Why Celonis' new Chief Customer Officer believes AI will drive top-line growth, not just cost savings

Process intelligence vendor Celonis has appointed Deutsche Bank and SAP veteran Dilipkumar Khandelwal as its new Chief Customer Officer, bringing with him a distinct perspective on enterprise AI strategy. While most companies focus on using artificial intelligence to reduce costs and improve operational efficiency, Khandelwal believes they're overlooking a more significant opportunity.

His thesis centers on revenue generation rather than cost reduction—using AI to create entirely new business models and income streams.

Khandelwal brings 25 years of experience across both sides of the enterprise software equation, building products at SAP and implementing them at Deutsche Bank. This dual perspective has shaped his understanding of why many enterprise AI projects fail to deliver transformative results. Speaking with diginomica, he explains: 

I've been on both sides of this journey, and it makes you realize that there are things which you can do a little bit more early in the products which can be much more impactful when you sell to customers. What you feel that you've built - you assume that customers consume it in this way - but when you come to reality, you see the complexity at the customer level.

It's that complexity - the messy, interconnected reality of how large organizations actually operate - that Khandelwal believes is holding back enterprise AI from reaching its potential.

The context problem in enterprise AI

Enterprise AI faces a fundamental challenge that consumer applications like ChatGPT don't encounter: the need for deep contextual understanding. While AI chatbots can handle straightforward tasks like writing emails or summarizing documents for individual users, enterprise environments present layers of complexity that generic AI struggles to navigate effectively.

Khandelwal uses a banking example that illustrates the problem:

Imagine when you run a mission-critical system in a banking world - I would like to be able to talk to a bot and say, 'Transfer 200 euros to Derek.' It looks so very easy, but it needs to understand who Derek is, the data. It needs to understand what Derek wants to do, from which account I'm talking about, to Derek in London, while I am in India.

That seemingly simple transaction actually requires the AI to understand customer relationships, navigate regulatory requirements, handle system integrations, and apply business rules that might span multiple departments and countries. Get any of that wrong, and you've got problems.

The challenge becomes more pronounced for large organizations operating legacy systems. Khandelwal experienced this complexity firsthand in his role as a global CIO:

When you operate as a global CIO, you land up with two or three challenges. One is you are operating in a world which is very complex - complex in terms of the systems, complex in terms of the products which you run, complex in terms of siloed organizations.

Process intelligence as the foundation

Celonis approaches this challenge by positioning process understanding as the foundation for effective enterprise AI. Rather than treating AI as a standalone solution, the company has evolved beyond traditional process mining into what it calls object-centric process mining, creating what it describes as a "living digital twin" of business operations.

The distinction is important for AI applications. Traditional process mining tracks a single object—such as a customer order—as it moves through an organization. However, this approach fails to capture how different processes interact with each other. Object-centric process mining maps these interdependencies, providing AI with the contextual information necessary to make informed recommendations that won't create unintended consequences elsewhere in the organization. Khandelwal explains:

If you're able to leverage data in the right way, if we are able to put some understanding of the data in the right place and some understanding of the process, of how it runs, you can create a lot of applications or usage on top to start creating value on the data. When you start creating value on the data, you start to create completely different business models.

The approach is already showing results for existing customers. Take Campari, which has used process intelligence to optimize operations across 35 legal entities in nearly 200 countries. The company has realized $5 million in value, boosted its touchless invoice rate from 50% to 80%, and automated 30% of incoming payment allocations.

Or consider Lufthansa Group, which runs Celonis across more than 50 use cases to understand passenger journeys from boarding to baggage handling. By mapping the entire process across multiple systems, the airline can spot problems before they happen rather than just reacting after passengers are already frustrated.

The bigger prize: turning operations into revenue

Those examples show clear operational wins, but Khandelwal's vision goes well beyond cost-cutting. He's talking about using process intelligence to build entirely new products and services on top of operational data - essentially turning your operations into a revenue center rather than just a cost center. He says: 

I look at it from both sides. As a process mining solution, it can be very good in terms of streamlining the processes, creating value and impacting the bottom line, but at the same time, also helping enterprises to not just look at the bottom line, but to look at the data in a completely different way. When you start leveraging and building products on top of it, you can easily impact the value generation and the top line.

Think about it this way: a logistics company that really understands its delivery processes could offer guaranteed delivery windows as a premium service. A manufacturer with deep supply chain visibility could provide real-time inventory optimization services to its suppliers. The possibilities start multiplying once you have that foundational understanding of how your business actually works.

This thinking aligns with Celonis' recent strategic moves, including launching Solution Suites that package connectors, process data, business context, and pre-built apps for specific departments. The company has also expanded its AgentC AI capabilities and introduced an Orchestration Engine for coordinating tasks across different systems. Khandelwal notes:

The changing business model is also about the penetration of AI in an enterprise. I'm telling you the penetration of AI in an enterprise will only happen when you understand what is happening at the back. 

The trust deficit in enterprise AI

Khandelwal's appointment as Chief Customer Officer isn't just about bringing in someone with solid enterprise credentials - it reflects Celonis' recognition that technology alone won't solve the enterprise AI puzzle. His focus on value realization addresses a persistent problem in enterprise software: the gap between what gets promised during sales and what customers actually experience after implementation. Khandelwal says: 

From my experience, the right way to create value for the customer is to just deliver what we promise in pre-sales. It's a simplistic way - I'm going to create this value for you. Once they implement it, they get the same value realization that you promised, which is the return on investment.

That emphasis on delivering promised value comes at a critical time. After years of AI hype and mixed results, enterprise buyers are becoming more skeptical. We are seeing an increasing amount of "AI fatigue" setting in among enterprise buyers who've been burned by implementations that didn't live up to expectations.  Khandelwal explains:

For me, the biggest thing is: how do you start creating value for the customer based on what we promise? That's why for me, trust is a very important factor in what is built, and that decides and shapes how the customer journey should look like.

What's refreshing about Khandelwal's perspective is that it represents a more mature understanding of what enterprise AI can and can't do. Rather than positioning AI as some kind of universal solution, he sees it as a tool that needs proper foundation and context to be effective.

When asked whether customers understand the complexity of enterprise AI implementation, Khandelwal admits: 

I think we need to keep on explaining this multiple times, and that's the narrative and the story.

My take

As enterprise buyers get smarter about AI procurement, they're increasingly looking for vendors who can demonstrate clear business value rather than tools that promise the world and then under deliver. Khandelwal's background spanning both software development and enterprise implementation gives him credibility on both sides of that equation.

His focus on top-line growth rather than just cost savings also signals where the process intelligence market is heading. While early process mining implementations were mostly about finding inefficiencies and cutting costs, the next phase involves using that understanding to create new value.  Khandelwal concludes:

As you become larger and larger, you work on different systems. You don't work in a single system - you work in complex systems. By that default nature, you require someone like Celonis, and I think that's why I feel very positive. Celonis is rightly positioned to take that market.

Whether he's right about that positioning remains to be seen, but his perspective on enterprise AI - grounded in real-world complexity rather than theoretical possibility - feels like a necessary corrective to some of the more breathless AI rhetoric we've been hearing lately.