Editor’s Note — This article is sponsored by BILL. As with all sponsored content in Fintech Takes, this article was written, edited, and published by me, Alex Johnson. I hope you enjoy it!
Every day, small business owners have to act like CFOs, controllers, and procurement managers before they ever get to do … the thing that motivated them to become entrepreneurs in the first place.
The florist initially set up shop to design bouquets, not reconcile expense reports — just as the controller at an up-and-coming local franchise didn’t sign up to chase invoices. Even the independent accountant doesn’t want to do accounting for their own business; they want to do accounting for their customers.
For years, the industry’s answer to this problem was to go faster. Automate workflows. Streamline approvals. Shrink the time between “to-do” and “done.”
But what if the real breakthrough isn’t speed?
What if it’s subtraction?
From Digitization to Overload
A generation ago, running a business meant spending Saturdays in the office above the hardware store, surrounded by invoices and manila folders that never stayed in a neat pile for long.
Then came the first wave of software in the late 1990s and early 2000s, promising to make the small business back-office work less painful.
And it worked! Software for small business owners, first delivered via CD-ROM and then later via SaaS, significantly reduced the amount of time small business owners and their employees spent on back-office tasks and gave them more flexibility as to where and when they got those tasks done.
The cure for paperwork was software. However, as the software multiplied, it created a new problem.
I’m not sure if this stat that I’m about to share it is technically correct, but it sure is directionally correct and I’ve loved it ever since someone told it to me:
The average small business has zero employees and ten SaaS subscriptions.
Using software, we solved one problem (the inconvenience of going into the literal back office), but created another (fracturing the back office across numerous, purpose-built software tools).
In order to solve this new pain point, we did what we always do in software.
Rebundle
Over the last decade, the market has tried to fix this problem through rebundling.
Whether you’re a horizontal platform like Intuit QuickBooks, or a vSaaS platform like Brightwheel, engineered to serve daycares, the idea is the same: bring everything a small business owner needs to run their business into one hub. One operating system.
This represents an important step forward, but it’s not the end of the journey for small businesses.
Why isn’t it, you ask?
Because unlike large enterprise companies, small businesses will never be able to afford a robust, well-staffed back office operation.
Enterprise companies will always have a back office because optimization is part of their DNA. At a big enough scale, even the smallest improvements matter. Finding an extra 75 basis points of performance lift (in cost, profitability, efficiency, etc.) justifies investing in the headcount necessary to get it.
Small businesses are different. There’s no competitive advantage in managing an efficient back office when you barely have one. For them, the back office isn’t a strategy; it’s a tax.
The paradox is that the companies most in need of sophisticated cash flow management are the least able to staff it.
For Walmart, better cash management leads to incrementally better margins. For a local pizzeria or HVAC installer, better cash management is often the difference between surviving as a business and not.
That imbalance is what makes this moment so interesting.
Do It For You
Every big technology shift has its own logic.
In the early era of small business software, the rallying cry was “do it yourself.”
Then came “do it with you” in the form of rebundled software platforms designed to shepherd small businesses through payments, invoices, and approvals.
Now, we’re entering the “do it for you” phase, powered by LLMs and agentic AI.
It’s a deceptively simple idea with big implications.
Few people understand that transition better than René Lacerte, the CEO and founder of BILL, who has spent the past two decades thinking about how to close the gap between the Fortune 500 and what he calls the Fortune 5 million (the millions of small and midsize businesses that form the backbone of the U.S. economy).
“The notion we really gravitate toward,” he told me, “is helping the Fortune 5 million have the same capabilities as the Fortune 500 for a fraction of the price. Eliminating friction and unlocking creativity; that’s the focus.”
LLMs are not yet ready to run an entire business on their own. Given their probabilistic nature, they may never be.
So the question becomes: where can we start to use them to help small business owners get better outcomes without requiring more time (or more SaaS tools) to invest into the back office.
The answer is to begin with the kinds of tasks that are time-consuming but clearly defined, like reading receipts or pulling key details and matching them to meetings on a calendar. The goal isn’t perfection; it’s consistency. If an AI system can get 90% of those tasks right, humans can focus on the 10% exceptions that actually require judgment.
“You have to take baby steps,” René told me.
One new AI agent BILL’s team has released can collect and pre-validate W-9 forms directly from suppliers, automating what would otherwise be say, three months of work. Another AI agent can categorize receipts automatically, surfacing only the ones that need a human eye.
“It’s not about accelerating workflows,” he said. “It’s about getting rid of them. The goal is to eliminate friction. Instead of ten steps, make it two.”
That shift, from speed to subtraction, is what separates this next generation of small business software from those that came before.
And it only works when LLMs have the context and scale necessary to learn the nuances and understand the edge cases in the jobs they are being asked to do. This is why many market observers speculate that incumbents may have an advantage, relative to startups, when it comes to harnessing this technology.
Take BILL as an example. The company processes payment volume equal to roughly 1% of U.S. GDP. Each week, that translates to billions of dollars moving across millions of transactions from their customers. And for the first time, more than half (54% to be exact) of those payments are sent and received between businesses that are both on BILL’s own network.
In other words, more than half of the money moving through BILL’s platform never leaves its ecosystem.
That matters because when both sides of a transaction, payer and payee, are inside the same network, BILL’s systems (and now its AI agents) have full visibility into the data, timing, and risk of those payments.
Who Goes First?
If the ultimate destination, when it comes to agentic AI, is the elimination of human-driven back-office systems and processes, it will start down market and work its way up.
“Entrepreneurs don’t know what they don’t know,” René said. “They’re not attached to old workflows and they don’t necessarily want faster workflows – they just want less work.”
And let’s not forget, small business owners in no way signed up to do any back office work. That remains the key point.
Large companies have (and will continue to) put AI at the front of their business: a Netflix or Amazon-esque recommendation engine, a customer service bot trained on millions of interactions, etc..
Small businesses succeed by doing the opposite. They win by keeping the human in front and pushing the technology to the back.
Think about the Etsy seller who includes a handwritten thank-you note with every order. There’s no way they can spend hours reconciling invoices and still make time to write those notes! They can choose to have AI write the note or have AI reconcile the invoices.
Every small business owner will choose the latter, and that’s how they win.
AI’s role here isn’t to replace that creativity but to liberate it; to pull small-business owners out of their workflows and back to their craft.
Small businesses run on attention and artistry, and the back office has always been a tax on both.
Now, LLMs and agentic AI are poised to offer a different kind of progress altogether.
Not a faster back office. A smaller one.
And ideally (and eventually), no back office at all.
