The CFO’s Role in AI-First SaaS Leadership: In Conversation with Brad Wolfe

There is a version of the AI conversation that sounds clean, fast, and almost too easy. Companies will automate more work, reduce costs, move faster, and unlock new efficiencies across the business.

Brad Wolfe has spent enough time inside the machinery to be more cautious.

Wolfe brings a rare mix of CFO, operator, and systems experience to the AI conversation. He is currently Chief Financial Officer, Operating Partner, and CFO/COO for PE-backed companies through Wolfe Packs Consulting, and leads AI strategy and consulting for lower middle-market companies through Wolfe Packs AI.

His background includes senior finance roles across PE-backed SaaS, software, cybersecurity, hardware, and tech-enabled businesses, including Acre Security, Elm Street Technology, and FalconStor Software. Across his career, he has worked through 50 acquisitions, more than 80 integrations, and more than 20 ERP, CRM, billing, and data warehouse implementations, giving him a practical view of the systems and operating models AI is now expected to transform.

Today, working with three private equity groups, Wolfe is helping build and test an AI-enabled shared services model for smaller portfolio companies, using platforms including Salesforce, Zuora, NetSuite, and Snowflake as the operating backbone. As he described it in the interview, the goal was not just to create shared services, but to use that environment as “a testing ground for AI.”

But his view is notably different from much of the current market noise. AI, he argues, is not simply another software tool that makes existing work more efficient. It changes the operating model itself. And once that happens, the questions become much more serious than which tools to buy or where to run a pilot.

“You have to look at AI and look at your business and say, are you AI defensible? Are you AI resistant?” Wolfe said. “Before you do anything else.”

For SaaS companies, that question cuts to the centre of strategy. A company may be AI defensible because it has proprietary data, deep customer stickiness, regulatory protection, or a product that is so embedded in a customer’s workflow that AI is unlikely to displace it quickly. AI resistant is different. It may give a company time, but not immunity. The moat exists, but AI may still find a way around it.

That distinction matters because it changes the investment case. If a company is not defensible or resistant, then pouring money into AI may not solve the deeper problem. If it is, then the next set of questions becomes more operational: where can AI create advantage, what needs to change first, and who is actually accountable for making sure it works?

Why the CFO Is Moving Closer to AI

One of Wolfe’s more pointed views is that AI should not be left solely to technical leaders. In the portfolio companies he works with, AI is led by CFOs.

That may sound counterintuitive until he explains what AI actually touches. It changes the control environment. It affects corporate governance. It requires ROI discipline, baselines, roadmaps, risk management, data accuracy, and a clear understanding of where capital should be allocated. In Wolfe’s view, those are not side issues. They are the work.

“AI changes your control environment,” he said. “It makes corporate governance much, much more important.”

This is especially true in lower middle-market SaaS companies, where the CFO often already carries responsibilities that might sit with a COO in a larger company. Sales, operations, and finance have traditionally run as separate models, often at different speeds and with different owners. AI compresses those models. It connects data, workflow, decision-making, and execution in ways that make it harder to keep functional boundaries clean.

Wolfe believes this will reshape the C-suite. The CRO may become more focused on sales. The CMO may become more focused on marketing. The CHRO may become more focused on people. But the analytical function, in his view, becomes more centralized, and much of it may sit closer to finance.

The reason is not that finance is more important than every other function. It is that finance is where judgment, controls, compliance, and business model discipline come together. Revenue recognition still matters. Debt covenants still matter. Auditability still matters. A beautiful AI-generated output is not useful if no one can verify whether it is correct.

That reality changes how companies should think about AI ownership. The issue is not simply who is most excited about the technology. It is who can be trusted to understand the business consequences when it breaks.

The Work Moves, It Doesn’t Disappear

Wolfe is also careful not to describe AI as a straight cost-reduction story. In fact, he pushes back hard on the assumption that AI will simply replace large portions of the workforce, especially in the lower middle market.

In the shared services model he helped build, employees were told that the company would use AI extensively, but would not lay people off because of it. The commitment came with a condition: people had to be willing to retrain, do new things, and work with the company through the transition.

The bigger change, Wolfe said, is that work shifts from the back end of the funnel to the front end. People who once cleaned, organized, verified, and controlled information after the fact are now needed earlier in the process. They are still doing work rooted in accuracy, control, and judgment, but from a different seat.

The reason is simple. AI is only as useful as the information going in and the verification around what comes out. Wolfe believes the “control surface” in an AI-native environment is much broader and deeper than in a legacy company. That makes people with strong finance, governance, process, and data integrity instincts more important, not less.

