The next trillion-dollar solution won't sell software. It will sell work.

This idea has been getting a lot of attention, but every article I’ve read is missing the most important part.

The shift from software to services isn't just a business model play. It's a fundamental restructuring of who holds power, who captures value, and who gets left behind. And most people have absolutely no idea it's happening.

I run an AI implementation firm called Black Box Theory AI. We don't sell tools. We don't sell licenses. We help companies plug AI into their operations so the work just... gets done. What I'm seeing on the ground right now is both further along and earlier than people think, depending on where you look.

But the most important part of all is how it’s actually happening. Let me break that down.

The Software Model Is Cracking

For 25 years, software ate the world. Buy a license, get a tool, hire humans to operate it. QuickBooks costs you $10K a year. The accountant running it costs $120K. (direct example from Sequoia Capital)

The software was never the expensive part. The human was. The problem was the solution couldn’t be solved without either, making both necessary. This also made both extremely hard to replace or scale.

AI flips this completely.

The cost of intelligence work is collapsing toward zero. When Claude or GPT can do 80% of what a junior analyst, paralegal, or bookkeeper does, the tool and the operator merge into one thing. You don't need QuickBooks AND an accountant. You need something that just closes the books.

That's not software. That's a service.

And the companies that figure this out first are going to build moats that look nothing like traditional SaaS.

The Black Box Theory Shift

Here's what I keep coming back to in my own work.

We're moving into a world where the buyer doesn't see how the work gets done. They don't need to. They don't want to. They just want it done.

Think about it. When you hire an accounting firm, you don't watch them work. You don't care which version of Excel they use. You hand them your financials and they hand you back closed books. That's a black box. Input goes in, output comes out, and you trust the process because the result speaks for itself.

AI services work the same way. The client doesn't need to know which model is running, what the prompt chain looks like, or how many agents are coordinating behind the scenes. They need the work product. Clean. On time. At a fraction of the old cost.

This is what we're building at Black Box Theory AI. And what we have been building for the last 5 years through a number of verticals. Not another dashboard. Not another copilot with a chat window. An actual implementation layer where AI handles the work and the client gets the outcome.

The "black box theory" isn't a bug. It's the entire point. It’s what will unlock smarter, more efficient scalability globally.

The Real Motives Nobody's Talking About

Many of the original articles frame this as a clean business evolution. Copilots become autopilots, autopilots capture the work budget, everyone wins. That's a nice story. But there are deeper forces at play that deserve honest conversation.

The margin capture is staggering. When you replace a $150/hour human with an AI that costs fractions of a penny per query, and you charge services-level pricing, the margins are unlike anything SaaS ever produced. We're talking 85-95% gross margins on work that used to be 30-40% margin at a services firm. The companies that get here first aren't just building businesses. They're building money machines.

That said, there is a lot more at stake for consideration as an AI actively doing this. And this is important for any decision maker considering this to understand. There are laws, regulations, servers, privatized data pools, regulations, managerial systems, updates, retrainings, and so much more that need to be handled at scale. Beyond the fact that the AI alone will take massive training, standalone cybersecurity, and custom system handling, each of these additional elements will be the lifeline of both the AI firms and the partners they work with. A majority of our capital as a firm thus far has gone into this since the inception of our company.

Power is consolidating fast. Every profession used to be distributed across thousands of small firms. Thousands of accounting practices, law firms, insurance brokers, recruiting agencies. AI services collapse that fragmentation. A single AI-native company can serve a market that used to require 10,000 small shops. That's a massive concentration of economic power into very few hands if not managed correctly.

The workforce displacement is real. I'm not going to dress it up. When the work gets automated, the people who used to do that work need somewhere to go. The "services not software" framing sounds friendly, but the end result for millions of professionals is the same. Their current role gets absorbed. That's not a reason to stop building, but it is a reason to be honest about what's happening and consider what the potential outcomes could be.

Here's the part that makes me optimistic about all of this.

The Other Side: Radical Empowerment

The same force that consolidates power in some places distributes it in others.

