Banks urged to overhaul dispute management

Industry experts are urging banks and payment firms to overhaul dispute management as AI-enabled fraud and friendly fraud increase. UK scam losses reached £1.3 billion in 2025; US imposter-scam losses were $3.5 billion the same year. The United Nations office estimated Asia Pacific losses at $88.3 billion in 2025.

Banks face three trends that are increasing dispute volumes and complexity: criminals using AI to scale scams, a rise in first-party misuse known as friendly fraud, and a wider range of payment methods that fragment dispute workflows. Friendly fraud occurs when customers request chargeback refunds for legitimate purchases and is projected to account for 28% of all chargebacks worldwide by 2031.

Firms have increased investment in AI for transaction detection. Panels convened by industry organisers in association with Pegasystems noted that stopping bad transactions is no longer sufficient and that chargeback and claims management systems need upgrades to handle higher volumes and more complex cases.

Upgrades described by panelists include using AI to speed case handling, apply consistent decision rules, automate routine work and reduce manual backlogs. Financial institutions need more contextual data to separate deliberate abuse from genuine customer errors. Introducing extra steps for some customers may be necessary; operators must balance added steps to avoid unnecessary customer churn while protecting merchants and card issuers from loss.

Generative models and automation are being used by criminals to craft impersonation scams and scale attacks. The same technologies are being applied by institutions to detect anomalies, triage cases and automate evidence collection. Panels will examine which AI techniques work at different dispute stages, from behavioral models in transaction screening to natural-language systems that parse customer claims and merchant evidence. Discussions will also address whether current models can keep pace as automated agents begin transacting on behalf of consumers, a trend called agentic commerce, and whether fraud teams are structured to respond to high-volume automated attacks.

Dispute outcomes must comply with card network rules, local consumer-protection laws and internal service-level targets. Panelists said that better integration across fraud detection, customer service and legal or compliance teams is needed so that evidence and decisions travel with a case rather than being recreated at each step.

Organisers listed speakers for the sessions including Scott Andrick, senior director at Pegasystems, and Scott Hamilton, global payments and liquidity expert and contributing editor. Panel discussions are set to focus on practical changes to triage disputes, deploy AI where it increases accuracy and redesign processes so dispute handling keeps pace with new payment methods and automated commerce.