Main content

Celonis and Emporix aim for end-to-end process automation with launch of new Orchestration Engine

Derek du Preez Profile picture for user ddpreez July 10, 2024
With Celonis’ Process Intelligence Graph providing the context and situational awareness, the new Orchestration Engine moves beyond simple task-based RPA towards end-to-end automation.

Cute robot holding megaphone on blue background © Lerbank-bbk22 -
(© Lerbank-bbk22 -

Celonis and Emporix have today jointly announced a new platform that aims to provide enterprises with the ability to automate processes end-to-end. Whilst many organizations already utilize robotic process automation (RPA) to automate simple tasks, such as calculating a price or creating a quote, this new Orchestration Engine aims to allow buyers to go further and fully automate the ‘sales’ or ‘purchasing’ processes across the enterprise. 

Essentially the Orchestration Engine sits above simpler processes or tasks that are often contained in silos and ‘conducts’ the movement and completion of work across the organization. For instance, rather than just automating the process of creating a quote in the sales process, the Orchestration Engine claims to be able to - using contextual awareness within the Celonis Process Intelligence Graph - automate the processes of creating a quote, sending it the customer, creating the order, shipping the goods, creating an invoice, sending the invoice to the customer and then receiving the money. The entire end-to-end sales process. 

The Orchestration Engine uses process insights - known as trigger events - from the Celonis Process Intelligence Platform to deliver near instant responses and initiate a sequence of layered actions across associated systems and tools. Using customized execution templates, it orchestrates actions to optimize performance against specific business objectives, which Celonis says allows it to continually learn and adjust through the monitoring of real process execution. 

There are a number of components to this that are worth exploring. We spoke with Leander Hartenauer, Head of Technology Partnerships at Celonis, who said that Celonis on its own provides an analytical capability for organizations, whereby it brings together process data, creates a common data model, and then abstracts that on the taxonomy level so that companies can compare what happens in SAP versus what happens in Oracle, for instance. 

Prior to now, Celonis has been used to help organizations prioritize actions in their source systems based on process intelligence. Companies could use business apps or action flows to take action on its process insights, but often these still required humans in the loop and human decision making. Hartenauer said: 

That is always still humans taking decisions and prioritizing. That works well when you have these operational teams in place, but they are often quite siloed. There’s someone that’s responsible for invoices, someone for returns, for orders, customer service, etc. If someone is off sick, this just doesn’t happen then, right? 

A lot of customers said that we’d made a great start, but wanted us to automate it further. 

Second to this, Hartenauer said, customers also recognized that Celonis’ Process Intelligence Graph, which moved the platform away from case centric to object centric process mining, and allowed for multi-dimensional process analysis, saw that there was opportunity for multi-dimensional action capability. In other words, with the Celonis platform understanding how different processes were interacting with each other, customers then wanted the ability to be able to automate actions based on that understanding: 

It’s not only about prioritizing orders in the order management domain - it’s about orchestrating multiple tasks across multiple departments to ensure that the process from start to finish, really end to end, is optimal. Optimal can mean human in the loop, or it can mean fully automated, that is the customer’s choice. 

[Previously] with our platform you could trigger single actions, or you could automate single tasks, or maybe multiple single tasks at the same time - but you could not say ‘I want the process to start here and I want to monitor the end to end process even if it takes three months’. 

Acting as a conductor 

The key thing to understand is that with RPA, the task automation happens in isolation and the contextual understanding of that task doesn’t carry over to the next task in the process. With the Celonis Process Intelligence Graph, there is a contextual understanding of each task or action along the full process. This understanding can then be used in conjunction with the Orchestration Engine to decide if certain actions should be taken and if work should progress to the next stage. As Hartenauer explained: 

In Celonis I always know how the process is going. With the Orchestration Engine I can course correct, it’s almost like having a conductor of an orchestra. I can say: ‘Okay, this part of the process runs in the normal SAP function, I expect this outcome. If this outcome happens, great, then let's go to the next one, if it doesn't happen, then I can course correct.’ 

Each step is stateful -  it has all the memory of what happened before. The Orchestration Engine always knows what happened before. It will know how you got to the ‘state’ you are in today, you have all the history. This lives within the workflow itself, which is why it can decide really quickly, it has all the context. 

The unique combination for us is the Orchestration Engine with our intelligence - because with the intelligence, the Orchestration Engine can always get access to the Celonis knowledge of what happened before, plus what it did itself, plus our business knowledge with Symbio and Celonis process management. Therefore, it really becomes dynamic. 

For instance, if an organization has a customer that wants to return a certain order, you may need to know: what’s the customer scoring? What's the customer's payment history? What are the goods in this order? Did the customer pay their last invoices? What is their credit scoring? As Hartenauer said:

All this information we have in Celonis, the Orchestration Engine has access to, to make a decision. 

And therefore, Celonis provides the context, the situational awareness, and then the Orchestration Engine can decide dynamically. I would say that deciding dynamically is something that is new. Because typically, if you look at workflow platforms, or an ERP system, or a CRM system, all these processes, they're always static. There are some elements of being dynamic, but not at this scale. 

The ambition here, Celonis hopes, is achieving an autonomous enterprise. The idea of ‘autonomous enterprise’ is something that has been spoken about for years and years - and some have questioned whether it is feasible. However, with advancements in AI technologies in recent months, the conversation is back on the table. An autonomous enterprise is thought of as an organization that applies AI and automation to essentially ‘self-drive’ its operations and services. Celonis says that it is an organization that can “detect events, anomalies and opportunities, and continually adjust processes, in real-time, to resolve those issues or seize those opportunities – maximizing performance against key metrics”. 

So with the introduction of the Celonis/Emporix Orchestration Engine, how far away are we from the autonomous enterprise being a reality? Hartenauer said:

I wouldn’t say we are far off. The question is, how far off are we in terms of the customer daring to go that far in terms of their digitization journey? 

My take

Whilst the idea of an ‘autonomous enterprise’ always grabs attention, I think the reality is that enterprises are less occupied with that and are more interested in introducing as much automation as possible to processes, whilst keeping people in the loop to ensure things are running smoothly. This feels like a big step for Celonis. The process intelligence capabilities it has delivered in recent years have time and time again shown to reap large efficiencies for customers - but those efficiencies have largely been down to customers taking out friction where they can and getting their systems to run more smoothly. With the introduction of the Process Intelligence Graph, buyers are now able to see how all the different parts of their organization are working together. The next rational step is to introduce action to this intelligence, to automate the actions across processes, rather than the individual tasks. It will be interesting to see this in practice and we look forward to speaking to customers as they start to make use of the new Orchestration Engine. 

A grey colored placeholder image