Software Stopped Competing With Software: Why the $1.5 Trillion Labor Budget Is the New B2B Battleground

For as long as B2B software has existed, every company in it fought the same war.

You sized your market by counting the other vendors who did roughly what you did. You won by being faster, cheaper, or better than them. Your total addressable market was whatever buyers were already spending on software like yours, and your job was to take a bigger slice of it.

That map is now wrong. The most dangerous competitor your software company will face next year doesn't sell software at all. It's a line item on your customer's payroll, and a second one on their invoice from a consulting firm.

For Founders, CEOs, Heads of Product, and GTM Leaders at B2B software companies deciding what their company is actually selling over the next three years.

Something fundamental flipped in the last eighteen months, and most software companies are still pricing, positioning, and selling as if it didn't. The category has a name now: services-as-software. The strategy underneath it is bigger than the name suggests.

The TAM just inverted

Start with the size of the pools, because that's where the whole argument lives.

HFS Research, which has tracked the services economy for years, breaks enterprise spending into four buckets: roughly $2 trillion on services, $1.5 trillion on labor, $300 billion on SaaS, and $380 billion on infrastructure. Sit with those numbers for a second. The software budget you've spent your whole career fighting over is the smallest pool on the board. The labor and services budgets next to it are more than ten times larger.

For thirty years that didn't matter, because software couldn't touch the labor budget. Software made people more productive; it didn't replace the work itself. A contract analysis tool helped a lawyer review faster. It didn't be the lawyer. The boundary between "tool" and "worker" held, and software stayed politely inside its $300 billion lane.

Agentic AI erased that boundary. When a system can complete the work end to end instead of merely assisting, the relevant question stops being "is this better than the competing software?" and becomes "is this cheaper and more reliable than the human or the consultant who used to do this?" The moment that question becomes answerable with a yes, your TAM is no longer the software budget. It's the labor budget. HFS now projects services-as-software will grow into a $1.5 trillion market by 2035, pulling revenue out of traditional IT services as that pool shrinks. Foundation Capital has gone further, framing the prize as a $4.6 trillion opportunity measured against the cost of human-delivered work that AI can absorb.

This is the inversion. Software's ceiling used to be other software. Its ceiling is now human work, and that ceiling is roughly an order of magnitude higher.

Why this is happening now, and not as a slide in a 2021 deck

"AI will replace services" has been a venture talking point for years. What changed is that the buyers and the incumbents started moving real money, which is the only signal that matters.

Boston Consulting Group pegged the agentic-AI opportunity in tech services alone at $200 billion, and framed it explicitly as a disruption of delivery economics: the billable hour is the thing under threat, not just the software license. Fortune, reporting in April 2026 on the pressure hitting firms like Cognizant, called it a $6 trillion reinvention and landed on a phrase that should make every software founder pay attention: IT services firms now have to start underwriting outcomes rather than selling effort. When a 300,000-person services giant starts talking about guaranteeing results instead of staffing projects, the ground has already moved.

The incumbents see it too. At its Knowledge 2026 event, ServiceNow unveiled what it described as an autonomous AI workforce meant to "sense, decide, and securely act" across a company's operations. Deloitte's 2026 technology predictions put the collision plainly: SaaS and AI agents are merging into something that reshapes budgets, headcount, and the customer relationship at once. And the buy side is hungry for it. The professional-services sector has gone from dabbling in AI to leading adoption across industries, with the majority of legal and accounting professionals now using AI tools that barely existed two years ago.

Put it together and the timing isn't mysterious. The capability crossed a threshold, the incumbents started repricing around outcomes, and the buyers stopped treating "let the machine do the work" as science fiction. That's a market, not a forecast.

What changes when you sell against labor instead of software

Here's the part founders underestimate. Moving from "tool" to "work" isn't a marketing reskin. It changes four things at the core of your business, and getting any of them wrong leaves most of the opportunity on the table.

Your pricing anchor changes. A seat license gets benchmarked against competitors' seat licenses, which drags every deal toward the cheapest comparable tool. Work gets benchmarked against the fully loaded cost of the person or firm doing it today. The same product that looks expensive next to a rival's $40-per-seat plan looks like a bargain next to a $95,000 analyst or a $400-an-hour consultant. If you price against software, you're competing in the $300 billion pool. If you price against the labor it replaces, you're competing in the $1.5 trillion one. Most companies leave the bigger anchor unspoken and quietly cap their own ceiling.

Your buyer changes. The software buyer optimizes for fit, integration, and price within a software budget. The person who funds headcount and approves the consulting spend is a different human, often more senior, with a much larger number to reallocate. Selling services-as-software means getting in front of that person, the one deciding whether to hire three more people or buy a system that does the work of five. That's an operating-budget conversation, not a software-procurement one.

