Finance chiefs head to CES 2026 to source AI beyond lending
Senior leaders from Korea's top financial institutions are heading to Las Vegas next month to translate global AI progress into practical financial services. With margins pressured by low interest rates and tighter rules, the focus is shifting from hype to implementation-efficiency, risk controls, and new fee income.
KakaoBank will be the only Korean internet-only bank on the ground. Industrial Bank of Korea (IBK) will be the only Korean bank running an exhibition booth, while KB, Shinhan, and Woori groups are sending CEOs and digital strategy teams.
Who's going-and what they're looking for
- KakaoBank: CEO Yun Ho-young (Daniel Yun) and AI/strategy teams will track AI agents, blockchain-based fintech, AI-driven security, and on-device authentication. Yun has emphasized AI as core to the bank's identity, attending global tech events (including Money20/20 Asia) and pushing AI-powered search and AI-enabled fund transfers.
- IBK: Operating a standalone booth featuring the bank's technologies plus seven partner firms, including interactive programs that let visitors experience Korea's content investment process.
- KB Financial Group: Delegation led by Choi Jae-hong, professor at Gangneung-Wonju National University and an independent non-executive director, joined by working teams from affiliates.
- Woori Financial Group: Roughly 30 participants, including deputy CEO-level leaders.
- Shinhan Bank: Scaling back to a working-level visit after three years of running a booth. At CES 2025, Shinhan demonstrated an unmanned branch with a virtual banker and its in-house AI chatbot.
AI themes that matter for finance leaders
- AI agents and copilots: Frontline service, collections, SME onboarding, and internal ops. Look for systems that can trigger actions across core banking, CRM, and risk tools-not just chat.
- On-device AI and authentication: Lower latency, better privacy, and resilience when connectivity is weak. Useful for step-up auth, continuous KYC, and device risk scoring.
- AI-driven security: Real-time fraud detection, anomaly spotting, and behavior biometrics that reduce false positives without creating customer friction.
- Blockchain-based fintech: Tokenized deposits, instant settlement, programmable payments, and audit trails that support compliance and treasury use cases.
If you're sending a team, set a clear brief: reduce cost-to-serve, compress time-to-yes, and create net-new fee lines. That lens will filter most booth demos fast.
What each institution is signaling
- KakaoBank: Doubling down on AI as a product and operations lever, not just marketing. Expect focus on customer experience and measurable service innovation.
- IBK: Positioning as a partner hub for fintech and content investment workflows-useful signal for B2B collaboration models.
- KB, Woori, Shinhan: Strategy-led delegations suggest an emphasis on scouting and fast-follower execution rather than headline demos this year.
Due diligence checklist for vendor meetings
- Model governance: Who trains what, where models run, and how updates are controlled. Evidence of bias testing and incident response.
- Data boundaries: PII handling, redaction, encryption, and tenant isolation. Can they prove no training on your data without consent?
- Performance: Latency under load, multilingual accuracy, and edge/device constraints. Show live metrics, not slides.
- Integration: Native connectors for core systems, event buses, and audit logs. How fast to a sandbox demo inside your environment?
- Unit economics: Clear cost per task, not per token. Benchmarks versus your current manual or RPA baselines.
- Compliance readiness: Evidence for audits, explainability reports, and regional data residency options.
90-day pilots worth considering
- AI agent for self-service: Contain high-volume queries (password resets, card limits, basic disputes) with verified actions, not just answers.
- KYC/AML triage: Triage and summarize alerts with evidence links; keep humans on final decisions to reduce case time without compliance risk.
- On-device step-up auth: Combine behavior biometrics with passkeys for fewer OTPs and lower fraud.
- SME onboarding assistant: Document extraction, risk summaries, and auto-assembled credit memos inside your LOS.
- Ops copilots: Email-to-action for back office queues (chargebacks, address changes, statement requests) with full audit trails.
Risk and controls: what to pre-negotiate
- Model cards, eval reports, and drift monitoring access before pilot kickoff.
- Human-in-the-loop thresholds for high-impact actions; clear rollback plan.
- Shadow-mode periods with holdout groups to prove lift and spot edge cases.
- Vendor responsibility matrix covering outages, data incidents, and legal requests.
Why this CES matters for banks
The tech on display now connects cleanly to P&L: fewer call minutes, faster onboarding, better fraud catch rates, and lower model inference costs. On-device AI and agent frameworks reduce operational friction and compliance overhead, which is where the dollars are.
For executives, the move is simple: pick two production-grade use cases, align risk and IT on a 90-day run, and measure hard numbers. Everything else is theater.
Resources
Your membership also unlocks:
