Conversational AI &
Customer Experience
Summit 2025
4th Annual Edition — From Pilots to Production. Enterprise-Grade Conversational AI in Europe.
A defining year for enterprise conversational AI
The 4th Annual Conversational AI & Customer Experience Summit (CACES Germany 2025) took place on 5–6 November 2025 at the Leonardo Royal Hotel, Munich. Organised by Altrusia Global, the two-day event brought together over 350 delegates from 250+ organisations across 14+ industries — the most-attended edition to date. Twenty-nine keynotes, two panel discussions, and dense networking sessions traced the industry's decisive move from experimentation to enterprise-grade deployment.
Key theme of 2025
The industry has moved decisively from piloting AI to scaling it. Speakers consistently distinguished between proof-of-concept and production — covering governance, hallucination control, data sovereignty, EU AI Act compliance, and the human skills required to manage hybrid AI–human teams.
- Governance & ComplianceEmerged as the single biggest enterprise concern, with multiple sessions directly addressing the EU AI Act (OTP Bank, Booking.com, Allianz).
- Voice AI Centre-StageElevenLabs, Telnyx, Hyundai, NIO and BMW joined forces in Panel 1 to map the road from chatbots to in-car AI co-pilots.
- Agentic AIDominated the product track — Flix, aiaibot and Telefónica shared real deployment case studies with measurable CX improvements.
- Human-Centred DesignA recurring thread, from "Invisible AI" (ON) to "Designing for Doubt" (Convocat) to Dr. Ali Fenwick's psychological critique of current AI models.
- Data DemocratisationStrong interest — IKEA's GenAI CX assistant serving 220,000 co-workers and Norlys' chat-first BI interface were standout case studies.
Who was in the room
A high-quality practitioner audience, weighted toward decision-makers and senior implementers, spanning regulated industries, technology vendors and research.
Seniority distribution
Headline figures
Industries represented
| Industry | Notable organisations |
|---|---|
| Automotive & Mobility | BMW Group, Hyundai, NIO, CARIAD SE, AutoScout24 |
| Banking & Financial Services | OTP Bank, UniCredit, Allianz, DekaBank, Raiffeisen Bank |
| Retail & E-Commerce | Zalando, LEGO Group, Otto, Zooplus, C&A, Mytheresa, PUMA, BIRKENSTOCK |
| Telecommunications | Telefónica, Sinch, T-Systems, Telnyx |
| Technology & Software | Siemens, Bosch, AMD, Sprinklr, Prodyna, Crealog, aiaibot |
| Healthcare & Life Sciences | Dentsply Sirona, Straumann Group, Sandoz, Merck |
| Insurance | Helvetia Versicherungen, AXA Schweiz, Esure Group |
| Travel & Logistics | Booking.com, Delivery Hero, Flix, Havi |
| Energy & Manufacturing | OMV, Sika, Emerson, OroraTech |
| Government & Public | City of The Hague, Federal Authority UAE |
| Media & Entertainment | RTL Deutschland, SoundCloud, Red Bull, Coinpedia |
| Consultancy | Allianz Consulting, Sprint Reply, Horváth & Partner, NTT DATA |
| Start-ups & Scale-ups | Ideabay AI, AddAI.Life, Nebuly AI, NextLevel.AI, Pluno.ai |
| FMCG & Consumer Goods | Philip Morris International, Haleon, Lacoste Germany |
Eight themes that defined CACES 2025
Across both days, eight recurring threads emerged — each reflecting where enterprise conversational AI is moving from theory to operational practice.
From Pilots to Production
Enterprise practitioners from Booking.com, Telefónica, aiaibot, and Flix shared frameworks for taking conversational AI from proof-of-concept to production at scale — covering evaluation pipelines, monitoring, and continuous improvement.
The EU AI Act Is Here
OTP Bank, Allianz, and UserTesting outlined practical governance approaches: risk classification, human-in-the-loop mandates, audit trails, and navigating "grey zones" in high-risk customer-facing applications.
