Cash and working capital
Billing, collections, payroll, procurement, taxes, debt service, project mobilization, and exceptions should roll into a living liquidity view.
KSIG Advisors / Insights / Finance operating system
CFO transformation · AI · Automation · Controls
For a middle-market company, AI in finance should not begin with “What tool should we buy?” It should begin with “How should this business operate?”
The shift
Andreessen Horowitz recently argued that today’s scarce CFO skill is increasingly the ability to design the operating system the company runs on: the data, workflows, agents, skills, and controls that convert context into intelligence and decisions.
That framing is particularly relevant in founder-led, construction, and other project-based businesses. The 13-week cash forecast may be wrong because AR is stale. Margin reporting may be late because committed costs are disconnected. WIP may be unreliable because field and accounting systems do not agree. Collections may move slowly because responsibility is unclear. Equipment decisions may sit outside the capital plan.
Those are not isolated spreadsheet problems. They are symptoms of how the business moves information.
Start with the process
A modern finance operating system should connect the economic chain of the business. In a project-based company that may mean estimate → contract → schedule → committed costs → WIP → change orders → billing → collections → cash.
AI can help gather information, reconcile records, prepare scenarios, draft commentary, route work, and surface exceptions. But finance still has to define the rules: which system is authoritative, what requires approval, what constitutes an exception, who can change sensitive information, and when an agent may recommend versus act.
The objective is not to automate finance. It is to make finance faster, more reliable, and more useful without losing accountability.
What the operating system should connect
Billing, collections, payroll, procurement, taxes, debt service, project mobilization, and exceptions should roll into a living liquidity view.
Estimate, job cost, committed costs, WIP, change orders, retainage, and cost-to-complete should reconcile quickly enough to influence operations.
Forecasting should move from periodic spreadsheet production toward a decision system that can be refreshed as the business changes.
Approvals, permissions, evidence, source data, and exception paths should be designed into workflows before automation is allowed to scale.
Finance should surface the few decisions that matter instead of forcing management to reconstruct the story from reports after the fact.
Build-versus-buy, integration, vendor dependence, usage pricing, security, and maintenance are financial decisions as much as technical ones.
Smaller team, higher leverage
a16z describes finance organizations becoming smaller and more leveraged as AI-native tools and internal automation take on work that once required manual reconciliation, reporting, and handoffs.
For the middle market, the practical upside is less time assembling data, moving information between systems, preparing repetitive reports, and chasing routine explanations. That creates more capacity for forecasting, margin analysis, pricing, working-capital management, lender readiness, capital planning, and operating decision support.
The CFO’s job becomes less about supervising the production of information and more about designing a system that produces reliable information continuously.
Control is not the brake
As software performs more finance work, traceability becomes more important. If an agent prepares a journal entry, the evidence and approval rules need to survive with it. If a forecast is updated automatically, management should be able to see what changed. If a workflow touches payment instructions, authority must be explicit.
The finance organization should define the boundary between preparation, recommendation, approval, and execution. That boundary is not administrative overhead. It is part of the architecture.
The CEO question
The question for owners and CEOs is no longer simply whether the CFO understands accounting. Financial truth remains foundational, but increasingly the CFO also needs systems judgment.
The modern CFO is not just reporting on the operating system of the business. Increasingly, the CFO is helping build it.
Source
Finance systems conversation
KSIG helps founder-led and growth-stage companies connect forecasting, cash, controls, reporting, systems, and decision cadence into a finance operating model management can actually use.
Discuss the finance operating system