That does not mean every employee will adapt easily. Some people have built careers doing a job a certain way and may resist being pulled into a more forward-looking or analytical role. Wolfe is pragmatic about that. Companies can provide training, runway, and support, but they cannot avoid the reality that some roles and mindsets will need to change.

His broader point is that AI does not remove the need for human judgment. It increases the need for it in different places.

“There’s no point in having AI produce 10,000 things an hour if we can only review a hundred an hour,” he said.

That sentence captures one of the practical realities many companies are discovering. Output is not the same as value. More code, more analysis, more content, more reports, or more recommendations can create new bottlenecks if the organization is not designed to absorb them. AI may accelerate one part of the system while overwhelming another.

The Real World Is Not a Lab

Wolfe is skeptical of the idea that AI has already been solved at scale. He has seen promising lab use cases and narrower process improvements, but he has not seen many companies solve the full operating problem.

That is partly because companies are messy. Data is messy. Customer inputs are inconsistent. Legacy systems were not built for AI-native workflows. Even a state-of-the-art stack with Salesforce, Zuora, NetSuite, Snowflake, and strong process discipline may still not be clean enough for AI to work at the speed and reliability companies imagine.

“What AI requires is absolute pristine process, procedure, systems management, org structure, controls,” Wolfe said.

The danger is not just that AI gets something wrong. It is that it gets something wrong in a way that looks plausible, moves quickly, and is difficult to explain. Wolfe described sandbox testing where AI behaved unexpectedly and the team spent days trying to understand why.

His analogy is memorable: AI can behave like “a two-year-old with a 10,000 IQ.” It has extraordinary pattern recognition, but not the broader world model or common sense that people often assume is there.

That is why Wolfe is wary of companies unleashing agents across the business without strong controls. To him, that can be like “starting fires in your house and hoping it works out well.” Prompts, policies, and enthusiasm are not the same as control. Nor is it enough to buy licenses and tell employees to do something productive.

He sees a middle path. Companies should not ban AI outright because employees will find ways to use it anyway. But they also should not give everyone open access to tools without training, closed environments, policies, and a clear methodology for identifying use cases. The goal is not to teach everyone to write clever prompts. It is to teach people how to identify problems AI may help solve, while protecting the company from shadow AI, data leakage, inconsistent analysis, and self-inflicted control issues.

That distinction matters for CFOs and boards. In private companies with debt, and especially in public companies, leaders are still accountable for the accuracy of what they sign. “We thought AI was doing it the right way” is not much of a defence.

AI Plus People

What makes Wolfe’s view stand out is that it is neither anti-AI nor blindly optimistic. He believes AI is powerful. He believes companies cannot ignore it. He also believes the current market has done a poor job of acknowledging the human and organizational cost of change.

He is especially critical of the idea that AI can devalue a lifetime of skill and replace it with vague promises about stipends, retraining, or a better future somewhere down the road. People do not live in the long run. They have mortgages, careers, identities, and expertise built over decades.

For Wolfe, the better model is “AI plus people.” Not AI as a replacement for people, and not people pretending AI is just another productivity tool. The human side remains in charge. AI can improve pattern recognition, research, analysis, finance presentation, and process insight, but it does not have a point of view. It does not understand the business model. It does not know where value is unless people define that clearly.

That may become one of the more important leadership tests in PE-backed SaaS. The winners will not simply be the companies that adopt AI fastest. They will be the companies that understand where they are defensible, clean up the unglamorous infrastructure, put real controls around use, and redesign work without hollowing out the judgment that makes the business valuable.

The talent implication is just as important. Companies will need leaders who can work across finance, operations, technology, product, and people. They will need executives who can ask better questions, not just approve new tools. They will need teams that are willing to retrain, but also boards and management teams disciplined enough to define what AI is for.

Near the end of the conversation, Wolfe described the current moment as both fun and terrifying. Like being on a roller coaster. That may be the most honest description of where many SaaS leaders are right now.
The ride has already started. The real question is whether the company is securely on the rails.

About the Lancer Group

The Lancer Group helps PE-backed SaaS and software companies find leaders who are ready for AI, not just fluent in the language of it. As AI reshapes products, operating models, team structures, and value creation plans, companies need executives who can separate signal from hype, ask better questions, and turn new capabilities into measurable business advantage. The right leader does not simply understand AI as a tool. They understand how it changes the way software companies build, sell, scale, and compete.