We’re Seeing More Demand From The Midmarket: Camunda CTO

Camunda Chief Technology Officer Daniel Meyer discusses why more midmarket companies are looking harder at automating business processes as agentic AI adoption rises.

(Daniel Meyer, chief technology officer at Camunda, on stage at 2025 CamundaCon, New York City)

Business process automation and orchestration platforms are garnering more attention as organizations, including midsize, look to streamline and automate their workflows amid increasing agentic AI adoption.

Largely deployed among the largest end of enterprises, more midsized and smaller organizations are looking for solutions to implement their own agentic AI-based automation and as a way to keep up with competitors.

Camunda offers a BPMN (Business Process Model and Notation)-based workflow automation platform.

Daniel Meyer, Camunda’s chief technology offer, along with Anna Schroth, corporate communication manager at the company, spoke with MES Computing about business process automation and the midmarket and why it’s technology that goes robot hand-in-hand with agentic AI.

[Camunda] seems to be a very complex solution for the average midmarket business. Are you mostly going after the Fortune 500 or is there a fit for midmarket?

Meyer: Organizations with 5,000 employees and up, that’s historically our main focus.

Why is that? Because, if I reach a certain scale as an organization now, I’m having a certain process and complexity that mandates processes to be managed and to be streamlined.

We do get a lot of demand now more recently from more midmarket and smaller organizations, and that’s interesting, some of those are more tech forward.

And the challenge they all have is, how do we scale our customers or our revenue and all of that without a proportional scale in employee base?

Schroth: Just a bit of Camunda history, the company was founded because [the founders] saw the need for this orchestration within the automation space. They started as a process consulting company, and then they realized, ‘Oh, there is a lack of software.’ So we start building the software and that’s how they started developing Camunda.

When your customers come to you, what are they looking to automate?

Meyer: That’s a really interesting question because Camunda is a horizontal platform. You can automate everything with it, but then the question is, like, what specifically, right?

And it’s a spectrum of different things. In financial services, it is use cases like customer onboarding, account opening.

It is things like underwriting. It is payment fraud. So, all those use cases where I need to provide fast service to my customer, and I need to do that at a certain level of efficiency, because otherwise I cannot treat every customer as one.

And you see the same in other industries. For example, in telco, a big thing is about provisioning, order execution, order management, and those kinds of use cases. I get a new phone, I want to activate my SIM card. Let’s get that done within one minute, right? I need to orchestrate various steps across the telco infrastructure, and that’s something that, historically, customers are doing.

[For] smaller to midmarket organizations, a good place to start is customer support. We worked with a fairly small company in the UK, and they’re a utility. They provide water to commercial clients. And they have, obviously, customer support; there’s no water, or there’s an issue with the water, or billing, or all kinds of different problems. And they have automated that with Camunda.

Are customers building their own AI agents, or does Camunda do that for them?

Meyer: Yes. Both. With Camunda, I can directly build enterprise-grade agents, agents that don’t just answer questions, but that also actually take action.

I can build my agent directly in Camunda, but I can also orchestrate agents with Camunda, and that can be agents you have built with Camunda or third-party agents

Do you find that U.S. firms are more hesitant to adopt business process automation?

Meyer: No. We have, like, nine of the 10 largest banks in the U.S. and some of them are significant. Are they leaders as far as adopting this? Yeah. For example, probably the company that has scaled Camunda the most, to my knowledge, is Goldman Sachs, o They have hundreds of processes. They run large chunks of their business with Camunda, and they’ve basically been very early on in starting to adopt Camunda.

What are the security guardrails that you have for your customers as far as these agents?

Meyer: There’s a number of things here. We talked a bit about enterprise-grade agents, that’s needed. So basically, there’s this tension between autonomy and trust, because only if I give some level of autonomy to the agent, then I can actually automate things that I could not automate in the past ... But then the more autonomy I give to the agent, the more I need to be able to trust it to not do the wrong thing.

So how do we do that? Inside Camunda, there’s a number of different things. The first thing is, with Camunda, you can visually model and design the behavior of the agent ... What are some steps where I’m dictating the order of activities that the agent must follow, and where do I grant autonomy to for the agent to decide on its own which steps to perform in which order and what data to use?

A design of the agents’ behavior, and that being a visual model that we can all look at and understand and validate, that way, we can build a lot of trust. So that’s the first piece.

Then you put that into production, the agent is running, and because it’s managed inside Camunda, we get full visibility and transparency into what it is doing, and what have all my agents done in the past. They get full audit, lock, and full traceability. And you can see literally each step that it took: What was the data? What? What did it consider? What was the thought, the chain of thought, the reasoning that the LLM did, and you can see what kind of reasoning it did in which step. And that is super important, because that way I get visibility and transparency. You can then also set up real-time alerts where you say, okay, if this happens, then I want to get an alert so that you can maybe shut it down, something like that. But also you can then have fine, granular control into, let’s say it does the wrong thing in a certain case, then you can intervene and change the behavior.

We had the example of doing a travel booking for a customer. And let’s say it’s stuck somewhere, and it’s just doing the wrong thing. You can go into our control plane, it’s called Operate, where you are operating your agents. And then in there you can say, ‘This is the step it’s currently performing; it’s doing the wrong thing. I’m going to take it and I’m going to move it to another step in the process, and I’m going to correct its behavior.’ That is extremely important, because if you don’t have that way of intervening, then it won’t work.