[0:00] Welcome back to EtherLink AI Insights. I'm Alex, and today we're diving into a topic that's reshaping how businesses handle customer service across Europe, specifically in Den Hogg. We're talking about AI voice agents and how they're transforming the customer service landscape for enterprises right now in 2026. Great to be here, Alex. And timing is really important here. This isn't theoretical anymore. Den Hogg has become this epicenter for businesses looking to adopt AI voice agents, [0:31] partly because it's the administrative hub of the Netherlands. And partly because companies there are facing real pressures, rising support costs, customer expectations through the roof, and traditional call centers just aren't cutting it anymore. So let's set the scene. What exactly are we seeing in terms of customer behavior that's driving this shift? I imagine it's not just about companies wanting to be fancy with AI. Absolutely not. The data is actually pretty striking. 70% of Dutch consumers still prefer voice or phone contact [1:04] for complex issues. That's important because it tells us customers aren't abandoning voice. They're actually demanding better voice experiences. And here's where it gets interesting. AI call center automation adoption in Europe has jumped 45% year over year. Enterprises are realizing that voice agents can handle 60 to 80% of inbound calls without escalating to a human when they're properly trained. That's a huge range, 60 to 80%. What determines where a company lands within that spectrum? [1:35] Domain expertise, honestly. A voice agent trained on your specific workflows, your industry vocabulary, your customer pain points. That's what gets you toward 80%. A generic voice bot handling basic FAQs might only hit 60%. The agentic systems we're talking about, the sophisticated ones, they can understand context, recognize emotion, and make real-time decisions about when to escalate. That's the game changer. And what's the ROI looking like for companies making this leap? [2:07] The numbers are compelling. Organizations deploying AI voice agents are seeing cost savings of 40 to 50% per interaction compared to human agents while maintaining 85% or higher customer satisfaction. For a mid-sized Denhag company handling 5,000 inbound calls monthly, that's $50,000 to $75,000 in savings every month. But, and this is crucial, that's only if the system is actually intelligent, not just automated. [2:39] Right. I want to dig into that distinction because I think a lot of people can flate automation with intelligence. What's the actual difference between a traditional chatbot or IVR system and what you call agentic workflows? This is huge. Traditional systems are rule-based. They match keywords to responses. Customer says the word billing, route to billing department. Agentic workflows are fundamentally different. An agentic AI system understands intent and context [3:10] across a multi-turn conversation. It makes real-time decisions about escalation or resolution. Critically, it integrates with your back-end systems, CRM, billing, inventory, all on its own. And it learns and adapts from patterns over time. So it's not just following a script, it's actually thinking? In a meaningful way, yes. Think about a customer calling about a billing dispute. An agentic system doesn't just say transferring you to billing. [3:41] It can access the customer's account history, understand why they're upset, propose a solution, and escalate only if necessary. That's the difference between automation, theater, and actual value creation. Now, I know one of the big things hanging over businesses right now, especially in Den Hogg, is EU regulation. The EU AI Act just kicked in at the beginning of 2025. How does that fit into the picture here? It's actually become a procurement gate. 78% of European enterprises now require AI Act compliance [4:16] documentation before they'll even consider a vendor. That's not hype. That's coming from enterprise procurement teams. For Den Hogg organizations in banking, insurance, healthcare, and government, compliance is an optional. It's table stakes. And voice agents handling customer service? Where do they land on the risk spectrum under the EU AI Act? They fall into the high-risk category. That means transparency requirements are serious. You need to disclose that customers are interacting with AI, [4:47] document how the system makes decisions, and ensure there's human oversight in place. It sounds restrictive, but here's the reality. Companies that build compliance into their systems from day one have a massive competitive advantage. Their vendor locked into solutions that regulators won't scrutinize six months from now. So compliance actually becomes a moat. It protects the business? Exactly. A Den Hogg bank that deploys an EU AI Act compliant voice agent isn't just getting cost savings. [5:19] It's also buying peace of mind. No regulatory surprises. No emergency rip-and-replace. That's worth real money when you think about the cost of compliance failures or having to rebuild everything in six months. Let's talk about the practical realities for