptible to errors, inconsistent responses or even hallucinations. Therefore, it is crucial to implement proper governance to ensure accountability and continuous improvement of agents.

Process Intelligence overcomes these challenges

Process intelligence (PI) is fundamental to addressing these obstacles and ensuring that AI agents are scalable, reliable and useful. It provides specific context for business processes, enabling AI agents to understand and act upon internal business workflows and data.

Manuel Haug, CTO of Celonis, during his media briefing

While consumer AI relies on large amounts of publicly available data, enterprise AI lacks a ‘reference layer’ that connects internal data. This is where Process Intelligence comes in, providing an ‘intelligence graph’ that acts as a semantic layer, connecting internal enterprise data and providing valuable context. This allows AI agents to speak the specific ‘language’ of the business and make more informed decisions.

In practice, the relationship between IP and AI will generate new business models, new ways of reaching markets, as Eugenio Cassiano, Senior Vice President of Strategy and Innovation at Celonis, summarised during the conference.

How does Celonis work with AI agents?
Celonis, a pioneer in process intelligence, facilitates the creation of AI agents for companies through three key steps:

1. Discovery of relevant problems: Process intelligence identifies areas with potential for improvement within workflows.

2. Agent creation: Provides valuable insights into how problems are solved today, generating clear work instructions for AI agents.

3. Execution and monitoring: Once agents are up and running, process intelligence allows their decisions and actions to be monitored, overlaying them with traditional process steps to ensure proper governance.

The future of AI agents

The use of AI agents to maximise productivity, reduce costs and improve customer experience is on the rise. Companies that use process intelligence to power these agents will be better equipped to identify problems, build more effective agents at scale and monitor their performance, giving them a competitive advantage in the next wave of AI progress.

According to the company, this approach will not only revolutionise business processes, but will also enable organisations to act more efficiently and responsibly, ensuring continuous and sustainable improvement in their operational performance.

The Cosentino case

Spanish multinational Cosentino, a manufacturer of high-quality natural and manufactured surfaces for architecture and design applications, played an important role during the Celosphere 2024 congress. It was one of the first companies to develop an AI agent to significantly reduce the time needed to review and release credit blocks in the order management process, using data from existing business systems and digital process twins. This agent was able to optimise decision-making for decision-makers, who could accept or reject the agent’s recommendations with a single click.

During the event, we had the opportunity to chat with Rafael Domene, Global CIO of Cosentino, who was quick to point out the benefits of applying this AI agent to such processes: ‘Credit and order management managers are now able to process 5 times more requests per day without increasing the risks associated with these types of transactions’.

This fact, in a sector as globalised and competitive as the one in which Cosentino operates, becomes essential to ensure the continuity of the Spanish company’s operations and business.

Fernando Ranz, Vice President and Country Manager of Celonis for Iberia and LATAM, and Rafael Domene, Cosentino’s Global CIO

In their words, ‘The implementation of a Celonis-powered AI assistant for credit block management has been a game changer for our order management operations, streamlining our processes and leading to faster and more reliable results. This innovation underscores our ongoing commitment to driving business success by leveraging advanced technology, such as AI.’

Celonis’ CTO also made reference to this project with Cosentino: “Celonis” process intelligence provides the data and business context needed to make enterprise AI more effective. What makes it unique is that it spans multiple systems, provides complete visibility into past and present business processes and enriches them with deep business insights such as business rules, industry benchmarks or reference models. Forward-thinking companies like Cosentino are using Celonis AI technology to facilitate innovation and accelerate business transformation,’ he concluded.

Celonis and the potential of the Spanish market

Cosentino has become an example of what can be achieved through the application of AI agents such as those that can be developed through Celonis AgentC, but it is not the only one. We also had the chance to talk to Fernando Ranz, Vice President and Country Manager of Celonis for Iberia and LATAM, who told us how the company’s office in Madrid has grown to house not only the sales force but also part of the engineering and international support teams: ‘we are currently more than 500 people and we have about 90 open positions, which means that our workforce will continue to grow’.

He said: ‘There are many sectors with great potential for implementing process intelligence where a lot of processes are still carried out manually and our technology is able to reduce the usual bottlenecks through automation using intelligent agents.

Ranz explained that the manufacturing sector has always been the main focus for Celonis, a company that was born in Germany, a country where this sector is one of the most powerful worldwide: ‘BMW is one of our most important customers. It has allowed us to get to know the car production process very closely, so it is a very natural environment for us. However, we are in a phase of expansion into new sectors in which we still have a lot of experience. ‘

Cassiano explained: ‘Unlike traditional data exchange technologies, such as EDI (electronic data interchange), APIs or email, which are designed to share transactional data, Celonis Networks provides continuous, up-to-date process information. With this shared actionable truth, companies can actively collaborate to identify and solve problems in business-to-business processes, streamlining operations and minimising risk’.

In other words, Networks injects a new level of transparency between organisations so that every process can be optimised. In the case of product supply chains, where different organisations (suppliers, manufacturers, distributors and customers) come into play, process intelligence can provide a global view independently of each company. This optimises and automates the entire lifecycle, which can exponentially improve the efficiency of these operations.