Companies have never invested more in customer experience. And yet, according to the 2025 Forrester CX Index, 25% of US brands saw their satisfaction score decline for the second consecutive year, compared to just 7% that improved. In Europe, signals are moving in the same direction. The Ipsos BVA Observatory of Customer Services 2025 recorded, for the first time, a decline in consumer trust in customer service — down 3 points in one year in France. This is not a temporary dip. It is a structural paradox: the more companies spend, the worse the results. This article unpacks the mechanisms behind this gap and identifies the concrete levers to close it.
CX spending figures are staggering. Global companies have poured tens of billions into omnichannel platforms, CRM tools, AI chatbots, voice analytics solutions, and NPS dashboards. By 2024, Gartner estimated that customer service and CX technology budgets were still growing at a sustained pace, driven by the promise of generative AI.
Yet the results are stubborn. According to Forrester, 2024 represented a historic low for customer experience quality. Not one or two industries in difficulty: a broad, cross-sector trend affecting retail, banking, insurance, telecoms, and public services alike. In 2025, the Forrester Global CX Index found that 21% of brands declined, 73% stagnated, and only 6% genuinely improved — across tens of thousands of brands analysed in 13 countries.
According to PwC, 59% of consumers consider human interaction essential to a positive experience. Technology has not absorbed the human pressure. In many cases, it has amplified it.
Understanding this paradox requires looking precisely at what happens between signing a purchase order for a CX solution and the moment a customer hangs up with a negative impression.
Forrester put it with surgical precision in its 2025 outlook: "too many companies invest in CX to improve what their customers say about their experiences, rather than to actually improve those experiences." The distinction is fundamental. NPS surveys are deployed, real-time dashboards built, verbatim analysis tools installed. Everything is measured. Little actually changes.
The CX budget becomes a reporting budget, not a transformation budget. Scores are presented in the boardroom, irritants are identified, action plans are approved — and six months later, the same irritants appear in the next report. The problem is not the measurement tool. The problem is the absence of a closed loop between measurement and operational action.
The intention was sound: relieve agents of repetitive tasks so they can focus on high-value interactions. In practice, many organisations inverted the logic. They automated the front lines of contact heavily — precisely where customers are most vulnerable, most frustrated, most in need of a fast and human response — and kept humans for complex cases that technology could not yet handle.
The result: the chatbot stalls, the customer grows frustrated, they pick up the phone, land in a queue, and speak to an overloaded agent who lacks context. According to PwC, 59% of consumers consider human interaction essential to a positive experience. That is not a nostalgic preference. It is an emotional reality that tools cannot replicate.
A CRM platform. A ticketing tool. A live chat solution. A quality monitoring system. A survey tool. Analytics dashboards. IVR scripts. Each tool was selected after a rigorous procurement process. None of them talks seamlessly to the others.
The customer, meanwhile, lives a single continuous experience. They expect the adviser who picks up to know they sent an email yesterday, browsed the FAQ twice, and signed a contract three years ago. In operational reality, that continuity is rare. Technology silos create relationship silos. And the customer, who cannot see your information systems, simply feels unknown — despite years of relationship.
The race to deploy generative AI in customer relations has produced real gains on specific, well-scoped use cases: agent writing assistance, call summarisation, intent detection. But it has also produced rushed deployments in contexts where the solution was not yet reliable enough to handle customers without human supervision.
According to Gartner, only 48% of digital initiatives meet or exceed their initial business targets. The failure rate of CX tech projects is not marginal. It is structural. And the first to pay the price are customers who encounter a hallucinating bot, an off-topic automated response, or a digital journey with no exit to a human agent.
The paradox has a direct price that ROI analyses consistently underestimate. When a customer has to call back three times to resolve an issue that should have been handled on the first contact, that is not just a degraded experience. It is an operational cost multiplied by three.
Industry benchmarks consistently document that a 1% improvement in first contact resolution (FCR) reduces operational costs by approximately 1% and increases satisfaction by a comparable margin. The implication is fundamental: operational efficiency and customer satisfaction do not oppose each other. They reinforce each other — provided technology serves resolution rather than avoidance.
The cost of customer disengagement is equally well documented. According to PwC, 32% of customers would walk away from a brand they love after just one bad experience. And 65% say a positive experience is more influential on their loyalty than great advertising. For a mid-sized business, the delta between these two realities represents millions in annual revenue riding on a single variable: the quality of customer service.
Faced with this paradox, the temptation is strong to look for the next technology that will solve what the previous one did not. That is exactly the thinking that perpetuates the cycle.
The real question is not "which tool should we invest in?" but rather: "at exactly which point in the customer journey is our current investment creating friction instead of value?"
That requires an analytical reading of the real customer journey — not the theoretical one drawn up in a meeting room. The touchpoints where customers drop off, the channels where they call back after an unresolved digital interaction, the moments where agents improvise for lack of context: this data exists in your systems. It is rarely aggregated in a way that produces a clear decision on budget allocation.
Forrester identified this as the top priority for CX leaders in 2025: directly linking CX improvements to business outcomes. This does not require a multi-year programme. It is a change in evaluation logic that can be applied to existing metrics, starting by cross-referencing first contact resolution rates, callback volumes following digital interactions, and transactional NPS by channel.