Voice AI & Emotional Intelligence
ElevenLabs, Telnyx and the automotive panel (Hyundai, NIO, BMW) positioned voice as the natural interface for the next decade. Latency, real-time edge compute, and emotional tone matching emerged as the key differentiators.
Data Sovereignty & MCP
CreaLog's keynote on the Model Context Protocol resonated strongly: standardised, LLM-independent pipelines allow enterprises to maintain data control while accessing the latest AI capabilities — a governance imperative, not just a technical choice.
Psychology, Trust & Invisible UX
Dr. Ali Fenwick (Hult) argued AI needs psychological frameworks before more code. ON's Kris Hedstrom and Convocat's Maaike Groenewege independently made the case that the best AI UX is one users barely notice — trust is built through restraint, not feature density.
AI-Native Telcos
Panel 2 explored what it means to redesign telecom customer journeys around LLMs rather than simply adding AI to legacy IVR. Norlys, o2 Telefónica, and CreaLog shared distinct maturity models, agreeing that responsible personalisation at scale remains the unsolved frontier.
Agentic AI in Customer Operations
Multiple sessions moved beyond chatbots to autonomous agents capable of multi-step reasoning, tool use, and proactive outreach. NextLevel AI, Flix, and aiaibot presented concrete ROI data across healthcare, insurance, e-commerce, and travel.
Democratising Data via Chat
IKEA's CX AI assistant for 220,000 employees and Norlys' single conversational BI interface both challenged the dashboard paradigm. Semantic layers and consolidated metrics were presented as necessary infrastructure for reliable, explainable conversational analytics.
The sessions delegates rated highest
Based on the post-event feedback survey, the following speakers were repeatedly named as favourites or "most insightful".
| Speaker | Organisation | Session |
|---|---|---|
| Prof. Michael McTear | Ulster University | Beyond Prompting: GenAI Competence Framework |
| Maaike Groenewege | Convocat | Designing for Doubt in the GenAI Age |
| Tim Stierman | IKEA | Unlock GenAI Data Analytics Power |
| Dr. Ali Fenwick | Hult Business School | AI's Blind Spots: Needs a Psychologist First |
| Raquel Serrano | Booking.com | Human + AI Collaboration in CX at Scale |
| Fabian Kortekaas | City of The Hague | Digitalizing the Citizen Experience |
| Kris Hedstrom | ON Running | Designing AI UX that Builds Trust |
| Anna Bredow & Vasily R. | Flix | Transforming CX with Agentic AI |
| Jörg Schoba | OMV | Self-Service AI Platform Approach |
| Daniel Arnst & Tina Rahman | Telefónica | Lessons from Scaling Aura Bot |
| Yanek Brinkmann | ElevenLabs | Voice AI, Emotion & Context |
| Christoph Esslinger | Ideabay AI | Opening & Closing (both days) |
What delegates told us
In delegates' own words
"Maaike made the presentation dynamic, very appealing and emotional — she talked about the pros AND cons of AI. It's important we are also aware of AI's impact on people's lives, not only businesses."
— On Maaike Groenewege, Convocat"Jörg Schoba — practical, down-to-earth showcase on implementing AI at a diverse organisation, with practical examples and both business and technical insights."
— On Jörg Schoba, OMV"Maaike's was such a good wrap up to everything. I left with tons of food for thought."
— On most insightful session"Raquel Serrano at Booking — very insightful, with real examples and numbers about its impact. Very interesting and well done."
— On Raquel Serrano, Booking.com"Besides Adonis, I enjoyed the keynotes from Prof. Michael McTear (for its depth and detail) and Anna and Vasily from Flix (for their excellent delivery and insights)."
— On most insightful session"Deploying GenAI in Practice — very open about the persisting difficulties encountered in implementing customer-facing AI solutions."
— On Telefónica's Aura Bot sessionThe organisations that powered CACES 2025
CACES Germany 2025 was made possible by a coalition of platform providers, AI vendors, and consultancy partners spanning enterprise voice, agent infrastructure, QA, and design.
CACES Germany 2026
Konrad-Zuse-Platz 14
81829 München, Germany