ETEGY / Insights / Finance Transformation / Before you deploy agents
Insights · Finance Operating Model

New systems, old workflows — and now agents on top?

Finance has modernized its systems — a new ERP, a cloud close, an FP&A platform — while the workflows around them stayed manual, disconnected, and undocumented. Layer AI agents onto that and you don’t get speed; you get fragmentation, faster. The platform is new. The operating model underneath it never changed.

A new system is not a new operating model

Finance transformation is usually sold as a platform: migrate the ERP, move the close to the cloud, stand up a modern FP&A tool. The systems get modernized — and then the real work still runs on spreadsheets, email approvals, manual reconciliations, and tribal knowledge stitched between the platforms. The tools changed; how work actually moves did not.

That gap is where the return goes missing. A modern close still takes too long because the workflow feeding it was never standardized. Forecasts stay slow and mistrusted because the data lineage behind them is undocumented. The platform can be state-of-the-art while the operating model it sits on is exactly as fragmented as before — and no dashboard makes that visible until the numbers are late.

The CFO data points to one gap

Finance leaders are clear on the priority and the direction of travel. What the research also shows is where it stalls: modernized technology sitting on manual, disconnected workflows — the exact surface onto which agents are now being deployed.

50%

of finance leaders name digital transformation of finance their top priority for the coming year — the clearest signal on the CFO agenda.

Deloitte · CFO Signals
54%

say integrating AI agents into finance will be one of their top finance-transformation priorities — even as workflows remain manual.

Deloitte · CFO Signals

Underneath the enthusiasm sits the constraint the same research keeps surfacing: systems have been modernized while workflows stay manual and disconnected, limiting the return on the investment. Gartner, meanwhile, puts cost optimization at the top of the enterprise agenda — which makes an unstandardized finance function, and the rework it generates, an expensive place to layer automation. Automate the gap and you scale it.

Modernized systems vs. standardized workflows
Where finance has invested — and where the operating model was left behind, function by function
Close & consolidateProve
Systemmodernized
Workflow
FP&A & forecastGet
System
Workflow
AP / AR & transactionsDo
System
Workflow
Controls & reportingProve
System
Workflow
System modernized Workflow standardized Illustrative · the gap is where agents amplify dysfunction

The gap between a modernized system and a standardized workflow is the return that never lands — and the surface onto which agents get deployed before it is closed.

50%
Say digital finance is their #1 priority
54%
Prioritizing AI agents in finance
Manual
Workflows still run under modern systems

Don’t automate the gap — standardize it first

The tempting sequence is to modernize the platform, then point agents at whatever is still slow. But an agent, like any automation, executes the workflow it is given. Hand it a manual, undocumented process and it will run that process — faster, at scale, and with exceptions no one can trace. The sequence is the whole game: standardize and make the workflow traceable, then automate it.

Modernize, then automate the gap
Point agents at the workflow as it is
  • New platform, same manual handoffs
  • Agents encode undocumented steps
  • Exceptions multiply with no owner
  • Close and forecast stay slow and mistrusted
  • No one can trace what the agent did
Standardize first, then automate
Make the workflow traceable, then deploy
  • The finance operating model is read and standardized
  • Data lineage documented before automation
  • Agents inherit a governed, stable process
  • Every output traces to a governed origin
  • Close and forecast become defensible, audit-ready

This is the same discipline that makes any automation safe: standardization is the prerequisite. In finance it is also a control requirement — a forecast or a close that can’t be traced to its origin is not just slow, it is a risk the audit committee owns.

GSDPI, applied to the finance function

The same five stages that read any operating model sequence a finance transformation correctly — putting the workflow-and-lineage work before the automation, not after go-live.

  1. G

    Get Read the actual state

    Map how the close, forecast, and transaction cycles truly run — the spreadsheets, the email approvals, the manual reconciliations between the modern platforms — not the process diagram in the ERP deck.

  2. S

    Sort Standardize & document lineage

    Standardize the workflow and document the data lineage behind every number. Decide how the work should run and where each figure originates — before any agent is pointed at it.

  3. D

    Do Automate the governed workflow

    Deploy automation and agents onto the standardized process, not the manual one. Every automated step is configured against a governed workflow with a documented source.

  4. P

    Prove Trace every output

    Each forecast, close entry, and report traces to a governed origin — defensible to the audit committee and the board. Speed never outruns control.

  5. I

    Improve Govern the model

    Keep the finance operating model a living, governed system as agents take on more. Each change traces to the model — automation compounds control instead of eroding it.

Where finance value leaks — point by point

The problem a CFO experiences as “we modernized and it’s still slow” resolves into specific breaks — nearly all of them in the workflow, not the platform. The six that carry the most weight:

The close runs on spreadsheets between systemsProve

A modern consolidation tool still waits on manual reconciliations and email approvals feeding it. The close is slow not because the system is weak, but because the workflow around it was never standardized.

Forecasts can’t be traced to their sourceGet

FP&A produces a number the business doesn’t trust because no one can show where it came from. Undocumented data lineage makes every forecast a negotiation instead of a fact.

Agents get pointed at the manual processDo

An agent is deployed onto the workflow as it is — encoding the manual steps and workarounds at machine speed. Exceptions multiply, and no one can trace what the agent actually did.

Controls become an afterthoughtProve

When outputs can’t be traced to a governed origin, controls are reconstructed after the fact for the auditors. What should be built into the workflow is bolted on at quarter-end.

Each platform is an islandSort

ERP, close, FP&A, and reporting were modernized separately, and the handoffs between them stayed manual. The finance operating model is a set of disconnected tools, not one governed system.

Modernization is treated as the finish lineImprove

The platform goes live and the program ends — but the workflow was never standardized, so the value keeps leaking. Without a governed model, the new system slowly ossifies around the old habits.

How ETEGY reads it

Every leak above traces to one root: the systems were modernized while the finance operating model was left manual and undocumented. We read the actual state, standardize the workflow and its lineage, then let automation and agents act on a model that is governed and traceable.

The ETEGY read

A Zero-Based read of the finance operating model across the GSDPI lifecyclebefore agents are deployed — so automation lands on a standardized workflow with documented lineage. Then the Traceability Ratio holds every close entry and forecast to a governed origin. Standardize first; automate second. See the ZBT Discovery →

ETEGYInsight
The Finance Readiness Scorecard
Finance · ScorecardPDF
Readiness scorecard · PDF

The Finance Automation-Readiness Scorecard

Five dimensions of the finance operating model, each scored 0–3 for whether it’s ready to automate — with a verdict on where agents will pay and where they’ll amplify. Built to circulate: take it into your finance leadership team.

Open the one-pager →
Coming soon · the book

Zero-Based Transformation — be first to read it.

Be first to know when Zero-Based Transformation™ publishes.
Occasional launch updates only.

Open the signup form →

An agent runs the workflow you give it. Give it one you’ve standardized.

An agent inherits the close, forecast, and controls exactly as they run today — exceptions and manual saves included. See what it would take on, before automation makes it permanent.