In the 2025 Forrester CX Index, only 10 "elite" brands out of 469 analysed across 13 countries maintained their top status. Names like Chewy, USAA, Navy Federal Credit Union, ING, H-E-B. Very different organisations in size and sector, but with one consistently documented trait: they did not spend more than the rest. They spent differently.
Three axes set them apart.
They treat resolution as a financial metric. Every contact not resolved on the first call has a calculated cost, embedded in board-level dashboards. FCR is not an operational KPI reserved for supervisors. It is a figure that reaches the executive committee alongside revenue data.
They maintained or reinforced the human element on high-emotional-stakes journeys. Complaints, disputes, cancellations, vulnerable customer situations: these are the moments where humans are not substitutable, where a skilled and empathetic adviser can reverse a situation and turn a dissatisfied customer into a loyal one. Automation is reserved for transactional journeys with low emotional complexity.
They simplified before they automated. Deploying a chatbot on a poorly designed customer journey produces a confusing chatbot. High-performing organisations first revised the journey logic, reduced unnecessary steps, clarified messaging, and then automated a process that was already working well when handled manually.
Customer service outsourcing has long been associated with cost reduction. That reading is now outdated, and the numbers confirm it. The best-performing BPO providers consistently deliver higher satisfaction rates than many in-house teams — precisely because their business model is structured around operational performance, not technology spend.
A specialised outsourcing partner brings three things that a CEO or CFO cannot easily build internally. First, a critical mass of sector experience: contact patterns, seasonal peaks, friction point typologies, scripts that work — all acquired and capitalised across hundreds of thousands of interactions. Second, the capacity to absorb volume without degrading quality, through an HR and technology architecture built for exactly that purpose. Third, a results-driven management model: the provider is evaluated on NPS, FCR, and resolution time — not on budget spent.
This is precisely the model that senior executives should examine when they find that internal CX technology investments are not producing expected results. Not as a shortcut, but as a structural response to a structural problem: managing customer relations as a profit centre, with clear indicators and accountability on outcomes.
For a detailed breakdown of how to make this decision, our guide on measuring outsourced contact centre performance through KPIs and SLAs covers the criteria by organisation size, maturity, and strategic context.
The CX budget paradox is not inevitable. It is the direct consequence of an investment logic that prioritised tooling over diagnosis, visibility over resolution, and technology over human capability.
For a CEO or CFO managing these budget envelopes, three questions deserve systematic consideration before any CX investment decision.
What is your first contact resolution rate, channel by channel? If you do not know this figure precisely, your next investment has no solid foundation. It is the starting point of any serious analysis.
What percentage of customers who engaged digitally called back within 24 hours? This rate shows exactly where digital fails to resolve — where invested money is generating additional cost rather than satisfaction.
Is your provider or team evaluated on customer outcomes or activity metrics? Calls handled, average handling time, answer rate: these are activity indicators. Transactional NPS, FCR, post-digital callback rate: these are outcome indicators. Accountability must rest on the latter.
To go further in reading CX performance indicators and understand how to combine them effectively, our article on NPS, CSAT, CES: which customer experience metrics to choose provides a complete analytical framework.
Why don't the companies that invest the most in CX necessarily get the best results?
Because investment is typically directed toward measuring scores rather than actually improving interactions. Forrester has documented this pattern: companies measure what customers say about the experience more than they transform the experience itself. Without a closed loop between measurement and operational action, budgets produce dashboards, not satisfaction.
What is the financial impact of a poor customer experience?
According to PwC, 32% of customers would leave a brand they love after a single bad experience. Conversely, a positive customer experience drives 65% of consumers toward greater loyalty. For a mid-sized business, the gap between these two outcomes represents millions in annual revenue sitting on the quality of customer service alone.
Can outsourcing customer service improve customer satisfaction?
Yes, provided the provider is evaluated on outcome indicators rather than activity metrics alone. The best BPO providers consistently deliver satisfaction levels above many in-house teams through their sector expertise, volume absorption capacity, and customer-performance-centred management model.
How do you distinguish a good CX investment from a bad one?
A good CX investment is tied to a specific, measurable customer outcome indicator defined before and after deployment: first contact resolution rate, transactional NPS on the relevant channel, post-digital callback rate. An investment made without a pre-defined outcome indicator is an unmanaged investment, regardless of its size.
Can artificial intelligence solve the CX investment paradox?
AI can contribute on well-defined, specific use cases. But deployed on a poorly designed customer journey, it automates a bad experience rather than creating a good one. The priority remains to simplify the journey before automating. The tool comes after the logic, not before.
The CX budget paradox is not a mystery. It is the predictable result of an investment logic that confused tooling with transformation. The organisations that improve do not spend more. They invest where it genuinely changes the experience the customer perceives: resolution on the first contact, context continuity across channels, human presence at the moments that matter.
If you want to assess your current customer relations setup and identify underexploited performance levers, the Armatis team works with executive teams on this kind of operational audit and recommendation approach.
Armatis is a European specialist in customer relations and business process outsourcing (BPO), operating across multiple continents with thousands of employees serving companies of all sizes and sectors. The company designs and manages end-to-end customer service operations: multichannel contact centres, complaints handling, technical support, back-office and digitised processes. Backed by integrated technology infrastructure and the ability to adapt to any sectoral and regulatory context, Armatis helps its clients combine operational performance, quality of experience and cost control, wherever they need it.
Contact our teams to discuss your challenges and find out how we can support you
Join the leaders who trust our multilingual and technological expertise.