AI's Biggest Market Isn't Software. It's the $Trillions Spent on Human Labor.
Hey—It’s Tom
Every week I sit down and try to write my best newsletter yet.
Maybe today I did it.
Estimated read time: 2 minute 17 seconds.
Everyone in tech is looking at the wrong number.
They’re watching SaaS revenues, cloud margins, API pricing wars. They’re debating whether OpenAI can justify a $730 billion valuation based on what developers will pay per token. Smart people, running serious models, all pointed at the same thing: software eating software.
Sponsored by Constant Contact
Marketing that’s always on—even when you’re not. Get automation tools that send your messages for you.
You can’t be everywhere at once—but your marketing can.
With Constant Contact’s automation tools, your emails, texts, and offers go out at the right time, every time—without you having to lift a finger.
Want to show your audience you care about them? Use an automaton template to send birthday messages. Working on driving sales? Set up an abandoned cart email automation to gently urge almost-shoppers to become actual shoppers. Looking to bring lapsed customers back into the fold? Send automations with offers and promos to entice them to buy again. You can even create custom automation paths that work best for your business.
It couldn’t be easier. Just set up automatic triggered messages based on your customers’ behavior, and watch those messages run while you focus on everything else.
SECOND ORDER readers, if you’re ready to save time and put pesky administrative tasks on autopilot, give Constant Contact’s marketing automation tools a try. Get started today—for free.
But the real play here isn’t software at all. It’s the roughly $61 trillion that the global economy spends every year on human labor. That’s the number that matters. That’s the ocean these companies are sailing toward and most people are still arguing about the dock.
Think about what actually costs money inside a company. Not the Salesforce license. Not the AWS bill. It’s people. It has always been people. Payroll is typically the single biggest line item in most companies’ operating budgets. An accountant reviewing contracts. A paralegal doing document review. A junior analyst building decks nobody asked for. A customer service rep explaining a return policy for the four hundredth time that week. A radiologist reading scans. These are not edge cases. This is the whole economy.
Software has been chipping away at this for forty years and it barely made a dent. Spreadsheets didn’t replace finance teams. CRMs didn’t replace sales teams. ERPs created entire new categories of consultants to implement and maintain them. Software made workers more productive but it didn’t replace the work itself because software, until very recently, could only follow rules. And most valuable human work isn’t rule-following. It’s judgment.
What changed is that AI actually has judgment now. Or something close enough to judgment that the distinction stops mattering in most business contexts. It can read a messy contract and find the problematic clause. It can look at an x-ray and grade the severity. It can write a passable first draft of almost anything and a genuinely good fifth draft if you push it. It can answer customer questions with context and nuance. Not perfectly. But good enough. And good enough at a tenth of the cost is a completely different market than perfect at the same cost.
This is the shift that is being underpriced by most of the conversation happening right now.
The software market, giant as it is, was always a margin game on top of labor. You buy the tool so your people can do more. The new game is different. The new game is what happens when the tool starts doing what the people do. You are not selling productivity software anymore. You are competing with the salary line.
Marc Andreessen famously said software is eating the world. That was true and it was a big deal. But labor is the world. Global GDP sits at roughly $117 trillion as of 2025. The ILO puts labor’s share at about 52 percent of that, which gets you to somewhere around $61 trillion flowing to workers every year. Software was eating the world but it was eating the edges, the distribution, the interface layers. AI is going after the core.
And the core is enormous in ways that the software comparison doesn’t capture. The global software market is around $750 to $800 billion. The global market for knowledge work alone, white collar wages in industries that look automatable, is somewhere north of $15 to $20 trillion just in the developed world. That’s a 20x difference. The addressable market for AI if it can genuinely replace, or more likely redistribute, even a fraction of that labor is not a software story. It’s something that doesn’t have a clean historical analogy.
There’s a version of this that sounds dystopian and a version that sounds like the biggest economic unlock in human history. Both are probably partially right, which is the uncomfortable thing nobody wants to sit with. The economy will create new jobs, it usually does, but the transition is real and the people most exposed are not the people with the most political power, which makes the policy response slow and inadequate. But that’s a different essay.
Leadership Can’t Be Automated1
AI can help you move faster, but real leadership still requires human judgment.
The free resource 5 Traits AI Can’t Replace explains the traits leaders must protect in an AI-driven world and why BELAY Executive Assistants are built to support them.
The point for now is simpler. If you are trying to understand where the value in AI is going to land, stop thinking about software multiples and start thinking about what things cost when a smart capable person does them versus what they’ll cost when AI does them just as well. The spread between those two numbers, multiplied across the entire global economy, is the actual market. And it is not a software company number. It is something larger than most people have a comfortable mental model for.
The software industry spent fifty years building picks and shovels for human workers. The next twenty years are about what happens when the picks and shovels learn to mine.
If you like what you read please do “add to address book” or reply (it helps with deliverability). If you don’t you can unsubscribe below.
See you tomorrow (Monday)—Tom
Ad