On May 29, finance leaders gathered in New York City for an exclusive CFO round-table to discuss what many see as the next evolution of AI in finance: Multi AI Agent Collaboration in Finance & Accounting.

Hosted at Antalia NYC Restaurant, the session brought together CFOs and finance executives for a structured conversation, focused not just on individual AI use cases, but on how multiple AI agents can work together across finance and accounting workflows.

Moderated by Rajeev Pathak and John Silverstein, the discussion followed a clear progression, from identifying use cases to understanding collaboration and control.

Identifying Use Cases Across Finance & Accounting

The conversation began with a practical question: 

“Where do AI agents actually fit within finance processes?”

Participants identified several high-impact areas where specialized agents can operate effectively, including:

• Accounts Payable and invoice processing

• Expense management and reconciliation

• Financial planning and reporting

• Cash flow and payments tracking

Rather than relying on a single system, the idea of multiple purpose-built agents, each handling a specific function, emerged as a more scalable and flexible approach.

Why AI Agents Need to Work Together

A key theme quickly surfaced: individual agents deliver value, but real impact comes from collaboration.

Finance workflows are inherently interconnected. For example:

• An invoice processing agent feeds into accounting and reporting

• Expense data from payments influences budgeting and forecasting

• Cash flow insights depend on inputs from multiple systems

Without coordination, these agents operate in silos. But with collaboration, they enable end-to-end workflow intelligence, where information flows seamlessly across processes.

How Multi-Agent Collaboration Works

The discussion then moved into how these agents actually interact.

Finance leaders explored different models of collaboration, including:

• Sequential workflows: One agent’s output becomes another’s input

• Parallel processing: Multiple agents working simultaneously on related tasks

• Orchestrated systems: A central layer coordinating actions across agents

The takeaway: Successful multi-agent systems require structured communication and orchestration, not just isolated automation.

Human Oversight: Staying in Control

While autonomy is increasing, participants emphasized that human control remains critical.

Key mechanisms discussed included:

• Approval checkpoints for sensitive financial decisions

• Clear visibility into agent actions and outputs

• The ability to override or intervene when needed

Rather than replacing human judgment, multi-agent systems are designed to support decision-making while keeping humans firmly in control.

From Concept to Capability

What made this discussion particularly valuable was its focus on real-world applicability. Multi-agent AI is no longer theoretical; it’s becoming a practical framework for managing complex finance operations.

The structured agenda helped build a clear narrative:
Use cases → Need for collaboration → Methods of interaction → Human control

This progression reflected how organizations are beginning to think about adopting multi-agent systems in a practical, scalable way.

A Collaborative Exchange Among Finance Leaders

As with all round-tables, the conversation extended beyond the formal agenda. Over drinks and dinner, CFOs exchanged perspectives, shared challenges, and explored how these ideas translate into their own organizations.

The peer-driven format made it possible to move beyond surface-level discussion into real, experience-backed insights.

Looking Ahead

The discussions in New York highlighted a clear shift:

The future of AI in finance isn’t just about smarter tools; it’s about systems of agents working together.

As organizations explore this next phase, the focus will be on:

• Designing collaborative agent ecosystems

• Ensuring seamless integration across workflows

• Maintaining control while scaling automation

Multi-agent AI is not just an evolution; it’s a step toward fully connected, intelligent finance operations.

Follow us on LinkedIn to stay updated on upcoming CFO round-tables and discussions.

Curious how multi-agent AI could work within your finance workflows? Book a demo with our finance experts to explore where it can drive the most impact for your workflows.