Your proof bar changes. Nobody asks a productivity tool to guarantee an outcome. They ask a worker to. The minute you claim to do the job rather than help with it, the buyer expects you to stand behind the result, which is exactly why the smartest incumbents are moving to underwrite outcomes. Service-level agreements, accuracy guarantees, and pay-for-results structures stop being aggressive sales tactics and become table stakes. The trust bar is higher because the promise is bigger.

Your competitor changes. You're no longer in a feature war with three other vendors. You're competing against a hiring plan, a BPO contract, and an incumbent consultancy with a thirty-year relationship and a golf membership. Some of those competitors are weak on price and slow to deliver, which is your opening. Others come with trust and accountability you'll have to earn the hard way.

The trap: selling a "work" product into a "tool" frame

The most common failure here is quiet and expensive. A company builds something that genuinely replaces a chunk of human work, then positions it, prices it, and sells it like software, because that's the muscle memory.

The result is predictable. The product gets slotted into a software budget, benchmarked against cheap tools, and bought by someone three levels below the person whose headcount it could replace. The company captures a sliver of the value it creates and wonders why deals feel small and price-sensitive. It built a worker and sold it as a widget.

The opposite trap is just as real and more dangerous: claiming to replace the work, getting bought on that promise, and then failing to deliver the outcome. Underwriting results you can't actually guarantee is how you turn a services-as-software pitch into a churn machine and a reputation problem. The category rewards companies that pick a job they can genuinely own and punishes the ones that oversell autonomy they haven't built yet.

A framework for making the shift

If the thesis holds for your product, here's the sequence that moves you from selling tools to selling work.

1. Reprice against the labor, not the license

Build your value story around the fully loaded cost of the work you replace: salary plus benefits plus management overhead, or the consultant's blended rate, or the BPO contract's annual spend. That number is your real anchor. Price to capture a meaningful share of it while still being an obvious bargain against the human alternative. Stop letting a competitor's seat price set your ceiling.

2. Sell the outcome, and be willing to stand behind it

Lead with the job done, not the features that do it. Then back it with structure the buyer can trust: an SLA, an accuracy threshold, a pilot that proves the result before they commit, a pricing model that ties at least part of your fee to the outcome. You can't ask for labor-budget dollars on a tool-budget promise.

3. Re-target the budget owner who funds the work

Map who actually controls the headcount or services line your product displaces. That's your economic buyer now. The software champion still matters for evaluation, but the deal is won or lost with the person deciding between hiring and buying. Build the relationship and the business case for them specifically.

4. Redesign delivery, and accept that some of it is services

Replacing human work usually means keeping humans in the loop, at least early: an oversight layer, an implementation team, a forward-deployed engineer who tailors the system to the customer's reality. That carries a services cost and compresses the clean software margin you're used to. Plan for it deliberately rather than discovering it in your gross-margin line. The companies winning here treat delivery as a designed part of the product, not an afterthought.

5. Pick one expensive, repetitive, well-defined job and own it completely

Don't try to absorb the whole labor budget at once. Find a single high-volume, high-cost, rules-heavy task that a system can do reliably end to end, and become the undisputed best in the world at that one job. A narrow wedge you can actually guarantee beats a broad promise you can't. Expansion comes after trust, not before it.

The honest counterpoint

This is not a free lunch, and pretending otherwise is how founders get hurt.

Selling against the labor budget invites a higher bar on reliability, liability, and trust than software has ever had to clear. The human-in-the-loop delivery that makes outcomes safe also drags on margins, and public-market investors still pay more for clean software economics than for services revenue, which means the transition can cost you multiple even as it grows revenue. Not every job is automatable to the standard a buyer will pay for, and the ones that are tend to attract a crowd fast. The opportunity is enormous and genuinely contested. Treat anyone who tells you it's easy with suspicion.

What to do this quarter

You don't need a strategic offsite to start testing the thesis.

In the first 30 days, pick your single highest-value product capability and write down the fully loaded cost of the human work it replaces today. If that number is several times your current price, you have a repricing and repositioning problem worth solving, and a much larger market hiding behind your current TAM.

By day 60, take that number into three live deals. Reframe the pitch around the labor it displaces, get in front of the person who funds that labor, and watch how the conversation changes when the comparison stops being a competitor's seat price.

By day 90, make the call on standing behind an outcome. Pilot one outcome-based or guaranteed structure with a customer who trusts you, and see whether your product can actually carry the promise. That single experiment will tell you more about your real position in this market than any analyst report.

The pools haven't changed size because of a trend piece. The $1.5 trillion is sitting there whether or not your company reaches for it, and the incumbents whose revenue it represents are already repricing to defend it. The software companies that win the next decade won't be the ones with the best features in the $300 billion pool. They'll be the ones who realized they were never really selling software, and went after the work itself.