enterprises rolling this out. What does implementation look like? Is this something a mid-market company can actually pull off? Absolutely. Platforms like Etherbot are designed exactly for this. They integrate voice, text, and backend actions [5:50] into a unified, agentex system that's already EU AI Act compliant. You're not building from scratch. That said, success still depends on understanding your specific use case. What are the 10 most common reasons customers call you? What's your existing tech stack? How much training data do you have? Those questions matter more than the tool. And what about the human element? Are we replacing customer service teams entirely or is this more about augmentation? Smart companies aren't replacing teams. [6:21] They're redeploying them. You don't need someone answering what's my account balance for the 500th time today. But you do need people handling complex disputes, building relationships, and making judgment calls. AI voice agents handle the high volume repetitive stuff. Humans handle the high value nuanced stuff. That's the sustainable model. So in theory, this could mean job displacement, but in practice, companies are finding it leads to job transformation. In theory, sure. [6:53] But the companies seeing the best ROI are treating this as a reallocation opportunity, not a layoff vehicle. Your team gets to focus on problem solving and customer delight instead of transactional calls. That's actually more engaging work, and it tends to reduce turnover, too. Let's zoom out a bit. What's the broader competitive landscape looking like in Den Hogg right now for enterprises considering this move? It's accelerating. McKinsey's 2025 AI index suggests [7:24] that organizations deploying agentic systems, not just chatbots, will reduce operational friction by 30% to 40% while improving satisfaction. That's a significant enough gap that competitors who move first gain a real advantage. And in Den Hogg specifically, you've got this mix of multinational corporations, government agencies, and mid-market firms all competing for talent and customers, speed and efficiency matter. So the window for first mover advantage is closing? [7:55] Not closed yet, but narrowing fast. Companies still have time to evaluate solutions, run pilots, and scale throughout 2026. But if you're waiting until 2027 to think about this, you're probably already behind. The voice first shift is happening now. What should a Den Hogg enterprise prioritize when evaluating AI voice agent platforms? Like if someone's listening and thinking, OK, maybe we should look into this. What are the non-negotiables? Three things in order. [8:27] First, EU AI Act compliance. Non-negotiable, especially in regulated industries. Second, multilingual capability. Den Hogg's business environment is Dutch, English, German, French. Your AI agent needs to handle that seamlessly. Third, back end integration. If the system can't connect to your CRM or billing system and actually resolve things, it's just a fancy chatbot. And does the cost of deploying these systems tend to be prohibitive? [8:58] Or is it accessible? Depends on your current setup. But most vendors now offer models where you pay per interaction or per licensed agent. You can start small, piloting on a single department or call queue and scale based on results. Given the 40 to 50% savings on large call volumes, ROI typically happens within months, not years. That's a pretty compelling business case. Let me ask you this. What's the most common misconception you see enterprises have about AI voice agents? [9:31] That technology solves the problem on its own. It doesn't. You need organizational alignment. You need to define what success looks like. You need training for your team on how to work with these systems. I've seen companies deploy amazing AI infrastructure and see mediocre results, because nobody internally understood how to use it. The technology is the easy part. The organizational change is the hard part. So implementation is as much about people and process [10:02] as it is about the tool? Absolutely. And that's actually good news, because it means smaller companies can compete with larger ones if they approach it strategically. You don't need unlimited budget. You need clarity and discipline. All right, let's wrap up. If you're a business leader in Den Hog listening to this, what's the one thing you should do this week? Map your current call volume and categorize calls by complexity. How many are simple questions that an AI voice agent could handle? [10:33] How many are complex and require human judgment? That one exercise will tell you whether voice agents make sense for your business and where to start. It takes maybe a day of analysis and gives you everything you need to build a business case. That's actionable and smart. Sam, thanks for diving deep on this. For listeners wanting to explore more, the full article with all the data, case studies, and compliance details is available on etherlink.ai. We'll link it in the show notes as well. Until next time, this has been etherlinkai insights. [11:06] I'm Alex, and thanks for tuning in.