{"id":125322,"date":"2026-03-17T14:00:00","date_gmt":"2026-03-17T21:00:00","guid":{"rendered":""},"modified":"2026-03-17T11:30:30","modified_gmt":"2026-03-17T18:30:30","slug":"how-frontier-firms-use-agentic-ai-to-gain-an-edge-in-capital-markets","status":"publish","type":"post","link":"https:\/\/www.microsoft.com\/en-us\/industry\/blog\/financial-services\/2026\/03\/17\/how-frontier-firms-use-agentic-ai-to-gain-an-edge-in-capital-markets\/","title":{"rendered":"How Frontier Firms use agentic AI to gain an edge in capital markets"},"content":{"rendered":"\n

This blog post is guest-authored by Thomas Shuster, Research Director, Worldwide Capital Markets, Wealth, and Digital Assets, IDC Financial Insights<\/em><\/strong><\/p>\n\n\n\n

As capital markets firms push toward the frontier, success increasingly depends on turning AI ambition into secure, repeatable operating impact at global scale. In this independent IDC guest blog, Thomas Shuster examines how agentic AI is reshaping capital markets operating models\u2014and why firms are gravitating toward platforms and partners that combine technological leadership, deep industry expertise, strong governance foundations, and proven experience delivering AI value across the end-to-end value chain.<\/em><\/p>\n\n\n\n

\n
 Explore infographic: Expanding the Frontier: Agentic Transformation in Capital Markets<\/em><\/a><\/div>\n<\/div>\n\n\n\n

When capital markets leaders talk about Frontier Firms, it is important to recognize that the term\u2019s definition has shifted. It is less about being first to experiment with new tools and more about translating AI investment into measurable, repeatable operating gains. That distinction matters as the operating environment tightens. Settlement cycles continue to compress, regulatory expectations change, and risk controls must remain effective as markets evolve. At the same time, technology teams are expected to modernize while continuing to support large legacy environments. In this context, agentic AI emerges as a practical marker of frontier operating models.<\/p>\n\n\n\n

From tools to operating models<\/h2>\n\n\n\n

Early generative AI tools improved drafting, summarization, and search. These capabilities were helpful but not transformative or differentiated. The step change occurs when firms shift from task acceleration to workflow redesign, deploying AI agents to execute multistep processes across systems under bounded human oversight.<\/p>\n\n\n\n

Frontier Firms focus on workflows characterized by high friction, frequent exceptions, and material costs when delayed. They redesign processes so agents perform the coordination and context gathering work that typically slows teams down: pulling data, checking policies, identifying breakpoints, proposing actions, and routing tasks to the right owners. Humans remain accountable for decisions but no longer act as the connective tissue that holds workflows together. This shift has important workforce implications because human effort moves away from manual orchestration and toward judgment, escalation, and decision-making.<\/p>\n\n\n\n

By contrast, non-Frontier Firms often attempt to layer AI onto workflows still defined by manual handoffs and fragmented systems. These initiatives may succeed in pilots but frequently stall when exposed to real-world operational variability.<\/p>\n\n\n\n

\n
Read IDC\u2019s Analyst Connection: How and Where Frontier Firms Are Bending the Curve on Innovation<\/em><\/a><\/div>\n<\/div>\n\n\n\n

Integration, not intelligence, is the limiting factor<\/h2>\n\n\n\n

Many operational breakdowns in capital markets stem from fragmented information. Trade exceptions can span execution data, reference data, allocations, settlement instructions, and counterparty communications. Know your customer (KYC) refreshes depend on sanctions data, beneficial ownership structures, customer documentation, and policy interpretation. These are inherently cross-system and, increasingly, cross-organization challenges.<\/p>\n\n\n\n

Frontier Firms treat data access as a core capability rather than a downstream integration problem. They invest in ecosystems that support secure, permitted access to internal and external data with auditability and clear economic and contractual rules. In practice, the operating framework often matters as much as the underlying technology. Questions of data ownership, computational rights, value sharing, and dispute resolution frequently determine whether an agentic use case can scale. Where these foundations are absent, teams compensate with manual workarounds that are slow, error-prone, and difficult to audit.<\/p>\n\n\n\n

Governance as an accelerator<\/h2>\n\n\n\n

There is a persistent tendency in capital markets to defer governance until a use case has demonstrated value. That approach breaks down with agentic AI. Agents act within workflows and can trigger downstream consequences if controls are weak.<\/p>\n\n\n\n

IDC\u2019s research shows that only about 4% of financial institutions believe AI agents should operate with full autonomy. More than 75% rate transparency as very or extremely important, with the share rising to roughly 88% among Frontier Firms. How frontier organizations operationalize trust reflects these preferences. They define which decisions require human approval, log agent inputs and actions, establish clear escalation paths, and design workflows that make overrides straightforward. Many organizations also prefer to rely on platform-level governance capabilities rather than bespoke controls for each use case.<\/p>\n\n\n\n

When done well, governance becomes an enabler rather than a constraint. It allows firms to deploy agentic workflows more broadly and with fewer surprises, aligning risk and innovation teams. Where governance lags, organizations often see the opposite outcome: Risk teams perceive AI as uncontrolled, innovation teams view governance as blocking progress, and value remains trapped in isolated proof points.<\/p>\n\n\n\n

\n
Read IDC\u2019s Analyst Connection How and Where Frontier Firms Are Bending the Curve on Innovation<\/em><\/a><\/div>\n<\/div>\n\n\n\n

Where Frontier Firms pull ahead first<\/h2>\n\n\n\n

IDC finds that Frontier Firms adopt functional and industry use cases almost twice as much as their peers. Expectations for automation are also rising. In IDC\u2019s resiliency and spending research, 87% of firms expect providers\u2019 agentic AI capabilities to eliminate manual and semi-manual workflows within 18 months.<\/p>\n\n\n\n

The gap widens most quickly where speed, exception handling, and control converge. In post-trade operations, many organizations still manage exceptions through email and informal handoffs, slowing resolution, and weakening auditability. Frontier Firms move toward agent-supported, structured case management. In onboarding and due diligence, event-driven regulatory expectations are making periodic refresh models brittle. While only about 10% of financial institutions used AI for regulatory compliance in the past year, nearly 90% plan to do so in the next 12 months. In research and intelligence functions, agents increasingly monitor sources, summarize changes, and map exposures, shifting human effort from aggregation to decision making.<\/p>\n\n\n\n

AI is reshaping business models<\/h2>\n\n\n\n

The frontier advantage is not limited to efficiency. IDC\u2019s research shows that organizations using agentic AI report a 2.3-time return on investment (ROI), with average payback periods of about 13 months. These attractive economics are accelerating investment. Building customized AI agents to automate business processes ranks as the top area of significantly increased IT spending among capital markets firms in 2026, which more than 80% of organizations have cited.<\/p>\n\n\n\n

As these agents mature, firms are also reassessing their application strategies. In IDC\u2019s survey, 84% of financial services firms agree that AI agents are emerging as a new layer of enterprise capability, prompting renewed scrutiny of investments in packaged applications.<\/p>\n\n\n\n

Closing thought<\/h2>\n\n\n\n

Agentic AI is not a shortcut around complexity. It is a way to absorb complexity without scaling cost and risk linearly. Ambition alone does not distinguish Frontier Firms. Differentiating them are data access, governance discipline, operating model design, workforce readiness, and organizational habits required to turn agentic AI into a durable source of advantage.<\/p>\n\n\n\n

Explore more insights on agentic AI in capital markets<\/h2>\n\n\n\n