AI Needs an Application Layer

Investors, in their infinite wisdom, have declared the application layer dead. Control Point incumbents and AI-native challengers alike are chum in the water for the great white shark coming up the stack: foundation models. The investor consensus has spoken! Momentum-money and mercenary talent are gone. Applications are dead!

We think differently.

Yes, AI is magic. It is pushing the technological frontier faster than anything I have seen. But intelligence alone doesn’t book the revenue, chase the invoice, or file the compliance report. AI needs an application layer to bring it to life for customers, and building your product as a system of action is how you can achieve this.

Outside of software development, that application layer largely doesn’t exist today. And in most industries, the foundation models aren’t likely to build it (we’ll explain why later in this article).

So in most industries, the application layer — the System of Action — is there for the taking. For both the category leaders and AI-native challengers alike.

Capturing this prize takes product focus and engineering excellence. Application companies now need to think like infrastructure companies to provide the scaffolding agents need to run reliably and get better over time.

Not everyone will make it.

Doing the Work ≠ Doing the Job 

Generative AI introduces non-determinism, and recursive non-determinism approximates reasoning. Run it long enough, and it starts to approach thinking. That alone would be interesting. But this is what makes it consequential: given data, context, and the permission to trigger tools, AI can do work in the real world. 

That is the frontier shift, but most people are missing the bigger opportunity: doing the work isn’t the same as doing a job. 

A job is a bundle of tasks: some need reasoning, some need to run the same way every time, and some still need a person. On its own, the model handles those tasks one at a time. Without a System of Action, the work gets done, but the job doesn’t. And all the learnings gained in the process are lost. 

That’s why so many businesses have lackluster experiences with the first wave of AI. Recently, MIT’s Project NANDA put a number on this: 95% of organizations are getting zero return on $30–40 billion of enterprise GenAI investment. 

It wasn’t a model quality problem. It was a context and learning problem. Tools that don’t retain context and don’t improve from feedback. The limiter seems to be that companies still think of AI as a tool, not as a coordinated system of action. 

You wouldn’t ask an employee to do their job without any onboarding or education about company context. And yet, that’s what we ask of AI. Roughly one in four enterprises still hadn’t connected their AI to internal data, per a December 2025 OpenAI report. 

This is staggering because case studies show that when you give AI context, learning loops, and guardrails (the aforementioned system of action), outcomes improve dramatically. Without hand-built skills, Claude’s accuracy on analytics questions is about 21%. With skills, accuracy hits 95% — and then drifted to 65% within a month. How did they stop this? They started treating the skills as ongoing learning cycles, rather than one-off launches. 

Anthropic’s analysis is the best argument for the application layer I’ve read to date: “an agent that doesn’t understand your business will answer what the user asked, but not what they meant.”

A system of action is how you train AI to do a job end to end. It thinks, then reaches for the tools (your workflows, your data, or your permissions) to run the deterministic parts, and then routes tasks requiring human judgment straight to decision-makers. Agents decide, act, produce the output, and observe the real-world consequences in a loop that rapidly improves them. 

So why isn’t every company doing this today? It’s simple: they can’t. (And the frontier labs know it). 

Your Opportunity: The System of Action Is Nobody’s Job 

To build a system of action, someone inside the business has to own every step. Someone has to own the definition of revenue recognition and defend it. Someone has to articulate how your company defines every term with appropriate nuance. 

And you can’t just automate it away. Anthropic tried bootstrapping the semantic layer and got plausible-looking definitions that encoded the exact ambiguities they were trying to remove.

To be clear: we are not saying AI and agents don’t work. In fact, we’ve seen firsthand how it’s transformed our business at Tidemark. But we’ve done the hard yards ourselves — enabling our tools and data, codifying our semantics, taxonomies, and processes, and building a harness for specific outputs and predictable execution.

It has transformed how we do work, but it was also a pain in the ass.

Now ask yourself what auto repair or HVAC shop is going to take that on. Or which 12-person accounting firm has a partner who is going to build and maintain an ontology of their firm and its clients. They don’t have one, and they never will. It’s a pain in the ass, and it is nobody’s job.

Turns out many of the largest enterprises aren’t excited to take this on either. 

Consider the foundation models’ investment in services organizations. Actions speak louder than any study. Collectively, the foundation models have committed roughly $10 billion to buying and building services companies so they can get this done for their customers. 

OpenAI alone has raised $4 billion for the OpenAI Deployment Company and its acquisition of Tomoro, staffing it with forward-deployed engineers from day one. 

Anthropic stood up Ode with Anthropic alongside Blackstone, Hellman & Friedman, and Goldman Sachs, aimed squarely at private-equity-owned businesses. 

When the model makers hire armies of humans to make the models work inside real businesses, they are telling you what the job requires: it requires an application layer!

