Despite the rise of digital tooling, banks continue to rely on spreadsheets for underwriting, financial crime investigations and risk reviews. Industry analyses show that spreadsheet reliance remains wide spread and risky. The Corporate Finance Institute highlights the persistent challenges: human error, version confusion, fragile formulas and data loss are all recurring problems in spreadsheet-driven environments (Corporate Finance Institute,2024). Denizon’s review of spreadsheet use in banks notes similar issues, warning that spreadsheets create “significant operational exposure” when used at scale in risk and compliance functions (Denizon, 2024).

In parallel, Coherent Global’s assessment of spreadsheet-related bank fines shows that regulators increasingly treat spreadsheet errors as indicators of weak governance (Coherent Global, 2023). The picture is clear. Excel is dependable until the moment it absolutely is not. And in risk and compliance, that moment matters.

This is the environment in which AI agents have emerged — not as a futuristic experiment, but as a practical answer to the workflow fragility banks can no longer ignore.

The shift that changed digital risk operations

For years, banks implemented automation in fragments: a sanctions checker here, a scoring model there, a templated report generator buried in a SharePoint folder. These improvements helped, but they did not fix the underlying problem. Risk and compliance are not standalone tasks. They are workflows.

A Chief Risk Officer at a major UK bank summarised this during an internal panel captured in industry reporting: “Our issue was not analytics. It was movement. Work did not flow.” It is a sentiment echoed across risk communities.

AI agents solved that movement problem. They move information, decisions and tasks across teams without waiting for a nudge, a reminder or a spreadsheet update.

Why Excel became the silent source of compliance risk

Banks rarely discuss it publicly, but anyone who has worked in risk knows the internal realities. Spreadsheets travel between multiple teams before a risk rating becomes final. Data updates hide inside inboxes that resemble archaeological digs. A policy change in London may take days to reach the offshore operations team. And as practitioners quietly admit, no audit ever failed because a spreadsheet formula was too elegant. Audits fail because no one can explain why three different versions of the same tracker exist.

Compliance analysts do not burn out from reviewing high-risk cases. They burn out from searching for supporting evidence across five teams. Transformation leads do not worry about work moving too fast. They worry that workflows move at the pace of Outlook reminders.

Excel became the digital equivalent of a filing cabinet everyone depends on and no one can control. AI agents do not replace Excel. They replace the operational chaos Excel was never designed to manage.

Where AI agents make the biggest impact

Banks modernising risk and compliance with AI agents are seeing measurable gains in workflow speed, quality and regulatory confidence.

1. KYC and onboarding workflows

AI agents validate documents, trigger risk scoring models, route cases and compile structured summaries for review. This shortens onboarding cycles and reduces early-stage risk exposure.

2. Transaction monitoring and alert handling

Instead of analysts spending hours collecting evidence, AI agents assemble timelines and build case packets automatically. Analysts spend more time interpreting signals and less time gathering data.

3. Sanctions and adverse media processes

Agents classify screening hits using domain models, enrich cases with supporting information and escalate only what needs human assessment. This reduces false positives and speeds resolution.

4. Reporting and regulatory submissions

OECD’s 2024 report on AI in finance highlights the growing expectation for consistent, auditable and repeatable data flows in risk reporting (OECD,2024). AI agents help institutions meet this expectation by generating structured outputs across Basel III, ICAAP, AML summaries and other regulatory submissions.

5. Internal coordination across teams

BCG’s 2024 analysis notes that 74 percent of companies struggle to scale AI value due to fragmented processes (BCG, 2024). Workflow agents directly address this issue by orchestrating tasks, dependencies and escalations across risk, compliance and operations teams.

The result is a risk and compliance function that finally feels like itis flowing rather than fire fighting.

What industry leaders are saying

Industry voices have been consistent about the direction of travel.

· Mark Brant, Chief Payments Officer at HSBC: “Operational risk often hides inside process delays. Fix the delays and risk becomes visible.” (Source: HSBC industry commentary)

· Bank of England & FCA Joint AI Report, 2024: Firms are expected to ensure their AI and data workflows are “reliable, timely and repeatable,” especially in risk and compliance programs. (Bank of England Report)

· Deloitte Financial Crime Report, 2025: “The next frontier in compliance transformation is workflow intelligence. Tools alone will not meet the rising volume and speed expectations.” (Deloitte, 2025)

· FSB (Financial Stability Board), 2024: Supervisors now expect firms to maintain consistent, auditable processes across jurisdictions, particularly in AML and sanctions. (FSB, 2024)

These perspectives reinforce one idea: risk is not only about decision quality. Risk is about how work travels.

Why this shift is accelerating in 2026

Three forces converged at the same moment.

1. Regulator expectations tightened

The Bank of England and FCA AI report highlights a clear shift toward requiring structured, traceable workflows.

2. Data volumes exploded

EY notes that the financial sector faces exponential increases in data volume and velocity, particularly in financial crime and operational risk (EY,2024). Manual operations cannot scale.

3. AI matured into workflow intelligence

What banks once called automation is evolving into AI agents capable of interpreting exceptions, managing dependencies and documenting every step. Itis a new operational layer.

Banks adopting AI agents consistently report:

· faster case handling

· fewer SLA breaches

· better audit readiness

· less operational fatigue

A rare combination in compliance.

What banks can modernise immediately

You do not need a multi-year transformation roadmap. You need three high impact starting points.

1. Fix onboarding workflows

a. This reduces risk exposure and improves customer experience simultaneously.

2. Modernise AML and sanctions case handling

a. Evidence gathering, enrichment and routing become structured instead of chaotic.

3. Introduce workflow agents for coordination

a. Teams stop drowning in email chases and SLAs improve within weeks.

Where LatentBridge fits

LatentBridge delivers AI-agent-powered workflow systems that modernise:

· KYC and onboarding

· AML, sanctions and adverse media

· Transaction monitoring

· Risk scoring

· Regulatory reporting

· Case assembly and document intelligence

· Cross-team workflow orchestration

Our accelerators include sanctions intelligence blocks, KYC extraction and validation packs, economic crime scoring models and workflow agents designed to keep cases moving. All deployed without requiring banks to over haul existing systems.

Risk and compliance finally move the way they were always meant to.

Final insight

Banks do not struggle with analytics. They struggle with movement.

· Excel was the bridge.

· AI agents are the operating system.

· The banks leading in 2026 are not the ones with the most tools.

· They are the ones with workflows that flow.