When lean teams collide with rising invoice volumes, leaders have a choice: slow down or scale up. In a recent CFO Leadership Council webinar, Ancestry and Blue Sky Farms showed what it looks like to choose the latter: pairing process expertise with domain-specific AI to turn accounts payable (AP) from a bottleneck into a growth-ready engine.

Why AI, and why now? 

For Ancestry, the push toward AI started with constraints and a clear-eyed view of what wasn’t working. 

“We run our teams pretty lean, and I think most people in finance and accounting can relate to that. There’s pressure on getting as much work done as we can with the fewest number of people,” said Randy Luck, Senior Manager of Accounts Payable and Travel. Their existing OCR-based solution had low accuracy and forced the AP team to review every single invoice, creating delays. Plus, the system could only handle limited document formats. The goal for Luck shifted from making just incremental tweaks to implementing a bigger technology and process change: “something quicker, more accurate, more reliable.”

For Blue Sky Farms, the AI catalyst was scalability. 2,500 invoices were already being entered each month manually, and with one of their major facilities doubling in herd size, that number was about to climb.

“[Do] you hire more employees, or do you deploy a tool that you can work with the same amount of employees, but get more done?” asked Reece Blaylock, IT Manager. Vendor trust was also central to the change: “We want to keep our vendors happy. When they send their invoice over, it's getting paid, and they're gonna get paid timely.”

Where they started, and why AP was the right entry point 

Both teams chose AP as the proving ground for applying AI because of the concentration of manual work, opportunity for measurable outcomes, and there was a real risk if things slipped: late fees, missed discounts, and strained vendor relationships can all negatively impact the bottom line. Ancestry anchored on building trust in AI prediction capabilities, so the teams didn’t have to scrutinize every transaction. “How do we get that trust to where we know the data is coming over and we don't have to look at it?” While Blue Sky Farms focused on cycle time and capacity — making sure scaling operations didn’t require scaling headcount.

Early proof points: from “will this work?” to “what’s next?” 

Within months, results were tangible for both organizations, and momentum continues to build across teams.

• Invoice accuracy jumped at Ancestry: “With the previous OCR, we were getting somewhere in the 60% range for accuracy. But when we switched to Vic.ai, I don't think we've had a month that's been lower than 93-94% accuracy.” That accuracy has reduced rework and increased confidence, especially across their 10+ global entities with varying AP requirements.
• No-touch processing also climbed steadily: “Today, we're at 72% for this month. We’ve been with Vic.ai for about six months now, and each month that has gone up.” These straight-through processing gains freed more capacity and validated the phased rollout.
• Invoice cycle times dropped at Blue Sky Farms: “We went down to three days to process an invoice; before that, it may have been five to seven business days.” This shift has reduced late fees and improved the capture of early payment discounts — the exact outcomes finance cares about.
• The team’s mindset also changed: “At first [staff] were scared that it’s going to replace [them], but now they're heavily reliant on it and want to use it, and they're excited to use it,” said Blaylock. After the first entity went live, “you could see the excitement: ‘Oh my gosh, we were done doing manual entry!’”

How they made it work: people, process, and precision 

Neither organization treated AI as a magic wand. They treated it like any high-stakes technology transformation — one that lives or dies by process clarity and people’s trust.

Luck’s approach started with listening: “Listening to the processors and saying what problems do you have; what can we do with this tool to make your life easier.” From there, he worked with IT to define the process, data requirements, and handoffs. “[We worked] with our IT team to really define the process: what we needed the AI to do and how we were going to communicate from our system to the AI system.” Clean vendor, PO, and GL data and clear project requirements reduced friction.

Blaylock distilled his playbook to two principles: prioritize creating process experts and scope the project tightly. “Find the expert or be the expert — not in AI but in the process, and know what the end results should look like and how the process is supposed to flow.” Also, don’t think one AI agent or application will do everything. “We wanted an AP automation agent, and now if we need to go build a separate AR agent, that’s doable. The more you specify what you want an agent to do, the better results you’re going to get.” That precision made requirements clearer, UAT faster, and success criteria obvious. His other mantra: “UAT, UAT, UAT.”

What they measured, and why those KPIs mattered 

Ancestry oriented their KPIs around accuracy and touchless processing, then layered controls for a multi-entity environment. “Accuracy, straight-through processing, and with 10+ entities globally, [meeting] audit and SOX requirements with rules and controls.” AI also removed repetitive steps like VAT entry: “Checking it is much faster than us having to manually enter all of those, and [Vic.ai] gets progressively better and better over time.”

Blue Sky Farms tied process metrics to financial impact. “Accuracy of the invoices [is] driving processing time and costs down,” said Blaylock. That “really drove our late fees down” and improved early payment discount capture. Their new baseline: invoices “from receipt to ready to pay” in roughly three days.

Lessons learned: what they’d do again

• Expect surprises and iterate calmly. Blue Sky Farms uncovered ERP customizations affecting other modules. “Don't be upset when something small goes wrong. Fix them appropriately.” Cross-functional collaboration sped fixes.
• Align data and ownership early. Ancestry emphasized defining what data moves where — across AP, IT, the ERP, and Vic.ai — before build. “Understand and define [data needs] before you get into implementation.”
• Use history to accelerate learning. “The one thing I really like that Vic.ai does is look at historical data, so that when you go live, there's already a head start.”

The human impact and the mindset shift 

The most striking change wasn’t only in dashboards; it was in the day-to-day. Team members who feared replacement now advocate for AI expansion. Leadership framing also helped the Ancestry team: “They communicated that [AI is] not a tool to replace our jobs, but to [make] us more efficient and make our jobs easier… ‘lean into this,’” said Luck.

What’s next: from AP wins to an autonomous finance roadmap 

For Ancestry, AP is the launchpad for AI use. “How do we now use AI to help with other functions within the finance and accounting group… journal entries, month-end close, reconciliations, cash forecasting.” 

For Blue Sky Farms, the focus is on leveraging agents that act, not just answer. Blaylock is already building agents that “automatically get data, act on that data more timely, answer questions, and hopefully save time and reduce headaches.” And there’s a new default question in every project review: “Every time we talk about a new project, the first question is, can AI do it? If not, do we need to create a [standard operating procedure] to do it?”

Top 5 webinar takeaways for finance leaders

1. Start with AI where the pain is clear and the wins are measurable. AP is often the right first move.

2. Anchor on a few KPIs: accuracy, no-touch rate, cycle time, late fees/discounts, and hours returned to the team.

3. Put process experts in the driver’s seat; pair them with IT early to define data flows and controls.

4. Scope narrowly and test relentlessly; invite processors into UAT to build trust and speed adoption.

5. Treat AI as domain-specific agents, not general-purpose tools. Precision beats breadth in finance workflows.

Watch the complete webinar recording to hear first-hand how Ancestry and Blue Sky Farms structured their pilots, aligned stakeholders, and scaled AI in AP.