For most of my career, technology was something the finance organization used. Increasingly, it is something the CFO helps design.
That may be one of the more consequential changes happening inside finance. Andreessen Horowitz recently described the evolution of the CFO from accountant, to banker, to FP&A leader, and now toward a builder and architect of the company’s operating system. Their definition includes the data, workflows, agents, skills, and controls that turn context into intelligence and decisions.
That description resonates because I am already experiencing it. After more than 25 years operating businesses and leading finance functions, I increasingly find myself doing work that once would have sat almost entirely inside IT or software development.
I use AI to write code. I build agents. I automate recurring workflows. I connect systems and data. I build and improve 13-week cash forecasting processes. I design controls, permissions, and approval paths around automated work. I evaluate whether a problem should be solved by buying software, integrating existing systems, automating a workflow, or simply building something ourselves.
None of that makes finance less financial. It makes finance more capable.
From reporting the business to designing how it runs
The traditional finance function is largely built around intervals: close the month, produce the statements, update the forecast, review performance, hold the meeting, make a decision, repeat.
AI begins to compress those intervals. The a16z analysis points toward continuously updated forecasts, more real-time finance functions, and software that increasingly performs work rather than merely helping employees manage it.
If information moves continuously, finance can increasingly identify a margin problem, liquidity constraint, collection issue, or spending anomaly when it develops instead of weeks later. The CFO therefore becomes responsible for more than interpreting the information. Increasingly, the CFO helps design how the information moves.
Data → systems → workflows → controls → intelligence → decisions → actions.
The financial statement remains important. But the operating architecture that produces the financial statement, forecast, and management decision may become even more important.
Domain knowledge becomes more valuable, not less
One of the more interesting consequences of AI is that technical capability is moving closer to the people who actually understand the business problem.
Historically, a CFO might identify a process that needed improvement, develop requirements, hand those requirements to IT, wait for resources, evaluate vendors, implement software, and eventually receive a solution.
Now I can identify a problem and begin prototyping the solution myself.
That does not eliminate engineering, IT, cybersecurity, or systems expertise. It changes the distance between business judgment and executable systems.
That may actually increase the value of experienced finance leaders. A junior employee may know how to use an AI tool. A senior finance executive knows what the output should look like, which exceptions matter, where controls can fail, what the economics mean, and which decisions should never be delegated.
a16z makes a similar observation: senior domain experts can extract extraordinary leverage from AI because experience supplies the context necessary to know what is worth building and what “good” looks like.
The finance stack is becoming part of the team
The last generation of finance software largely helped people manage work. The emerging generation increasingly performs parts of the work itself: gathering data, reconciling records, preparing scenarios, handling collections, supporting procurement, drafting analysis, and surfacing exceptions.
That changes the build-versus-buy question. “Build” may no longer mean a major engineering program. It can mean a finance team using AI to prototype a small internal tool or agent. “Buy” may mean paying for usage, outcomes, or automated work rather than traditional seats.
The CFO therefore needs to understand not just software cost, but workflow economics: where automation creates leverage, where integration creates dependency, where controls must remain explicit, and whether the resulting system actually improves decisions.
Automation increases the importance of controls
There is an important counterweight. The more work software performs, the more important governance becomes.
An agent drafting a journal entry needs supporting evidence. An automated forecast needs traceable assumptions. A workflow changing payment information needs authorization. An AI system making recommendations needs defined decision rights. An automated action needs an audit trail.
This is where finance has an advantage. Controls have always been part of the CFO’s world. AI does not eliminate that discipline. It expands where it must be applied.
Automate what should be automated, preserve human judgment where it matters, and make the authority around both visible.
Continuous finance changes the management cadence
Traditional FP&A grew up around periodic planning. But when revenue, usage, cost, and operating data can be refreshed continuously, the forecast can increasingly become a live decision system rather than a monthly artifact.
That is particularly meaningful for 13-week cash forecasting. The value is not the spreadsheet itself. The value is the operating discipline underneath it: current billing information, collections, payroll, purchasing, project commitments, debt service, taxes, and the exceptions that affect liquidity.
AI can reduce the work required to assemble that information, but the CFO still has to decide what matters, which assumptions are credible, and what action the business should take.
The CFO as architect
I do not think the CFO becomes the CTO. Nor should every CFO become a software engineer.
But I do think the boundary between finance, technology, and operations is becoming much more porous. The CFO increasingly needs to understand APIs, automation, AI agents, data architecture, system integration, technology economics, permissions, and control design for the same reason CFOs learned capital markets, FP&A, and ERP systems in earlier eras:
Those tools increasingly determine how the business operates.
Finance still owns the judgment. But increasingly, the CFO also helps design the machinery that turns information into decisions.
The job is still finance.
The difference is that more of finance can now be built.
And I think the CFO who understands both the economics of the business and the architecture underneath those economics will have an increasingly important seat at the table.
That is where I see the CFO role headed, if we are not already there.