Throughout this series, we’ve challenged some of the assumptions that have shaped customer operations for decades.

• That BPO value should include more than just the people and capacity provided
• That operational KPIs don’t necessarily tell executives whether customer operations are creating business value
• That resolving the underlying problem matters more than simply processing the interaction efficiently
• That technology can change the relationship between workload, capacity and cost

Those ideas sound compelling in theory. The more important question is what happens when you put them into practice.

At Kärcher UK, we now have an answer.

In January 2026, Kärcher UK and Quantanite deployed QiQ across Kärcher’s UK customer service operation. Within six months, the operation improved across major service and customer metrics, all while managing a higher overall workload.

But the most important finding wasn’t a performance metric. It was discovering what was actually causing customer problems.

The agents weren’t the problem

Before QiQ, Kärcher’s customer service operation was already performing strongly against traditional quality measures.

Voice QA averaged 95% during 2025. Email QA averaged 98%.

By conventional measures, the frontline operation was performing well. But like most contact centres, quality assurance relied on manually reviewing a sample of customer interactions.

That could tell Kärcher how agents were performing within the conversations being reviewed. It could not provide a complete view of why customers were contacting the business, why issues remained unresolved or where friction was being created elsewhere in the customer journey.

QiQ changed the scope of analysis from a sample of interactions to every voice conversation. And what it found challenged a familiar assumption about contact centre performance:

Approximately 87% of unresolved customer interactions stemmed from process dependencies, not agent performance.

The issues concentrated on documentation requirements, warranty validation, repair workflows, and missing proactive updates on repairs, collections, refunds, and replacements.

Agents were consistently demonstrating empathy and professionalism. The friction customers experienced largely came from somewhere else.

That distinction matters because it changes where improvement effort goes. If we assume the problem is frontline performance, we respond with more coaching. If the evidence shows the problem is operational, coaching the agent cannot solve it. The process has to change.

Customer conversations became a map of operational friction

This scenario is where the idea of the contact centre as a source of business intelligence becomes tangible.

Rather than analysing each conversation primarily as an agent performance event, QiQ connected customer sentiment, outcomes and recurring themes to the operational context behind them.

That allowed Kärcher and Quantanite to identify where customers were actually creating effort.

The resulting actions extended beyond the contact centre. Document collection and verification were streamlined. Repair and warranty workflows were redesigned. Proactive customer updates were introduced. Customer journeys were changed so enquiries reached the right team first time. And coaching became more targeted, using the behaviours identified in actual best-practice interactions rather than relying solely on generic quality feedback.

The customer conversation became the starting point for improving the wider operation.

Resolution improved without chasing shorter calls

One argument in this series is that traditional efficiency metrics can mislead when treated as outcomes.

Kärcher’s handle-time data illustrates why. Average handle time remained broadly stable: 9:01 in 2025 versus 9:20 in the first half of 2026.

Viewed in isolation, that might not look like an efficiency improvement. But transfers fell from 5.6% to 3.7%,  a 33% reduction. Agents were increasingly resolving interactions they might previously have transferred elsewhere.

At the same time, the longest-handle-time calls shortened by 21%, the gap to the FCR target closed by 45%, and repeat contact improved by 9.7%.

The goal was not to make every conversation shorter. It was to make the interaction more effective.

That is an important distinction. Kärcher wasn’t optimising for shorter conversations, but rather for resolution. It was optimising for resolution.

Better experience and greater efficiency moved together

The wider results reinforce another argument from this series: improving customer experience and reducing operational friction do not have to be opposing objectives.

Comparing pre-QiQ performance with the first six months after deployment:

• Voice transfer rates fell 33%
• Call abandonment halved, from 13.4% to 6.7%
• During the March–April seasonal peak, abandonment fell from 20.7% to 9.9%
• Peak average speed to answer improved dramatically. In April 2025 it reached 25 minutes 37 seconds; the worst month during the comparable 2026 period was 8 minutes 30 seconds
• Email response times improved from a peak of up to seven days to a consistent three to five days
• The gap to Kärcher’s NPS target closed by 47%
• The FCR gap closed by 45%
• And voice QA increased from 95% to 97%, while email QA held at 98%

Efficiency did not improve at the expense of quality. Both moved in the right direction.

More workload did not mean worse performance

The peak-season data is particularly important in the context of Gavin Atkinson’s argument about scale.

Between the comparable 2025 and 2026 peak periods, voice contacts handled decreased slightly, from 27,375 to 26,069. But email demand increased dramatically: from 9,813 to 18,547  an increase of approximately 89%. Across both channels, total interactions handled increased from 37,188 to 44,616.

That’s a 20% increase in overall workload.

Yet during that higher workload, abandonment fell, answer times improved, transfers declined and quality remained strong.

This does not mean technology removed the need for people or that workload suddenly became irrelevant. It demonstrates something more useful. Greater workload does not necessarily have to translate into deteriorating service performance.

When organisations have better intelligence about what is happening inside the operation, where friction is occurring and what needs to change, they can create more capacity from the operating model they already have.

The data also showed what good looks like

QiQ didn’t only identify failure. Analysing interactions across the operation also surfaced best-practice conversations and quantified the differences between them. Interactions materially outperformed Kärcher’s targets: calls were completed 400 seconds under the AHT target; 40% above the FCR target was performed; and +25.8 NPS and +5 CSAT were achieved, with zero repeat contacts.

That creates a different foundation for coaching. Instead of defining quality only through a static scorecard or a small sample of manually reviewed calls, Kärcher can identify evidence of what successful interactions actually look like in its operation.

Those patterns can then become a replicable framework for developing agents against demonstrated outcomes.

And because QiQ continues to analyse interactions, the effects of changes can subsequently be measured against the wider operation.

From conversations to continuous operational intelligence

At the beginning of this series, the argument was strategic.

The BPO model needs to move from supplying resources to solving problems. Customer operations need to move beyond operational activity and toward business outcomes. And technology needs to create a more scalable operating model rather than simply automate isolated tasks.

QiQ provided a way to operationalise this approach. Kärcher provides evidence of what happens when it is applied.

Perhaps the clearest proof is the 87% finding.

Traditional QA could tell Kärcher how its agents performed. Operational intelligence revealed that the overwhelming majority of unresolved interactions weren’t primarily agent problems at all.

That changed the diagnosis. The diagnosis changed the action. And the action changed the results.

As David Napper, Supply Chain Director at Kärcher UK, puts it:

“Traditional QA showed us a sample of our customer conversations; QiQ shows us all of them. It proved that most repeat contacts were process problems, not people problems.”

That is the larger opportunity. Customer conversations do not have to remain historical records of issues the business has already handled.

Analysed continuously and connected to operational context, they can provide evidence of what the business should change next. At Kärcher, that shift is already measurable.

Read the full Kärcher UK case study to see the complete results and methodology.