Full Episode 001 Transcript
Tim Pratt:
Welcome to the Tim Pratt Podcast. Today we’re diving into The CFO’s Biggest Misstep with Agentic AI — and why so many finance leaders are starting in the wrong place. We’ll walk through the 10 most important things every CFO needs to understand. Let’s dive in.
Why Most CFOs Start in the Wrong Place
Most CFOs are starting their Agentic AI journey in the wrong place. The instinct is to begin with forecasting and financial reporting because those feel like the highest-value areas, but according to recent research, this is usually a costly mistake.
BCG’s report ‘The CFO’s AI Agenda’ from May 2026 puts it clearly. “AI is advancing quickly in finance. The constraint is no longer capacity — but whether CFOs can build the data, processes, and skills to realize returns.”
And here’s the key line, “Most organizations try to layer intelligence onto processes that were never designed for consistency.” This is exactly why starting with forecasting and reporting is usually the wrong move. The complexity is high, the data is messy, and the results take much longer to show.
The Real High-ROI Starting Point
Instead of starting at the top with high-visibility projects, the smartest CFOs are beginning in the back office where processes are more rule-based and automation can deliver fast, measurable wins. Here’s a great real-world example from Coupa’s 2026 AP Automation Case Studies.
Take GameStop. They were manually keying every single invoice — all 750,000 of them each year. With recent acquisitions, the only way they could handle the increased volume was to keep adding more staff. So they deployed a comprehensive suite of Coupa solutions to automate and consolidate their global divisions onto one centralized platform — covering everything from sourcing and contracts through invoicing, expenses, and payments.
The results were outstanding: They achieved an 82% first-time match rate, reduced AP headcount by 20%, and cut average invoice processing time by 70%. This is exactly why back-office processes like procure-to-pay and invoice automation are the smartest place to start with Agentic AI. You get fast, measurable wins that build momentum for bigger initiatives.
The Hidden Cost of “Quick Wins”
Many teams implement automation and immediately cut headcount. But there’s a hidden cost to this approach. It creates fear across the organization, kills employee morale, and actually slows down real transformation. Quick wins can become long-term setbacks.
Fortune recently ran a very important article with this headline: ‘AI isn’t paying off in the way companies think. Layoffs driven by automation are failing to generate returns, study finds.’ The article explains that many companies assume AI will simply replace people and immediately cut labor costs. But a new study found that’s not what’s actually happening. In reality, many AI-related layoffs are not delivering the expected financial returns.
Then it gets to the most important part. Under the section “Where companies see returns with AI implementation”, one expert said: “That’s not where the value is” — referring to immediate layoffs. “That’s not where the productivity gains are going to be.”
This is the hidden cost of quick wins. When you automate something and immediately cut headcount, you create fear, resistance, and often slow down the entire transformation. The smartest companies are learning to be more patient — using early wins to upskill their teams instead of just reducing them.
The New Role of the CFO in AI
We’re witnessing a major shift. CFOs are becoming the true leaders of digital transformation — more than CTOs. Why? Because finance understands process, data integrity, controls, risk, and ROI better than any other function in the company.
Forbes recently captured this shift perfectly in an article titled: ‘From Record-Keeper to Rainmaker: The New Kind of CFO Leading the Agentic AI Revolution.’ It states: “For decades, the chief financial officer was measured by how well they controlled costs. In 2026, they are measured by how well they deploy intelligence.”
And according to a recent CFO.com article citing Oliver Wyman and the New York Stock Exchange, when CFOs were asked about the top forces driving change in their role, AI and digital transformation ranked #1 — cited by 66% of respondents.
This is a profound change. The CFO is no longer just the guardian of numbers — they are now the strategic leader of how intelligence is applied across the entire organization. This gives today’s CFO a more powerful and influential seat at the table than ever before.
Build vs Buy vs Bolt-On
When evaluating Agentic AI platforms, finance teams face three main choices: build it themselves, buy a full platform, or bolt agents onto existing systems. The key question is: how do you make this decision without creating even more fragmentation in your tech stack?
A recent Deloitte report titled ‘4 Shifts are Shaping Technology Infrastructure’ nails this challenge. It warns: “When technology modernization happens in pieces, leaders risk creating the next legacy system — locking into new constraints that limit what technology can do next.”
Later in the article, Deloitte drives the point home even further, explaining that bolting on new technologies without a clear architecture often leads to increased complexity, higher long-term maintenance costs, and reduced agility over time.
