Expert Voices: Michael Kirchberger on How Organisations Should Prepare for Agentic AI
Part 3 of our partner series ahead of DMEXCO 2026.
This September, Ibexa is heading to DMEXCO in Cologne, and we're bringing our partner network with us. This year, we're especially looking forward to the conversations around agentic marketing: how AI agents are starting to work directly within the marketing stack, and what that means for the teams running it.
To get a feel for how that shift looks from the front line, we asked our partners for their take. From building the foundation to thinking about the future of agentic marketing, this interview series takes us every step of the way.
In the third part of this series, we spoke to Michael Kirchberger from schoene neue kinder about how to prepare for the agentic era: who needs to own which decisions, why fragmented data quietly undermines every AI initiative, and the one habit that keeps projects from stalling before they've even started. Here's what he told us.
Where do you see the greatest potential for AI and agentic technology in business?
"We see the biggest potential in technical sales and after-sales. In both areas, highly skilled employees often spend their time manually pulling information together from several systems instead of doing the work they're actually trained for.
Take a machinery manufacturer we worked with. Sales engineers used to gather product data from the PIM, pricing from the ERP, and customer history from the CRM, all by hand, before they could put together a quote. With an AI agent now handling that groundwork, the sales engineer simply reviews and finalises the draft. Time spent per quote has dropped from around two hours to twenty minutes.
We see similar gains in service. At a packaging equipment manufacturer, a technician can now look up an asset's ID and get immediate access to its fault history, the right spare parts, and the relevant maintenance documentation. Previously, that meant searching across several systems and scattered PDF archives.
These use cases work well when the underlying process is clearly defined and the data is reliable. Without that foundation, AI adds very little."
What makes a Digital Experience Platform modern today?
"A modern Digital Experience Platform has to do far more than manage content. It needs to bring together content, product information, customer data, and digital processes across a company's existing systems.
In B2B especially, product data lives in the PIM, prices and availability sit in the ERP, and customer data comes from the CRM. A good platform pulls all of that together and makes it available wherever sales, marketing, service, and customers need it. That foundation is what makes personalised offers, consistent customer experiences, and meaningful AI applications possible in the first place.
At the same time, the platform has to stay manageable in everyday use. Marketing teams should be able to run content and campaigns independently across different countries. Product information has to stay consistent across every channel. Customers need to find the content, products, and services relevant to them quickly.
For us, a platform is modern when it fits a company's roadmap, meaning its future goals and requirements, not just its current ones."
When do customers realise their existing solution has hit its limits?
"Usually, it's the accumulation of small problems that eventually becomes unworkable. A landing page needs an IT ticket. A price change has to be updated in several systems. A new language version turns into its own project.
That's often how isolated workarounds appear. Sales teams build their own spreadsheets because they don't trust the data in the system. Marketing builds campaign pages outside the actual platform because it's faster.
Lately, we're seeing the same limits surface in AI projects. One machinery manufacturer wanted to extend its existing platform with an AI-powered service portal. Once we looked into it, we found the product documentation spread across four different repositories, part of it only available as scanned PDFs. The spare parts data existed solely in the ERP and couldn't be accessed through any interface.
The real project, in that case, wasn't the service portal at all. It was consolidating, structuring, and connecting the information first. We see this pattern often: a new digital initiative becomes the moment a company discovers how much of its existing system landscape can no longer keep up."
What mistakes most often delay digitalisation projects?
"In our experience, it's missing decisions more than anything else. Plenty of people are involved, but no one has the authority to prioritise decisively. As a result, approvals and directional decisions that could easily be made in a single workshop end up taking weeks.
That's why, before a project starts, we clarify which decisions will need to be made along the way, who's authorised to make them, and when additional stakeholders need to be brought in.
That alignment can feel like extra effort at the start. But it saves considerable time later and stops important decisions from being repeatedly postponed."
If you could give a company one piece of advice before starting a digital project, what would it be?
"Write the initiative down on one page before you talk to any vendors.
That page should cover who you're solving the problem for, what business outcome you expect, what should change in the process, and how you'll measure success. The key is to describe the problem, not a predetermined solution.
The moment vendors enter the conversation, it naturally shifts toward their particular solution. Suddenly you're discussing technologies, platforms, features, or delivery methods before the actual task has been properly defined. That applies just as much to software vendors as to agencies.
If you've clearly described the initiative beforehand, you can measure every proposal against it. Without that, you're more likely to end up choosing whoever gave the best presentation."
