Bank of America announced on March 26 a full-scale deployment of AI agents across Merrill Wealth Management and Bank of America Private Bank, reaching approximately 15,000 employees in the first phase. The system — built on Salesforce Agentforce — handles client meeting preparation, real-time AI note-taking during calls, and automated follow-up generation. A subset of roughly 1,000 financial advisers is now using AI agents to handle client queries and prepare live recommendations.

The headline figure: up to four hours saved per client meeting.

What the System Actually Does

The deployment, called the “AI-Powered Meeting Journey,” breaks a client interaction into three phases where AI handles the labour-intensive work.

Before a meeting, the system prepares a comprehensive briefing — pulling together the client’s portfolio, recent account activity, relevant market context, and any open action items from previous conversations. Advisers arrive informed without spending two hours digging through notes.

During the meeting, AI handles note-taking in real time, capturing key decisions, client concerns, and commitments made. No manual transcription, no post-meeting scramble to document what happened while it is still fresh.

After the meeting, the system generates follow-up summaries and recommended actions automatically. The adviser reviews and sends rather than composing from scratch.

The four-hour saving figure covers all three phases combined. For an adviser managing dozens of client relationships, that adds up to a meaningful shift in how time is allocated — away from administrative overhead and toward the high-value work of actual advisory.

Why This Deployment Matters

This is not a pilot programme. Bank of America is rolling this out to 15,000 employees in the first phase across two of its most prominent wealth management divisions. That is institutional-scale deployment, not an experiment.

The ROI case for AI agents in knowledge work has been debated for years. Productivity estimates from consulting firms, projections in think-tank reports, percentage-improvement claims from technology vendors — all of it has been somewhat abstract. This is different. A major financial institution has deployed, measured, and publicly stated a four-hour per-meeting efficiency figure attached to a specific workflow.

That figure is going to show up in a lot of board presentations over the next few months.

The Pattern Repeating Across Industries

What Bank of America has done with financial advisers is structurally identical to what we are seeing across professional services. The same pattern keeps appearing:

• A knowledge worker spends significant time on meeting prep, documentation, and follow-up
• Those tasks are not the job — they are the overhead required to do the job
• AI agents handle the overhead
• The knowledge worker spends more time on the actual high-value work
• Efficiency improves measurably

This pattern works in financial advisory, legal services, medical practice management, real estate, consulting, and anywhere else where a skilled professional spends a meaningful fraction of their day on administrative work that does not require their expertise.

The tools to do this are available now. The infrastructure — Salesforce Agentforce in Bank of America’s case — is enterprise-grade, compliance-ready, and proven at scale.

What Smaller Businesses Can Take From This

Bank of America has enterprise resources that most businesses do not. They have Salesforce at enterprise pricing, internal technical teams, compliance infrastructure, and the budget to build this properly. That can make it feel distant from the reality of a 50-person professional services firm.

But the core workflow is not complicated. Meeting prep, note-taking, and follow-up generation are not uniquely complex at Bank of America’s scale. A financial adviser at a regional wealth management firm has the same bottlenecks as a Merrill adviser. The four-hour figure does not scale down by 90% just because the firm is smaller.

The difference is implementation support. Bank of America had Salesforce and internal engineers. Smaller businesses working with Omni Ops get the same outcome without the internal build team.

---

What This Means for Enterprise DNA Clients

The Bank of America deployment makes the ROI case for AI agents in client-facing professional roles about as concrete as it gets. If your business involves advising, consulting, or managing ongoing client relationships — and your team spends hours each week on meeting prep and documentation — this is the playbook.

If you want the playbook other teams are using with Claude and Codex right now, grab the free Working With Claude field guide. Download it here.