2026 解锁未来：金融服务行业代理型AI应用实践指南

A playbook that explores the power of Table of contents Executive Summary 01Introduction 02Sector spotlights •Banking•Insurance 03Becoming a Frontier organisation Authors Executive Summary Financial services organisations are operating in a period ofprofound transformation. Macroeconomic uncertainty, risingcompliance demands, increasing fraud risk, shifting customerexpectations and a rapidly ageing workforce are all reshapinghow banks, insurers and capital markets firms must operate. At Agentic AI represents a step change from traditional analytics,automation and generative AI. Rather than completing isolatedtasks, agents can reason, act and orchestrate multistepprocesses with minimal human intervention—accessing data, The impact of agentic AI is already visible across the industry. Inbanking, significant value is already emerging in financial crime,regulatory compliance, customer engagement and lending.Agents can gather and validate Know Your Customerinformation, assemble source-of-wealth documentation and In insurance, agentic AI is transforming underwriting, claims,customer service and workforce sustainability. Underwriters canbase decisions on continuously updated risk signals drawn frompolicy data, claims activity, climate models, behavioural insightsand geospatial imagery. Claims can be triaged and resolved Capital markets organisations are leveraging agentic AI for highvalue analytical and operational activities. Multi-agent workflowscompress time to first analysis from days to minutes byingesting financial statements, earnings calls, market news andalternative data into structured research drafts. Agents can also Across all subsectors, a common pattern is emerging: agenticAI unlocks significant value when it transforms decision-heavy,high-complexity workflows, transforms customer experience,and connects previouslydisconnected parts of the enterprise.Yet scaling these capabilities requires more than isolated pilots.Financial institutions must build reusable, enterprise-wide AI Achieving this transformation requires a clear, actionableroadmap. Leaders should define a bold but responsible visionfor AI; raise organisational literacy and confidence; redesignoperating models around value chains, not isolated tasks;modernise data and cloud foundations; embed Responsible AI Agentic AI is no longer a distant promise. It is a practicalcatalyst for transforming financial services today. Institutionsthat act now—scaling responsibly, governing effectively and Introduction Market situation and industry challenges Financial services companies are facing a rapidly accelerating pace of change. Even though theindustry has shown greatresilience in the face of recent shocks, it is contending with manychallenges that could increase risks in the near future. Among the challenges that are convergingin a highly complex market environment: ongoing digital transformation, an ageing workforce, Meanwhile, new regulatory requirements have come into force across different jurisdictions withthe aim of safeguarding the stability and security of the global financial system. But these arealso increasing compliance pressures and costs for banks, insurers and other institutions. PwC’s Many new regulations affecting the industry focus on cybersecurity and cyber resilience due tothe growing risk of fraud, ransomware, hacking and other threats. For example, worldwide lossesdue to fraud in banking are projected toreach $58 billionby 2030. Rising fraud and compliance $58 billion by 2030 Given all these developments, it’s clear that financial services organisations need to prepare for afuture business environment that could look very different from today’s. And they’re increasingly What are AI agentsand agentic AI? AI and agentic AI represent different layers of capability. Beforeexploring how financial institutions use them, it’s important toclarify what these systems are and how they differ. Artificial intelligence (AI) refers to systems and technologies thatuse data and advanced algorithmic models—including machinelearning and deep learning—to execute tasks traditionallyrequiring human intelligence, such as predicting outcomes,making or recommending decisions, automating processes and Agentic AI goes beyond the capabilities of traditional AI, whichis limited to identifying anomalies and scoring risks withoutbeing able to take independent action. It also outperformsgenerative AI, which can only create content in response to In contrast, agentic AI can operate with minimal humanintervention. While critical decisions in financial organisationswill continue to require human accountability, AI agents can Agentic AI is redefining what is possible for financial services institutions offering innovativesolutions to longstanding industry challenges. Within the financial services industry, manycompanies are emerging as early adopters of AI, driven by margin pressure, regulation, and By automating complex tasks, enhan