For the first time, AI gives contact centers the opportunity to lift customer experience and efficiency at once. So why do most AI projects still fail to deliver on both? Because AI alone is no basis for a real transformation. In this blog, our Simon-Kucher Elevate experts explain what separates the winners from the rest and how to get the contact center transformation right.
For most of its history, contact centers have lived under an iron law of trade-offs: you can streamline your customer service, or improve customer satisfaction, but rarely both. With the rise of AI, a new possibility has emerged: lifting customer experience and efficiency, while lowering operational costs. But only when done right. Done wrong, AI delivers none of the three and adds cost and complexity on top.
We called the central risk the Ferrari Paradox: buy the most powerful AI technology on the market, drop it into a thin-data environment, and it starves. But even the right car, perfectly fueled, wins no race alone. It needs to follow a clear vision and ambition – and it needs the team around it: the pit crew, the engineers, the telemetry, a race strategy. For a contact center, that car is the technology. The vision is the strategic ambition that defines what the contact center should achieve. And the team is the operating model: the processes, structures, governance, and competencies that make the technology work. Usually, it is the vision and the operating model that are forgotten first.
Why do so many AI transformation projects in customer service fail?
The pattern is consistent: an organization sets out to modernize its contact center, picks an AI use case, and bolts it onto the way things already work. Technology goes live, and little or nothing changes in the operating model. In one project, an organization deployed an AI solution that successfully generated new opportunities, but the downstream team had little incentive to act on them because existing targets and KPIs had not been adapted to the new way of working.
In another case we discovered that the contact center had run various AI features for six months without the team knowing they even existed!
In both examples, the technology was available and worked; however, not much changed, because the relevant components of the operating model around it were not redesigned to use it.
These are not failures of technology but failures of transformation. Typical causes include:
- The project is use-case or technology driven, but not guided by a clear strategy and target picture;
- Project is often led by IT, with no real involvement of the business partners;
- Impact is never measured against a baseline or target values;
- Objectives are not aligned with the incentive systems;
- Operating model and its components such as structure, governance, processes or skills have not been adapted and aligned;
- Second and third use cases are never planned before the first is built and tested.
Each solution is bolted on alone, and by the third the pieces no longer fit.
Too special for AI, or too trusting of it
Behind these failures sits a misjudgment about what AI can do in contact centers and customer service, running in two opposite directions:
- Too complex for AI: Leaders steeped in the call center era see only exceptions and edge cases, conclude AI can handle one narrow request, and draw the target picture far too small. They forget that modern AI learns over time, absorbing past conversations and taking on cases well beyond the first cautious pilot.
- Too much trust in AI: The opposite error pushes AI onto complex, deeply human customer interactions where it harms the client relationship, like the airline chatbot that cannot parse an accent, never reaches a person, and traps you in a loop. That is not efficiency; it is friction dressed up as innovation.
The key is to sit between the two. A simple address change or a bank account update belongs to automation; the high-value customer and the business-critical case need a person. Drawing that line is the actual job – and not a one-time exercise. As AI capabilities evolve, the line needs regular reassessment.
The right sequence: defining clear ambition first
The solution is not having two months of strategy workshops before anything ships. It is to settle a clear business ambition first, then move fast and adapt. Start with the question almost everyone skips: What is this project aiming for – cutting costs, lifting customer experience, both, or turning service into a sales channel? With that discussed, the rest follows in order:
- understand the data and system architecture you have;
- prioritize the use cases that combine impact with feasibility;
- pilot them, and adapt processes, skills, and operating models as you go.
Skip the first step – defining a clear ambition – and the rest falls apart. Without it, there is no basis for prioritizing use cases, measuring success, or aligning the operating model.
As a project example: In a quick customer service audit for a multi-utility provider whose service costs had jumped 47% in two years, we found that 43% of the most important cost-driving KPIs were not being measured at all. The company had no visibility into its own performance – and without that visibility, neither strategic steering nor AI implementation is possible.
The winning formula is simple: first ambition, then data and use cases, then the operating model. Never the tool first.