A solopreneur with the right AI stack can now deliver what used to take a 50-person agency. I've seen it. I'm helping build it. One person, properly equipped, can run operations that would have required an entire department two years ago.

Small businesses that could never afford a real CFO, a legal team, or a marketing department can now access those capabilities through AI services at a fraction of the old cost. The quality floor has risen. The cost ceiling has dropped. The gap between what a Fortune 500 company can do and what a five-person startup can do is shrinking fast.

Geographic barriers are dissolving. Someone in a small town across the world can now compete with Manhattan firms for the same outcomes. It's not about where you are. It's about what you can deliver.

That's not just a market shift. That's a genuine democratization of capability. AI doesn't just replace workers. It creates entirely new categories of work, new types of companies, and new opportunities for people who were never in the game before.

The question isn't whether disruption is coming. It's whether we build systems that spread the upside or concentrate it.

Where We Actually Are (Not Where Social Media Thinks We Are)

Here's my honest read on the market right now.

99% of people have no idea any of this is possible. I'm not exaggerating. The average business owner hasn't even had a real conversation with an AI, let alone considered replacing their outsourced accounting or legal work with one. The gap between what's technically possible and what the market has adopted is the widest I've ever seen it in any technology cycle.

In some verticals, the autopilot model is already working. Legal document generation, medical coding, basic IT support. The intelligence work is getting done by AI right now, today, at scale.

In other verticals, we're still firmly in copilot territory. Management consulting, complex sales, anything requiring deep relationship context. The judgement-heavy work isn't going anywhere soon, just evolving.

And in most verticals, we're still at zero. The median company hasn't started. They're not even evaluating. They don't know what to evaluate.

That's the gap. And it's exactly where our firm lives.

At Black Box Theory AI, we're on the cusp of being ahead of this curve. Not because we're smarter than everyone else, but because we're actually in the trenches implementing AI for real companies with real workflows and real revenue on the line. Most of the AI conversation online is still theoretical. We're doing it and have been for years.

The Intelligence vs. Judgement Framework

The original article nails one thing: the distinction between intelligence work and judgement work is the single most important framework for understanding where AI will hit first.

Intelligence is following rules. Complex rules, sure, but rules. Translating clinical notes into billing codes. Drafting standard contracts. Reconciling accounts. Screening resumes against requirements. The rules might fill a 500-page manual, but they're still rules.

Judgement is knowing which rules to break. When to take on technical debt. Whether to settle a claim or fight it. Which candidate will thrive in your specific culture. That's pattern recognition built on years of getting things right and wrong.

AI has crossed the intelligence threshold. It can do most intelligence work autonomously right now. The judgement threshold is further out, but it's moving faster than anyone expected.

The smart play is to start with intelligence-heavy work, prove the model, and let the AI compound data about what good judgement looks like in that domain. Every completed engagement makes the system smarter. That's a compounding advantage that no software license ever provided.

What This Means If You're Building

If you're a founder, the playbook is becoming clear. Every improvement in the underlying models makes your ecosystem faster and more effective while your competitors scramble to justify their per-seat pricing.

Start where outsourcing already exists. If you already pay an external firm to do the work, you’ve already accepted that it can be done externally. You don’t need to restructure. You just need to swap vendors.

This will become your black box. You don't need to understand exactly what's happening inside. You need the output to be right. Focus on reliability, accuracy, and speed.

And think about both sides of the equation. The margin opportunity is real, but so is the responsibility. The companies that build this well, that spread the upside, that create new opportunities instead of just eliminating old ones, those are the ones that will last.

The Bottom Line

The software era gave us tools. The AI era gives us outcomes.

We're watching the largest shift in how work gets done since the internet. Most of the world hasn't noticed yet. By the time they do, the companies that moved first will already be entrenched.

The question isn't whether AI will replace services. It will. The question is who builds the systems that do it right, and whether the result is a world where more people can do more things, or one where a handful of companies own the work layer for entire industries.

I know which one I'm building toward.

Follow Scott Caputo

All my AI projects are managed at blackboxtheory.ai

I share my personal thoughts/opinions at scottcaputo.com