Every customer service leader we speak to right now is hearing the same thing from the boardroom: "We need to do something with AI." And most of them are sitting there thinking: "Great. But what exactly?"
So they start Googling. They try ChatGPT for drafting email replies. Someone sets up a basic FAQ bot. An IT team starts scoping a multi-agent orchestration project. All of this happens in parallel, with no shared strategy and no clear picture of what the organisation is actually ready for.
We see this pattern constantly. And it rarely ends well.
When we audit customer service operations in Belgium, we find that most companies are still at level zero or one on our AI maturity model. They have the ambition to be at level five. That gap between ambition and reality is where projects go wrong.
Level five means autonomous AI agents handling complex interactions across channels, learning and improving on their own. Level zero means your agents are still copy-pasting between systems and manually typing up call summaries. You can't jump from one to the other. And trying to do so usually ends with a failed pilot, a frustrated team, and a lot of wasted budget.
The companies that succeed with AI take a different approach. They start by understanding where they actually are, and then they figure out the most realistic next step.
.png)
One of the things we consistently share with clients is a breakdown that comes from some of the large consultancy firms: roughly 70% of AI success comes from your people and processes. About 30% is technology.
That ratio surprises people. Most AI conversations start with the technology: which model, which platform, which vendor. But the organisations that get results start with the people. How will this change the way our people work? What processes need to be redesigned? How do you build trust in a new tool?
If you skip that part, you don't end up with a smart assistant. You end up with a chatbot that confidently gives wrong answers. We've all seen the headlines: delivery companies where the bot tells customers to take their business elsewhere, or worse. That's what happens when AI gets deployed without the right foundation.
This is something we feel strongly about, and it comes up at almost every client. AI initiatives in customer service are too often driven by the IT department. IT picks a tool, builds something, brings it to the business team, and the response is: "That's not useful to us."
We've seen it happen first-hand. An IT team at one of our clients built an AI agent and presented it to the business. The business looked at it and said: this doesn't solve any of our problems. Months of work, wasted. Not because the technology was bad, but because nobody had asked the business what they actually needed.
The business has to be the driving force. They know the customer pain points, the process bottlenecks, and the daily frustrations that AI could realistically address. IT plays a critical role in making it happen, but they shouldn't be deciding what to build. That's a recipe for solutions that look impressive in a demo but collect dust in production.
One thing we notice is that companies tend to think about AI in customer service as a binary choice: either you build a customer-facing chatbot, or you don't. In reality, there's a much broader range of use cases.
You can use AI to capture data better, by letting it transcribe and summarise calls so agents can focus on the conversation instead of typing notes. You can use it to assist agents, helping them find the right answer faster or suggesting next steps. You can improve reporting, for example by running sentiment analysis across every call instead of relying on a survey that 7% of customers fill in. You can use it to deflect routine inquiries through self-service. And you can use it to improve processes after the fact, analysing interactions to find patterns and optimisation opportunities.
.png)
Each of these use cases has different requirements in terms of data readiness, process maturity, and organisational buy-in. The point isn't to do all of them at once. The point is to pick the one that makes sense for where you are right now.
Here's one more thing we want people to think about. Most organisations approach AI in customer service purely through the lens of cost reduction. Fewer agents, shorter calls, lower cost per contact.
That's a valid goal, but it's only half the story. Think about who in your organisation has the most direct contact with your customers. It's your service team. They talk to customers when things go wrong, when emotions are high, and when trust is on the line. That's also when customers are most open to hearing about solutions.
If you equip your service team with the right tools and the right data, they become your best channel for retention and even upselling. Your cost center becomes a profit center. We've seen this happen at several clients, and the results are tangible.
If you're being told to "do something with AI" and you're not sure where to begin, that's actually a good place to be. It means you haven't committed to the wrong thing yet.
Start with an honest look at your current situation. Where is your data? How mature are your processes? What are your agents struggling with today? The answers will point you toward a first step that's realistic, achievable, and sets the foundation for everything that comes after.
Not sure where you stand? Let's have a conversation and figure out the next step together.
