AI is changing the business model, not just the tech stack
The following is a guest post from Dean Quiambao, partner at Armanino. Opinions are the author’s own.
For the past two years, companies have been racing to adopt artificial intelligence. Employees have been encouraged to experiment, departments have launched pilots and businesses have invested heavily in tools promising to make their organizations more productive.
But as CFOs begin preparing their 2027 budgets, a more important question is emerging: What has all this experimentation actually changed about the business?
The answer should go well beyond faster reporting, automated workflows or reduced administrative costs. AI is beginning to change the fundamental economics of how companies operate, how they allocate resources, and how they deliver services.
And that means one thing: CFOs need to start thinking differently about their role in shaping the business.
Experimentation is giving way to accountability
Not long ago, the prevailing advice was simple: Get everyone using AI. Give employees access to the technology and let them discover what it can do.
I think it was worth moving fast, because companies needed to experiment to understand the technology’s capabilities. But it also created an environment where employees pursued disconnected projects and AI spending grew without clearly defined business outcomes.
A recent moderated panel discussion with Pranav Sood, CFO of Airwallex, shows a new phase of approaching AI investment.
Rather than allowing unlimited experimentation, Airwallex has narrowed its focus to a defined pipeline of initiatives, prioritizing the projects most likely to deliver meaningful results.
One particularly interesting consideration is the trade-off between spending on AI tokens and adding employees. For CFOs, that’s a fundamental shift in how we think about resource allocation. AI has become an operating expense that must be managed as diligently as headcount or other parts of a business’s infrastructure.
If 2026 was the year of AI experimentation, 2027 needs to be the year of accountability.
Stop automating processes that shouldn’t exist
One of the biggest mistakes I see companies making is treating AI like every other technology upgrade. They take an existing workflow and ask how AI can perform those same tasks faster. But what if the workflow itself no longer makes sense?
Consider the traditional month-end close. Instead of asking how AI can accelerate each individual step, finance leaders should consider whether information can be reconciled and validated more continuously, reducing the need for the same manual process altogether. This is a fundamentally different way of approaching technology.
At our recent panel, we discussed how automating an inefficient process can simply reproduce the same problems at greater speed. The real opportunity comes when companies start with the desired business outcome and work backward to determine how people, data and technology should come together to deliver it.
This requires CFOs to move beyond evaluating technology purchases and become more actively involved in redesigning how their organizations operate.
AI is changing the economics of professional services
We’re already seeing this transformation unfold across finance and accounting.
Companies such as Rillet and Ramp are challenging assumptions about what financial software can do and how quickly organizations can adopt new operating models.
In particular, I’ve watched the conversation around Rillet change dramatically. Not long ago, executives were asking whether AI-native accounting platforms could replace basic bookkeeping software. Today, the questions are much more sophisticated. Can these systems support larger enterprises and the revenues that come with them? Can they meet the reporting, internal control and compliance demands of public companies?
That change in expectations is significant. Professional services firms will need to reconsider how they deliver value when AI takes on more technical and administrative work. Business models built around the time required to complete routine tasks may face increasing pressure as technology reduces the labor involved.
But that doesn’t mean human expertise becomes less valuable. In many ways, the opposite is true.
AI can help prepare financial analyses, reconcile transactions and identify patterns. What it cannot independently replace is the judgment required to sit across from another executive, understand a company’s challenges, negotiate competing priorities and determine the right course of action.
For CFOs, this raises questions about how organizations develop talent, structure teams and reward employees. The value of a finance professional may increasingly depend less on how much work they can process and more on the business decisions they can help improve.
Connecting businesses, not just systems
In separate but recent conversations with executives from Rillet, Tabs and Human Interest, I heard leaders say the next generation of AI will require agents that can communicate and work across different companies’ platforms. That has implications for the way audits and tax returns are handled.
Today, completing that work often requires significant coordination between clients, accounting firms and multiple technology systems. Now imagine AI agents that can securely coordinate parts of that process across those systems, with professionals maintaining oversight and judgment. That changes the way collaboration takes place and how firms price their services. It also means firms have to know where their competitive advantage comes from in this paradigm. Standing out is more than a case of adopting the most sophisticated AI tools. Tech, proprietary data and professional expertise have to come together in a more effective way to give customers what they need.
What CFOs should be doing now
As organizations finalize their 2027 budgets, CFOs have an opportunity to bring greater discipline to AI adoption.
Start by identifying the business problems worth solving rather than searching for opportunities to deploy technology. Determine which initiatives can meaningfully improve the operating model, and establish clear accountability for their execution.
Then assess whether the organization has the infrastructure to support those investments.
Does the company need to replace existing systems or integrate AI capabilities into them? Is its financial data sufficiently organized and reliable? Does it need specialized talent, such as a finance engineer, to connect systems and build scalable capabilities?
These are no longer questions that should be left entirely to the technology department.
CFOs also need to understand how their cost structures may change as AI spending becomes more closely tied to usage and how the organization should balance those investments against traditional labor and technology expenditures.
For companies that haven’t moved aggressively into AI, there’s still an opportunity. Early adopters have learned valuable lessons about what works and what doesn’t. Finance leaders can use those lessons to move forward with greater intention.
The CFO’s role has always evolved alongside the needs of the business. Today, that evolution is accelerating. CFOs must seize the opportunity to help design the operating models that will determine future performance.
