ConstraintCognitive automation2025-05-08
Klarna's CEO said the company would again hire humans for customer service, saying cost had been 'too predominant' in its AI-first approach and quality had suffered
Customer service representativeoccupation page →Event date / reported
2025-05-08 · reported 2025-05-09
Evidence stage
ConstraintFailure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.
Tasks this bears on
Answering repeat questions
The same fifty questions, asked in a thousand different ways.
Automating✓ Evidence-backed
Reviewing what the bot said
Auditing automated conversations for wrong answers, bad tone and promises the company cannot keep.
New task✓ Evidence-backed
Where this applies
Same company as the February 2024 deployment; the assistant kept handling roughly two-thirds of inquiries. Secondary source quoting a Bloomberg interview of 8 May 2025.
What this means
A constraint record on the same deployment: 14 months on, the CEO said cost had weighed too heavily and quality had suffered, and the company would hire people again — while the assistant kept roughly two-thirds of inquiries. It does not undo the capability shown in 2024; it says the deployment conditions include a quality floor a company can miss, and it is the clearest evidence on this site that reviewing what the bot said is real, paid work.
What it does not yet show
One executive's interview, reported second-hand, is not a market reversal, and 'hire humans again' came with no numbers, roles or timeline. The bot still handled most of the volume, so this is a correction of the mix, not a return to the pre-2024 model, and it says nothing about what other companies' quality metrics show.
What you can check
Ask your quality or operations lead whether anyone is currently scored on reviewing bot transcripts, and if so how many conversations a week; a 'nobody' answer tells you the oversight task has not been funded at your employer yet.
Does it change the assessment?
No. The impact index is never moved by a single event. What this record did: the 2 linked task judgements above now rest on evidence instead of inference.
Source
CX Dive (reporting a Bloomberg interview) · verified 2026-09-10 · Claude (VOLO agent) — source text fetched and cross-checked · interpreted 2026-09-10 · Claude (VOLO agent)