Why People-Centred AI Is Your Only ROI Strategy

TL;DR: Contact centres pursuing AI-driven headcount reduction consistently fail to achieve ROI, whilst those investing in people-centred AI implementations see measurable returns. Klarna's failed automation strategy (which led to staff rehire) demonstrates that replacing experienced agents erodes institutional knowledge and customer satisfaction. Research shows successful AI implementations augment human agents rather than replace them, transforming cost centres into value centres and delivering 3.5x to 8x returns on investment.

[Research includes references from McKinsey & Company, Indeed.com, IDC, Gartner, Accenture and others]

Why People-Centred AI Delivers ROI That Replacement Strategies Never Will

• AI implementations that augment human agents deliver 3.5x to 8x returns, increase issue resolution by 14%, and reduce handle time by 9%.
• Replacement strategies destroy institutional knowledge and customer relationships, creating costly attrition-recruitment cycles.
• Agent fear and resistance kill more AI projects than technical problems, with only 25% of call centres successfully integrating AI.
• Training staff to work alongside AI transforms cost centres into value centres, creating new tactical business departments.
• One example of a company that built human capital first saw projected revenue increases of $400M to $700M and 25% productivity gains.

We hear leaders tell the same story repeatedly. They implement AI. The business case promises savings. Six months later, they face agent turnover, customer complaints, and missing key deliverables and performance.

In this research, several sources were explored to consider:

• What happens when AI replacement backfires
• Why business cases fail
• What the ROI data shows
• How to turn cost centres into value centres
• The human factor in failure
• The training and preparation gap
• Testing assumptions the right way
• The patterns of success
• Frequently asked questions, answered
• Key takeaways

What Happened When Klarna Replaced 700 Agents With AI?

Klarna thought they'd solved customer service. In 2023, they replaced roughly 700 customer service employees with an AI assistant. The chatbot handled two-thirds of all queries. The CEO called it a major efficiency win.

Then customer satisfaction dropped. Operational problems started surfacing. Quality suffered because, as the CEO admitted, "Cost unfortunately seems to have been a too predominant evaluation factor."

By mid-2025, Klarna was rehiring human agents. They had to rebuild the capacity they'd just dismantled.

Leaders see Klarna as a warning. Yet many companies repeat the same mistake.

Bottom line: Klarna saved money in the short term but spent more rebuilding what they had destroyed. Cutting staff without preserving institutional knowledge creates expensive problems.

Why Do Business Cases Fail in Real Implementation?

The disconnect starts with how we build business cases.

Most AI business cases are paper exercises. They project savings based on headcount reduction. They calculate efficiency gains in a vacuum. They assume everything works exactly as planned.

Amara's Law applies here. Named after American scientist Roy Amara, it states that we overestimate what technology does in the short run and underestimate what it does in the long run.

The business case shows immediate cost savings from cutting staff. What the case doesn't show is what happens three years later when you need new skills but lost the people who carried your institutional knowledge.

You end up in a cycle: massive attrition now, massive recruitment later. Save money upfront, spend more to fix what you broke.

The core issue: Paper-based business cases overestimate short-term gains and ignore long-term consequences. This creates an attrition-recruitment cycle that costs more than it saves.

What Does the Research Say About AI ROI?

The data paints a different picture than vendor pitches.

Organisations report $3.50 return for every $1 invested in AI. Top performers achieve up to 8x returns. The key: successful implementations use AI to augment human agents, not replace them.

One company with 5,000 customer service agents implemented AI properly. They increased issue resolution by 14% and reduced handle time by 9%. McKinsey estimated that this approach increased productivity by 30% to 40% on current function costs.

The numbers worsen when you look at customer experience. Nearly one in five consumers who used AI for customer service saw no benefits. A failure rate almost four times higher than AI use in general.

The evidence: Augmentation delivers 3.50-8x returns because customers respond more positively to empowered agents.

How Do You Transform Cost Centres Into Value Centres?

Leaders who achieve real ROI from an AI approach approach their work differently. They're creating value centres.

This post is for subscribers only

Become a member to get access to all content

Subscribe now