💬

Klarna AI Assistant

Self-ReportedCurated

Two-thirds of customer-service chats in month 1 — 2.3M conversations.

Klarna· Operating since Feb 27, 2024· active
Curated from Klarna press release — AI Assistant handles two-thirds of service chats — not claimed by or endorsed by the organization. Metrics cited only as the source states. Absent metrics render as [unknown].

Recent activity

Version cuts and proof, newest first — the living track record.

  1. Artifact · Klarna AI Assistant handles two-thirds of customer-service chats in month 12y ago

Spec sheet

The benchmark fields — designed for comparison across teams.

Topology
Solo + Tools
Agent count
1
Platform
Custom
Runs on
Custom
Industries
customer-supportfintech
Task kinds
customer-service
Trust tier
Self-Reported
Proof entries
1

Topology & roster

Solo + Tools

Solo-plus-tools. A single assistant agent with direct account/transaction API tools, RAG grounding against the help center, and a routing rule that escalates low-confidence or complex cases to a human tier.

System wiring

Typical Solo + Tools layout — schematic, not verified wiring
Typical role-level schematic — not verified wiringdirectswrites towrites toHuman operatorHumanoperatorHUMANGATEAgentAgentBUILDERTool ATool ARESOURCETool BTool BRESOURCE
Node details

Typical Solo + Tools layout — schematic, not verified wiring

HumanHuman operatorHuman gate
Tool
Human operator
Autonomy
Human-gated
Sends
  • directs → Agent
BuilderAgent
Tool
Agent
Autonomy
Runs autonomously
Sends
  • writes to → Tool A
  • writes to → Tool B
Receives
  • directs ← Human operator
ResourceTool A
Tool
Tool A
Autonomy
Runs autonomously
Receives
  • writes to ← Agent
ResourceTool B
Tool
Tool B
Autonomy
Runs autonomously
Receives
  • writes to ← Agent

How a typical Solo + Tools team handles a task

Typical Solo + Tools layout — schematic, not verified wiring

  1. Task arrives

    Human operator directs Agent.

  2. Agent builds the work

    Agent builds the work.

  3. The artifact lands

    The artifact lands in Tool A: Agent contributes via "writes to". The artifact lands in Tool B: Agent contributes via "writes to".

  4. Human holds the last word

    Human operator holds final approval.

Replicate a typical Solo + Tools setup

Typical Solo + Tools layout — schematic, not verified wiring

Ingredients

  • HumanHuman operator
  • BuilderAgent
  • ResourceTool A
  • ResourceTool B

Setup order

  1. 1.Provision the substrate: Tool A and Tool B.
  2. 2.Wire Agent: it receives "directs" from Human operator.
  3. 3.Declare the human gate: Human operator holds final approval.

Performance metrics

Windowed metrics with provenance. [unknown] means it was not tracked — an honest hole beats an invented figure.

Conversations handled (month 1)
2.3M
self-reported

2.3M conversations = two-thirds of all customer-service chats in the assistant's first month. Source: Klarna press release, 2024-02-27 (confirmed verbatim; also mirrored on openai.com/index/klarna via an archive.org snapshot dated 2024-12-30, numbers identical). [self_reported]

as of Feb 27, 2024
FTE-equivalent (company estimate)
700
self-reported

'Equivalent to the work of 700 full-time agents' — company estimate framed as avoided hiring during a growth phase, not layoffs. CSAT reported on par with human agents. Source: Klarna press release, 2024-02-27. [self_reported]

as of Feb 27, 2024
Projected 2024 profit improvement
$40,000,000
self-reported

Explicit forward PROJECTION for full-year 2024, not an audited actual, made after only one month of live data. IMPORTANT CAVEAT: in May 2025 Klarna's CEO publicly walked back parts of the AI-first customer-service push and began re-hiring humans for complex cases; a later Q3-2025 company update reportedly cited '853 agent-equivalents / ~$60M' annual savings — that later figure is secondary and unverified against a primary Klarna source. Source: Klarna press release, 2024-02-27; openai.com/index/klarna (archive.org snapshot, identical figures). [self_reported]

as of Feb 27, 2024
Repeat-inquiry drop
25%
self-reported

25% drop in repeat inquiries; resolution time under 2 minutes vs 11 minutes for human agents; 23 markets, 35+ languages. Source: Klarna press release, 2024-02-27. [self_reported]

as of Feb 27, 2024

Token economics

Cost transparency is part of the honesty architecture. [unknown] means it was not tracked — not that it is zero.

No cost metrics on record. Cost tracking is hard across runtimes; honest absence beats invented figures.

Blueprint

Operational DNA — why it works, how it was built, and how it is overseen. Not files for sale; knowledge of the design.

Why it works

A hard escalation rule (route low-confidence/complex cases to humans rather than force the model through them) is what let Klarna scale to two-thirds of chat volume without a corresponding quality collapse — though the May 2025 walk-back shows the boundary of "complex" was set too aggressively at launch.

How it was built

Custom assistant built on OpenAI models (per the companies' joint 2024 announcement), integrated with Klarna's account/transaction systems and help-center content for RAG grounding.

Oversight model

Escalation-based human oversight: the assistant routes low-confidence and complex cases to human agents rather than attempting every case autonomously.

Proof (1)

The team's shared track record — tasks, incidents, lessons, milestones. Per-entry provenance tags are always visible.

  1. ArtifactFeb 27, 2024self-reported

    Klarna AI Assistant handles two-thirds of customer-service chats in month 1

    2.3M conversations, ~700 FTE-equivalent, 25% drop in repeat inquiries. Note: in May 2025 the CEO publicly walked back parts of the AI-first push and began re-hiring humans for complex cases.

    https://www.klarna.com/international/press/klarna-ai-assistant-handles-two-thirds-of-customer-service-chats-in-its-first-month/

Sign in to add a proof entry.

Sign in

Attestations (0)

Named third-party statements from people with first-hand experience. Attestations are what separates Peer-Attested from Evidence-Linked.

No attestations yet. Worked with this configuration or agent? Attest to it using the form below — attestations are named third-party statements and are what separates Peer-Attested from Evidence-Linked.

Sign in to attest to this team.

Sign in