Green Sheet interviews Toffer Grant, PEX founder and CEO

As organizations face continued economic uncertainty and mounting pressure to improve efficiency, finance leaders are taking a more disciplined approach to AI investment. CFOs are increasingly focused on measurable operational gains, stronger risk management and practical tools that reduce manual work while supporting long-term financial sustainability.

Toffer Grant, founder and CEO of PEX, has a front-row seat to how finance teams are navigating that shift. PEX provides spend management and expense tools designed to streamline finance operations, including technology that allows employees to submit receipts by text message while automatically matching receipts to transactions behind the scenes.

In this Q&A, Grant discusses how finance leaders are evaluating AI investments, where companies are seeing the clearest returns, and why many organizations are moving away from broad experimentation toward more targeted, operationally focused use cases.

Green Sheet: Many companies are investing heavily in AI, but how can CFOs distinguish between meaningful ROI and expensive experimentation?

Toffer Grant: The starting point should be the business problem a CFO wants to solve, not the technology itself. The strongest gains to be had from deploying AI address a specific operational challenge and can be tracked with measurable outcomes. It's best to think small at first and experiment with employees to confirm adoption.

The finance team will need to see the data flow into their tools and confirm that information is coming through clearly. From there, layer on more complex goals and match the tools.

Meaningful ROI comes from connecting AI investments to business results. That could mean improving decision-making, accelerating workflows, increasing productivity or helping teams focus on higher-value work.

Good examples start with high-frequency tasks, such as receipt capture and transaction notes – both important but a low stakes entry point for employees and for finance with big potential for immediate efficiency gains.

It also depends on having reliable data and a clear plan for how success will be measured. Are receipts flowing into finance more quickly? Are they documented properly? Many organizations are still approaching AI as a broad innovation rather than targeted operational investment.

At the end of the day, CFOs should evaluate AI the same way they evaluate any major investment. What process are we improving? How will we measure success? And what changes in outcomes should we expect? If those answers aren't clear, it becomes much harder to separate long-term value from short-term experimentation.

GS: What are the biggest mistakes finance leaders make when evaluating AI-related spending and performance?

TG: AI-related spending is measured in tokens and time. Evaluating LLM tools like Claude and ChatGPT is specific to use cases, trust and what users want to gain from them. As frontier models are released, the latest and greatest tend to chew up tokens and budget more quickly.

Finance teams that are first starting to learn about AI tools should look to their existing providers for new AI capabilities. Many platforms are rolling out AI-enabled tools as free add-ons or enhancements to their platforms – a really easy and cheap entry point.

As users become more sophisticated and are comfortable using MCP tools, added benefits come from hooking an LLM to their existing tools for AI-generated analysis. Think: Hooking up QuickBooks desktop to Claude for financial insights.

Anytime Claude does work, it uses tokens and incurs cost. Being aware of this helps with budgeting experiments and avoids large bills. The key takeaway is: play with existing chatbots and tools that are free and embedded in existing tools and then start linking LLMs to them for enhanced experiences.

One common mistake is focusing too much on the technology itself rather than the operations it can support. The conversation should start with the problem being solved, the process being improved or the outcome the organization is trying to achieve. Without that foundation, it becomes difficult to determine whether an investment is actually creating value.

Another challenge is focusing too narrowly on immediate returns. Some benefits, such as productivity gains and reduced manual work, can be measured relatively quickly. Others emerge over time through better decision-making, improved forecasting, stronger resource allocation, and a greater ability to identify trends and risks earlier.

Finance leaders need a balanced approach to measurement. Looking only at short-term efficiency gains can understate the value of an investment, while focusing only on long-term potential can make it difficult to demonstrate accountability. The most effective evaluations consider both operational impact today and strategic value over time.

GS: In the payments industry specifically, where are AI investments producing the clearest operational or financial returns today?

TG: In the payments industry, some of the clearest returns are coming from AI's ability to analyze transaction data at scale and save employee time spent on tasks they may view as low priority but that remain important to finance. Organizations are using AI to identify anomalous transactions, detect potential fraud, monitor spending patterns and surface issues that would be difficult to catch through manual review alone.

We're also seeing strong results in spend management and transaction oversight. AI can help finance teams automatically categorize transactions, flag potential policy violations, and streamline reconciliation and reporting processes. These are areas where even small efficiency gains can add up quickly because of the volume of transactions many organizations manage.

The common thread is that AI helps teams move from manually reviewing payments and transactions to proactively managing exceptions and insights. That reduces operational costs, improves visibility into spending, and allows finance teams to focus more on analysis and decision-making rather than administrative work.

GS: How are payment companies balancing the promise of AI-driven fraud prevention and automation against concerns about cost, regulatory exposure and risk?

TG: Most payment companies recognize the potential of AI, but they are approaching adoption with a clear focus on governance and risk management. The strongest use cases today focus on fraud detection, transaction monitoring, customer support, and operational workflows where AI can improve speed and efficiency without introducing unnecessary risk.

Any AI system that influences financial decisions, transaction reviews or customer interactions needs to operate within established controls and compliance frameworks. Organizations are paying close attention to transparency, pattern recognition, auditability and human oversight, particularly in areas where errors can have financial, customer or regulatory implications.

Many are taking an incremental approach by deploying AI in targeted workflows, measuring results and expanding from there. That allows teams to capture operational benefits while maintaining the visibility and accountability that are essential in payments. AI helps speed up review processes, identify risks faster and automate routine tasks around the clock.

GS: Are CFOs becoming more skeptical of broad AI initiatives as economic uncertainty and budget pressure continue to grow?

TG: I wouldn't describe it as skepticism so much as a shift toward greater accountability. Early in the AI adoption cycle, many organizations were focused on exploring the technology and understanding its potential. Today, finance leaders are asking more detailed questions about business impact, implementation costs and measurable returns.

Economic uncertainty tends to raise the bar for any technology investment, and AI is no exception. CFOs are looking for solutions that can improve efficiency, reduce manual work, strengthen decision-making or help teams operate more effectively. Broad promises alone are no longer enough to justify investment.

In many ways, this reflects a more mature phase of AI adoption. Finance leaders continue to see significant opportunity, but they are prioritizing use cases with clear outcomes and a realistic path to value.

GS: As companies push to "do more with less," how are finance leaders balancing innovation with risk management and long-term financial sustainability?

TG: Innovation, risk management and long-term financial sustainability all exist symbiotically; they are interconnected parts of running a responsible business. Innovation creates value when it helps organizations operate more efficiently, make better decisions and adapt to changing business conditions. Risk management helps ensure those benefits can be realized consistently and responsibly over time.

Both of these feed into long-term financial sustainability, which in turn creates more future opportunities to innovate and manage future unknown risks.

That balance is especially important in the current macroeconomic environment. Finance leaders are under pressure to improve productivity and profitability, but they also have to protect the business from decisions that create complexity or risk down the line. At the end of the day, finance leaders are focused on creating durable value and building an organization that can remain resilient, adaptable and financially healthy regardless of what comes next.

Notice to readers: These are archived articles. Contact information, links and other details may be out of date. We regret any inconvenience.