Building the team around the car
Here the racing image turns literal. Just picture AI as the engine of the car. Whether it wins depends on everything around it:
- the structures and processes it runs on;
- the skills and competencies of the people developing and using it;
- the governance and incentives that decide whether its output is used;
- the analytics that prove it has worked;
- the culture that lets a busy team adopt it.
Change the engine and leave the rest untouched, and you have a faster car that still loses the race. A successful contact center transformation reshapes contact center operations and the team itself: fewer standard-case handlers, more people who train and supervise the AI, contact center agents who work beside an AI-driven copilot rather than a script, effectively empowering agents to handle complex cases alongside technology. Human connection in contact management stays – separate or part of an AI-human combination – where personal decisions and judgment earn their place, on the complex, high-value or emotionally charged cases, while the system handles the rest.
What ROI can companies expect from contact center transformation?
The honest answer: it depends on where you start and how far the operating model moves with the technology. But the target ranges are becoming clear in 2026. One of the largest field studies of generative AI in customer support to date – several thousand agents at a Fortune 500 software firm – measured a 14% average productivity gain from an AI assistant. The more revealing number sits underneath though: novice agents improved by 34%, while the most experienced agents barely moved. The tool alone does not lift the team evenly – but how roles, routing, and training are redesigned around it decides who benefits in the end.
Looking at the potential across the whole function, industry analysts put the achievable productivity gain in customer care at 30 to 45%. Surveys among service leaders point in the same direction: AI is expected to resolve every second service case by 2027, up from roughly 30% in 2025, and deployments of AI agents are expected to cut average handling time by around 20%.
Our own Simon-Kucher projects confirm that the upside is real when the transformation is done in the right sequence. Rebuilding one client's service strategy and operating model end to end lifted the Net Promoter Score by 35 points and raised first-contact resolution to 80% – results driven less by any single tool than by skill-based roles, clean processes, and digital support working together.
Numbers like these are potential, not entitlement. The gap between a 14% gain and a 45% gain is not the technology – instead it’s the ambition, the data, and the team around the car.
A growth lever, handled with care
Tempting as it is to call the contact center a growth engine, be honest about what it is for. Its job is often not to sell but first to solve the customer’s problem: this is the moment of truth where customer satisfaction is won or lost. Growth sits on top of that, not in place of it: turn every interaction into a pitch, and you erode the very customer trust that makes the next sale possible.
Handled well instead, the contact center creates room, so once routine customer inquiries are resolved cleanly, the agent has the context to make an offer that the customer sees as helpful. That kind of customer engagement – relevant, timely, and grounded in a resolved issue – tends to generate more revenue per customer, but only once the foundation is right.
Same steps, different questions: DACH and the Middle East
The steps of a contact center transformation barely change across industries or regions; just the specific questions during the process do. In DACH, for example, the pressure is the labor shortage and cost discipline. In the Middle East the personal relationship stays central, yet digital expectations run extraordinarily high and customers are loud when something breaks. Communication channels have shifted too: the telephone, long declared dead, is back, now paired with AI into an effective human-plus-technology channel with a clear roadmap for the future.
Customers move across multiple digital channels – phone calls, emails, live chats or chatbots, social media, and customers expect personalized interactions and real-time responses on each; an omnichannel contact center is the baseline, not a differentiator. What stays constant is that contact reasons and customer journeys within an industry are alike, which is why cross-industry experience pays, with telecommunications and insurance running ahead.
A business project, not a technology project
Which is why we hold one conviction firmly at Simon-Kucher: contact center transformation is a business and technology project, not a technology project alone. Optimizing a contact center end-to-end means working not only on use cases but also governance, structures, processes, people, data, and the commercial lens in an aligned manner – which is why in such projects, we pair Simon-Kucher Elevate’s digital depth with our specific business and industry expertise. Line up the technology, data, and operating model behind a clear ambition and transformation plan, and AI stops being a prestige purchase and becomes what it promised.
The Ferrari in the garage is impressive. But the race is won by the whole team: the right car, the right fuel, the right strategy, and a crew that knows what to do when it matters.
Want to build the team around your contact center? Reach out to our Simon-Kucher experts Tamas Harangozo and Christopher Jeckl.