This is the real danger with the ‘Bolt-On’ approach. You add agents here and there to move fast, but you risk creating tomorrow’s legacy system today.
Governance and Explainability
The number one reason finance teams hesitate with Agentic AI is governance and explainability. Auditors, regulators, and boards all want to know: why did the agent make that decision? You must design governance and audit trails from day one.
A recent Houseblend article titled ‘AI Agents in Finance 2026: A CFO Guide to Reality vs Hype’ highlights this challenge in its executive summary. CFOs are rapidly reshaping their view of AI agents — sophisticated AI systems capable of autonomous decision-making — as they navigate an era marked by generative AI and ‘agentic’ models.
And right at the top of the barriers they identify: Governance and explainability remain the #1 reason finance teams hesitate with Agentic AI. Without clear audit trails and transparent decision-making, even the most promising agents face significant resistance from risk, compliance, and audit functions. This is not something you can add later. The most successful teams are designing governance and explainability into their Agentic AI systems from the very beginning.
From RAG to Compiled Knowledge
Traditional RAG (Retrieval-Augmented Generation) is no longer enough for complex finance workflows. We’re moving toward ‘Compiled Knowledge’ systems — where agents don’t just retrieve information, they truly understand your company’s policies, procedures, and institutional knowledge.
A recent VentureBeat article titled ‘The retrieval build: Why hybrid retrieval intent tripled as enterprise RAG programs hit the scale wall’ highlights this exact challenge. Look at this survey data. The percentage of companies planning to expand RAG into more workflows dropped sharply — from 43.1% in January down to 26.7% in March, a decline of 16.4 percentage points in just three months.
This decline is telling. Companies are realizing that simply retrieving information isn’t enough for the complexity of real finance work. They need agents that don’t just find data — they actually understand context, rules, and institutional knowledge. That’s the shift happening right now.
Usage-Based Pricing Is Coming
Usage-based pricing and unpredictable token spend is coming fast. CFOs need to start preparing now — new forecasting models, new controls, and a completely different mindset around AI costs.
A recent Fortune article highlighted a major opportunity here. The title says it all: ‘CFOs could cut agentic AI costs up to 60% by fixing this overlooked data problem’. It explains:
“In the race to deploy AI agents, many companies are overlooking a costly problem hiding in plain sight: data without context.” This is a critical point. Without clean, contextualized data, your token spend becomes unpredictable and inefficient. The companies that get ahead will be the ones that fix their data foundation now — before usage-based pricing really kicks in.
The Rise of the Accounting Engineer
We’re about to see the rise of a powerful new hybrid role: the Accounting Engineer.
This person combines deep finance domain expertise with the ability to design, manage, and optimize agentic AI workflows. This skillset will be extremely valuable in the years ahead.
A recent Deloitte report titled ‘Finance Trends 2026: Navigating the expanded scope of finance’ highlights this shift in “Trend 5: Infusing tech talent in finance — Where data scientists and accountants meet.”
According to the survey, finance leaders are actively looking for people who can bridge both worlds. Sixty four percent of respondents said at least one technical skill is a top development priority through 2026. Leading the list: AI and automation skills at 28%, followed closely by Data analysis and technology integration at 27%.
This is the emergence of the Accounting Engineer — a powerful new hybrid role that combines deep finance expertise with agentic AI capabilities. The teams that develop this skillset fastest will have a massive competitive advantage.
The Long-Term Competitive Advantage
Here’s the final truth: companies that rush to cut headcount after automation often lose in the long run. The real winners delay those cuts, focus on upskilling their teams, and reallocate talent to higher-value work. That patience creates massive competitive advantage.
A recent article in The National CIO Review titled ‘Companies Cutting Staff for AI See No Clear Gains’ backs this up with fresh research. According to a Gartner survey highlighted in the article:
- “Layoffs are common after AI implementation… but they don’t drive better results.“
- “Budget savings do not equal value creation.”
This is the upside. Companies that focus on upskilling over headcount cuts create more capable teams, maintain institutional knowledge, and position themselves to win in the Agentic AI era.
And that wraps up the 10 most important things every CFO should understand about Agentic AI right now. If you’re a finance leader working on Agentic AI, I’d love to hear what you’re seeing. Drop a comment below or reach out on LinkedIn. I’ll see you in the next episode.