I'm done with AI. CFOs have heard enough about AI’s promise
I’m done with this BS: Here are 3 AI use cases with ROI.
I’m done with this BS:
AI is the future. AI will transform finance. AI is coming.
95% of AI projects fail because they’re too broad, too experimental, or too disconnected from business needs. It’s all just hype without ROI.
But there’s a 5% rare finance AI project that succeeds.
These are boring but essential finance workflows made faster, cheaper, and better.
In today’s edition, I’ll share with you 3 cases where AI is already producing real ROI in the CFO’s office.
Let’s dive in.
Shopify's VP of Finance Told me How They Are Using AI to Change FP&A
Shopify’s VP of finance, Jason (great guy), loves reading this newsletter and shared that the themes resonated with the shift he’s leading.
But while most companies are stuck in the "what if" stage, Shopify, the leading global commerce company powering millions of independent brands, is deep in the "how to" phase. They are actively building the future of finance, and Jason gave us an unprecedented look under the hood.
One of the team’s key responsibilities is analyzing the weekly performance of their merchant funding product, Shopify Capital.
The Old Way: For several days, multiple analysts painstakingly navigated Looker dashboards, delving into segmentations, cohorts, and other splits to construct the weekly funnel's narrative. This was a slow process, as extracting genuine insight proved time-consuming.
The New Way: They built an AI agent.
This isn’t a simple chatbot. It’s a sophisticated system designed to replicate and accelerate the entire analytical process.
Here’s how it works:
The Agent's Brain (Context): The agent was given access to three critical sources of information.
The Data: Direct query access to the core BigQuery tables.
The Product Context: Access to all internal product documentation. The agent knows when features were launched and what changed in the product.
The Human Process: The team documented their exact analytical workflow, every step an analyst would take, and fed it to the agent.
The Result:
The impact is profound. When the FP&A team gets into a meeting with product and engineering leads, the discovery phase is already done. The conversation starts at a higher level:
"So, what should we do?"
You’ll find the full interview useful.
Now, let’s move from e-commerce to enterprise scale.
Here’s how Microsoft rewrote the finance capacity mode
Microsoft: When Finance Work Stopped Scaling
Cory Hrncirik, Modern Finance Leader at Microsoft, described a familiar problem:
procurement cycles were drowning in manual sourcing and vendor analysis.
Collections were reactive. Forecasts lagged reality.
Every month, the finance team pulled together data from across the company. But by the time forecasts were finalized, they were already stale. Vendor sourcing took weeks. Collections slipped through cracks. The workload only grew, but headcount couldn’t.
Then Microsoft made a bold bet to use AI agents to take over entire workflows.
Instead of just layering automation on top, they built specialized bots.
A sourcing assistant for procurement.
An analyst agent for forecasting, even collection bots.
What once dragged on for weeks collapsed in real time.
Forecasts updated dynamically. Collections turned proactive.
The impact was superb:
One AI-driven sourcing agent alone saved $10M annually.
Finance freed up 15,000 hours a year.
Overall, the push unlocked thousands upon thousands of hours as well as millions of dollars, in Hrncirik’s words.
For Microsoft, this wasn’t about trimming a few minutes here and there.
It was about bending the headcount curve.
Handling far more workload without adding more people.
AI didn’t just streamline tasks. It rewrote the capacity model of finance.
The IRS: When Bureaucracy Met Bots
The IRS is famous for paperwork.
For CFO Teresa Hunter’s team, it was a daily crisis.
Her finance staff slogged through months-long processes:
Modifying contracts.
Reformatting documents.
Moving data between systems.
Backlogs piled up. Morale sank. Everyone knew the system was broken, but fixing it seemed impossible in the government’s legacy environment.
In 2021, Hunter tried something radical.
She brought in bots.
Using UiPath, her team deployed RPA with AI data extraction.
1,500 contract modifications that once took nearly a year? Done in 72 hours.
30,000 labor hours saved in just 3 years.
Error rates fell. Employee morale rose.
Hunter’s philosophy was simple.
Start with the ugliest pain points, automate them first, then scale. By attacking what staff hated most, she not only unlocked capacity but also built trust.
The result was more than efficiency.
For one of the largest bureaucracies in the world, finance processes that once defined red tape suddenly became a showcase for modernization.
Why These Projects Worked (and Most Don’t)
Common threads emerge from Shopify, Microsoft and the IRS.
Clear Pain Point: Each tackled a specific, costly bottleneck.
Process Readiness: Workflows and data were structured enough for AI to plug in.
Domain-Specific Tools: These weren’t generic pilots. They were targeted solutions built for finance.
Human-in-the-Loop: Teams treated AI like a junior assistant, not an oracle.
Executive Sponsorship: CFOs and finance leaders owned the project and measured outcomes in dollars, hours, and days saved.
Never chase hype.
Solve boring, painful, measurable problems first.
Identify the pain in hours and dollars.
Start with one workflow (AP, close, forecasting).
Clean your data and process it first.
Choose proven finance-specific tools.
Embed AI into daily workflows.
Keep humans in the loop.
Track ROI relentlessly (time, cost, error rate).
Celebrate and expand the wins.
Big News
The AI CFO Office community is stepping into the real world.
Our first in-person gathering will be in Dubai.
It’s intimate, invite-only, and designed for finance leaders shaping the future with AI.
You can join us. Apply for an invite.
The Bottom Line
95% of AI projects fail. But the 5% that succeed prove the ROI is real:
Microsoft saved $10M and 15,000+ hours a year.
The IRS cut a year’s work to 72 hours.
Shopify’s VP of finance changed FP&A
The hype is over. The question is simple:
Will your finance team join the 5% who deliver?
That’s all for today.
See you on Thursday!
Whenever you're ready, there are 2 ways I can help you:
Promote your business to 24,000+ CFOs and senior finance leaders by sponsoring this newsletter.
If you’re building an AI-powered CFO tech startup, I’d love to hear more and explore if it’s a fit for our investment portfolio.
Advertise with the AI CFO Office.
I’m Wouter Born. A CFOTech investor, advisor, and founder of finstory.ai
Find me on LinkedIn