I spent two decades running labor-heavy services businesses. Co-founding Mynd and Waypoint Homes taught me a lot about building operations at scale, but the lesson that stuck most was also the most frustrating: in businesses where people are your primary cost of goods sold, the margin ceiling is structural. You can hire well, build good processes, and still watch capital get absorbed by the daily operational grind. Every budget cycle came down to the same tradeoff between protecting the operation and feeding the growth engine, and the honest answer is that the operation usually won.
What I wish I had understood earlier is that this is not really an efficiency problem. It is an architectural one. The businesses that will look fundamentally different in five years are not the ones that found a smarter way to manage labor costs. They are the ones that stopped building around human labor as the default unit of operational capacity.
The math has shifted
That is what agentic AI actually makes possible, and I want to be specific about the mechanism, because most of what gets written about AI in business stops well short of it. Autonomous digital workers execute multi-step workflows without requiring a person at every handoff. They handle exceptions, reconcile data across systems, manage communications, and make contextual routing decisions in real time. When you build operations around them rather than bolting them onto existing processes, the cost structure changes in a durable way. Labor expense converts to software cost, gross margin expands, and the cost curve flattens as the business scales. More importantly, the strategic options available to you change. Capital that used to be consumed by operational capacity can go toward growth, product, and market expansion in ways that were not previously available.
For finance operations, this reframe matters in a particularly urgent way. Most growing businesses accumulate operational debt long before they recognize it as such. Finance teams end up stitching together workflows across email, CRM systems, ERP platforms, invoicing tools, budgeting software, and spreadsheets layered with years of macros and manual exceptions. Contract terms get re-entered by hand. Data gets pulled from one system and keyed into another. Teams race to close books, reconcile liabilities, and collect cash under constant deadline pressure, while every new customer, product line, or business change introduces another edge case that requires human intervention. The issue is not that finance teams are inefficient. It is that the operational architecture underneath them was never designed to scale cleanly.
A pioneering advertising technology company managing roughly $2 billion in annual partner payouts illustrates what this looks like when the architecture changes. Nearly 80 percent of its revenue depended on 775 external partners operating across eight channels and 23 legal entities. But the underlying operation relied heavily on manual processes like contract terms re-entered by hand, Excel macros calculating roughly $150 million in monthly payouts, and fragmented workflows with limited auditability.
The company rebuilt the payout operation around interconnected digital workforces responsible for contract ingestion, payment calculations, liability tracking, payment triggering, and anomaly detection. The results speak for themselves: 99.999 percent calculation accuracy, 84 percent faster contract ingestion, and 97 percent faster calculations, with the operation ramped in under two months. More importantly, the company transformed a fragile finance workflow into a finance operation that could scale with the business instead. The same pattern is emerging in other finance-heavy workflows. One large insurance company automated accounts payable processing from invoice intake through ERP submission, where digital workers auto-resolve 63 percent of exceptions while maintaining full SOX-compliant audit coverage across thousands of monthly transactions.
What changes inside the operation
The capital question is where this shift becomes most concrete for finance leaders. When a meaningful share of labor cost converts to software cost, budget opens up that has been structurally unavailable for years. In my experience running these businesses, that budget rarely just drops to the bottom line. It funds the product investment that kept getting deferred, the market expansion that never quite penciled out, the customer acquisition capacity that the operation could not support. The businesses that are pulling ahead right now are redeploying that margin in ways that compound.
Getting there requires a specific kind of starting point. The operators I see making real progress begin by identifying the two or three functions in their back office that are highest in labor cost, highest in volume, and most repeatable in structure. They model one function before and after, and they run it in production rather than in a pilot environment. The learning is faster than most expect. The compounding across additional workflows is where the economics of the whole business start to shift.
The margin ceilings that have defined labor-heavy service businesses for a long time are not inevitable. They are a product of building operations around a cost structure that no longer has to be the default. The founders and operators who internalize that distinction early are building businesses that can scale operationally without scaling friction at the same rate.
By Colin Wiel, cofounder and CEO of Qurrent
