Most Finance Leaders Are Waiting in the wrong AI Line

Picture a busy airport departure hall. Hundreds of travellers all heading to the same destination. Most are in the long queue. Shuffling forward. Checking their watches. Waiting for the line to move.

And over to the right, a handful of people walk straight through. No queue. No wait. Full access.

That is exactly where Finance and AI stands today.

Stages of Enterprise AI Maturity (W XO-Claude 2) (1)

New research across more than 80 Selling, General and Administrative (SG&A) business processes shows that the gains from getting AI right in Finance are extraordinary:

44%

Productivity gain
across all SG&A functions

42%

Finance cost cut
over 5-7 year horizon

40%

SG&A savings
~$180M for a typical $10B company

For a typical $10 billion company, that translates to approximately $180 million in cost savings and 1,379 FTE-equivalents of capacity redirected to higher-value work.

The problem is not the technology. The problem is that most Finance teams are standing in the wrong queue entirely.

THE 5 STAGES OF FINANCE AI ADOPTION

Research shows that AI adoption follows a predictable five-stage progression. Most Finance teams are stalled somewhere in the first three.

88%

Stage 1

Bought AI Tools

Many licences. No clear Finance strategy.

Most Finance teams have an AI subscription. Microsoft Copilot. ChatGPT Enterprise. A handful of vendor-bundled AI modules in the ERP. Sometimes all three at once.

What they do not have is a coherent answer to a simple question: what problem are we actually solving?

Buying AI tools is not a Finance AI strategy. It is the first step on a very long staircase.

66%

Stage 2

Ran AI Pilots

Small wins. Hard to scale. No clear path forward.

The Finance team ran a pilot. Journal entry automation. An AI-assisted variance commentary draft. A Copilot integration for the FP&A workbook. It worked in the demo.

Then the questions came. How do we govern this? What about audit trail? How do we scale it to 3,000 other processes? What happens when the data changes?

Pilots prove potential. They do not create transformation. 66% of Finance teams are stuck at this stage, running the same three pilots on repeat.

33%

Stage 3

Built AI Agents

Custom builds. Expensive. Fragile foundations.

One in three Finance teams has built an AI agent. A custom reconciliation agent. A close orchestration bot. An intelligent data extraction layer for AP.

And here is where most of them discover the hidden problem.

Agents are only as good as the knowledge they can access. An agent that cannot find the correct accounting policy, the right version of the intercompany elimination rules, or the current treatment for a complex transaction does not just slow down. It hallucinates. It acts on stale data. It gets the answer confidently wrong.

Most Finance teams skip the knowledge foundation and build agents directly on top of fragmented, ungoverned data. They are building on sand.

This is the Enterprise Knowledge Crisis. Read why it is the real blocker for Finance AI transformation: The Enterprise Knowledge Crisis and Why Agentic AI Is the Only Answer That Scales

23%

Stage 4

Scaled AI Agents

Some scale. Multiple agents. But fragmented impact.

A smaller group has pushed further. Agents deployed across Record-to-Report and Procure-to-Pay. Orchestration across multiple workflows. Meaningful automation at scale.

But the compounding effect has not kicked in yet. Finance Business Partners are still writing variance commentary manually. The CFO is still waiting for the synthesised answer. FP&A is still producing reports rather than driving decisions.

The agents are running. But Finance is still operating in the same model it always has. The transformation has not happened yet.

This is where FP&A meets Agentic AI and everything changes. Read more: The FP&A Marriage Made in Heaven

Stage 5

Redesigned the Finance Operating Model

End-to-end. Agentic. Maximum value.

Only 6% of Finance organisations have reached this stage. And the distance between Stage 4 and Stage 5 is not incremental. It is categorical.

Stage 5 Finance teams are not just running more agents. They have fundamentally redesigned how Finance works. The Agentic Finance Department runs with 10 specialist AI agents and 3 supervisors across all Finance towers, orchestrated through a unified intelligence layer, connected directly to ERP, bank portals, sub-ledgers and tax systems.

The Controller reviews agent decisions. They do not execute them at scale. Month-end close runs with 99%+ first-pass accuracy. Finance Business Partners spend 80% of their time in strategic conversations, not data gathering.

This is not a future state. It is running now.

See what end-to-end Agentic Finance actually looks like in production: Month-End Close and the Agentic Finance Department

THE FAST TRACK: HOW TO SKIP THE QUEUE

The 6% who have redesigned their Finance operating model did not do it by accident. They did not work through Stages 1 to 4 sequentially either. They found a different route.

Here is what the Fast Track looks like in Finance:

Step 1: Build the Knowledge Foundation First

Before a single agent touches a Finance workflow, your organisation needs a governed, unified intelligence layer across all Finance policies, procedures, reports and operational documentation. This is not optional. It is the prerequisite for everything that follows. Agents without governed knowledge do not just underperform. They create liability.

Result: 50-80% reduction in time spent searching for Finance information. 90%+ improvement in knowledge accessibility across the function.

Step 2: Deploy the Agentic Finance Department

Not one agent. Not a pilot. A complete operating model. 10 specialist agents and 3 supervisors across FP&A, Record-to-Report, Purchase-to-Pay and Order-to-Cash, orchestrated through IBM watsonx Orchestrate. ISO 27001 and PCI DSS compliant from day one. Open architecture, no lock-in.

Result: 33% faster close. 57% fewer forecast errors. 99%+ first-pass journal accuracy. Controllers reviewing, not executing.

Step 3: Let the Compounding Begin

Once the knowledge foundation is in place and the Agentic Finance Department is running, the compounding effect kicks in. Every process that an agent handles frees Finance capacity for higher-value work. FP&A stops producing reports and starts generating decisions. The CFO gets synthesised answers with root-cause attribution, not a spreadsheet.

Result: $180M in savings over 5-7 years for a typical $10 billion organisation. 1,379 FTE-equivalents of capacity redirected to strategic work.

THE QUESTION EVERY FINANCE LEADER NEEDS TO ANSWER

Which queue are you standing in?

The long one moves slowly. 88% of your peers are in it right now. You can wait. Or you can find the Fast Track.

The organisations that move decisively in the next 12 months will set a benchmark that competitors spend years trying to close. The research is clear: those who wait will spend that time trying to close a gap that keeps growing.

SEE THE FAST TRACK LIVE. IN 60 MINUTES.

In one session with our team, we will map your current Finance close process end-to-end, identify your top three automation opportunities ranked by impact and time-to-value, and show you a live Agentic Finance Department running a real month-end close.

What you will see in 60 minutes:

Your Finance close process mapped end-to-end, identifying the structured, unstructured and interactive work layers in your specific environment
Your top three automation opportunities, ranked by impact, feasibility and time-to-value
A live demonstration of the Agentic Finance Department running a real month-end close across ERP, bank portals, sub-ledgers and tax systems
How this is built on IBM watsonx Orchestrate: open architecture, no lock-in, compliant from day one
A 90-day roadmap sequenced correctly, scoped for your function, ready to execute