At the end of February, I hopped on a train to Berlin to spend three days at CCW2025. This event, held annually, brings together industry leaders, market players, and technology service providers related to all things customer experience and customer interactions.
Catching up with our partner
First stop on my agenda was meeting up with Creative Virtual partner, SOGEDES, at their stand to catch up in person with the team, as well as to see live their latest innovations in customer experience solutions. On show were their AI-powered customer interaction solutions which included chatbots, voicebots and omnichannel communication, as well as automation for contact centres.
All things AI, contact centres and customer experience
Next up was a walk around the exhibition halls to see technology in action and new customer experience solutions, listen to the keynote speakers talking all things trends, challenges and innovations and network with industry leaders as well as those at the coal face of building and deploying customer experience solutions.
What caught my eye and ear?
To whittle down what caught my eye and ear over the three days to a few points, was not easy given that over 200 exhibitors and over 50 speakers and discussion forums were part of this year’s CCW2025. But whittle down I did.
This year’s buzz was…not about the buzz but getting things to work. AI is a commodity now – for vendors and customers. Hence the talk of the town was:
- AI needs the Human in the loop.
- “Just throw AI at it” doesn’t work. There’s Do’s and Don’ts in AI projects.
- The EU AI Act and its implications.
Human in the loop
There will forever be a need for human control when it comes to AI systems. Human oversight integrated into AI systems is the compass to principles and ethics, which ensures accuracy, compliance and high-quality interactions between brands and customers.
Discussions about human in the loop, humans in control, human oversight, and human intervention – and many other phrases used to say the same thing – all boil down to the same goal. That is, ensuring better accuracy, greater trust and customer experience excellence. It is about combining the benefits of AI with human expertise to ensure personalisation, quality, and control of all customer interactions.
How to run an AI project
Often overlooked or ignored, but rightly spelt out by industry leaders, were the key requirements to a successful AI project which included:
- A clearly defined goal with success measurement
- Assembling the right team – skills, experience, ability
- Understanding and focus on content/knowledge management – data quality criticality
- Choice of AI model – LLMs (choice per use case), rule-based AI, hybrid approach)
- Human oversight embedded
- Integration and interoperability across a company’s business and operational support systems
- Continuous monitoring, updating, and adoption of innovative technologies and features
EU AI Act
The Act was approved just after last year’s (2024) CCW and so it was not surprising this was a topic of interest at this year’s event. Some rules are already applicable with full enforcement to take effect during 2026-2027.
The Act is forcing discussions on how companies can adapt to ensure compliance given its impact on AI deployments in customer service, contact centres and automation projects.
I have mentioned previously that trust, compliance, and responsible AI are topics that permeate every discussion on conversational AI, so it is no wonder, and right, that the EU AI Act was top of mind during CCW2025.
What I expected to hear about, but did not, or only very little.
Maybe talks and booths were planned before the following topics became commonplace, or maybe they aren’t above the radar yet. You can take them as my predictions of what is going to be hot at CCW 2026.
Agentic AI
It was a surprise that Agentic AI did not feature more noticeably at this year’s CCW, given that it is a rapidly growing topic within the context of AI-centric customer interactions:
- AI customer support agents where the entire interaction is autonomously managed
- AI in the contact centre which would see AI deployed to predict customer needs without input and enable dynamic decision-making that adjusts real time based on context
- AI in workflow automation where AI agents can manage complex customer service workflows autonomously and AI is used to streamline process tasks such as ticket routing, resolution paths, and agent scheduling
Neuro-symbolic AI
Ok, I wasn’t really expecting this to be a prominent topic. But it struck me how AI was identified (again) with only one technology, namely Large Language Models.
Neuro-symbolic AI, combining symbolic representation and their explicit rules and relationships with the number crunching power of large neural networks will greatly improve explainability, transparency and context awareness of AI solutions. It’s thinking fast AND slow, for those who know their Kahneman ☺. I guess, it will be the thing on the CCW 2026 – or maybe 2027.
While we wait for CCW 2026
Those topics are also the focus for Creative Virtual during the coming year. We will be bringing more innovative solutions to market, ensuring our customers have access to and can use the latest technologies to successfully deploy conversational AI solutions that deliver better customer experiences and enjoy growth and success.
Watch this space for more on:
- Agentic AI
- Human in the loop AI
- How to run an AI project
- EU AI Act
- Neuro-symbolic AI