This is your opening. As a Control Point incumbent or Native AI challenger, you can do it for them. 

You can build your customers the application layer required for AI to create value in the enterprise. And you have more right to do it than anyone, because you often had to solve those exact problems to build your product in the first place.

The Labs Won’t Be the Application Layer for All Industries

The fear, naturally, is that the labs will build this application layer for all industries.
It’s somewhat warranted, because they have built much of the application layer for software development. But this was inevitable: Code is their home turf: public training data, verifiable output. The labs are software developers building for software developers.

But most of the economy doesn’t look like software development. This means that in most industries, the application layer is still very much undecided, and the labs are not the experts. They will go after some high-value segments, but not all.
So where else will the labs play? Ask four questions:

• Is the prize big enough to make their roadmap?
• Does the technology have what it needs to solve the problem? Pre-training data to learn from, outcomes you can verify?
• Will the foundation models build the best product for the whole job, including the deterministic, unglamorous majority of the job, and carry the liability when an agent acts?
• And will your customers change how they work?

Software development answered yes to all four, and the labs built the best product for developers. 

Now ask the same four questions about dental practices, accounting firms, freight brokerages, and title companies. Even if all four answers are yes, ask whether the foundation models will have the interest and tenacity to build the best product for the specific industry.

Frontier labs, even with infinite capital, can’t build for every vertical at once. They will concentrate on the biggest prizes. Attacking software development made perfect sense. How far down the priority list is HVAC operating software?

Frontier labs will also concentrate where AI works well — where pre-training data is available, where outcomes are verifiable, where a wrong answer is cheap. But what about the rest of the job that isn’t as glamorous? When the right tool is deterministic software, classic ML, or just plain rules? That’s not what the labs are selling.

Customers don’t pay for AI for the sake of AI. They’re paying to have the whole job done. If you can’t do the entire task mix, you can’t “do the job.” Why would the labs build for “the rest of the job” when their pitch is that the model does it all?

Finishing the job also invariably means dealing with edge cases. Take Watsco, North America’s largest HVAC distributor. When the EPA’s refrigerant rules forced the A2L transition, Watsco had to add over 10,000 new “regulatory match-ups” — each one with dimensions, capacities, bills of material, warranty information — to cover over 50% of its products. Regulatory match-ups are no joke — use the wrong SKU, and you’re arguing with a manufacturer about whether the warranty still covers a big honking piece of capital equipment.

Building for each vertical means learning an industry from scratch — its language, its ontologies, its regulations, its commercial practices, its workflows, and its edge cases. Are researchers at the frontier labs fired up to code changing refrigerant rules? I doubt it. 

Finally, the frontier labs can only move as fast as customers want to adopt. Most industries aren’t as change-seeking as the Valley. Disruption and behavior change are a hassle, not a status to be celebrated. The dentist who has run the same front-office routine for 11 years is not looking for a new way to work; she is looking for Thursday to go smoothly. Change costs her chair time, retraining, and a week of rework.

You have a moment in time to build the agentic application layer for your customers, ahead of the industry and ahead of the labs. Do not read that as safety. Read it as a window.

Somebody is going to become the application layer — the System of Action — in your market. If it isn’t you, you’ve got problems.

Customers Want AI Where They Already Work

The good news is that customers are telling us where they want the application layer built. They want AI embedded in their workflows, working on their data, in UIs they already use.

In our own research, when we surveyed vertical software buyers in January 2025, 83% said they would switch to a Control Point agent that was 80% as effective. We got the same result when we ran the survey again in April 2026. Customers want AI embedded in the system they already run their business on. They do not want another login or another vendor.

The 10x System of Action Opportunity

If you make the transition and deliver this AI magic to your customers, what’s that worth?

Turns out it’s worth far more than the historical software opportunity alone.

Tidemark Fellow Ershad Jamil, former Chief Growth Officer at ServiceTitan, and I sized it here. Looking at the P&L of a representative small business, labor is roughly half of all spending. Software is just 5%. Roughly, this means the System of Action opportunity is 10x larger than the software opportunity!

This isn’t just theory — we’re starting to see it in the marketplace.

“Ironically, before long-duration agents, vertical AIs were stuck in co-pilot mode. Now you have vertical AIs that can actually take action and do work. You have a vertical AI booking a $7M deal with a customer who never paid more than $500K for all software tools combined.”   

– Anthropic insider.

Sit with that a second — one vertical AI company signed a deal 14x bigger than everything that customer had ever spent on software, combined.

That’s not just 14x+ revenue. It’s 14x+ TAM, 14x+ the dollars to fund CAC, 14x+ the dollars to fund the next agent. You have 14x+ the relationship, 14x+ the sway, 14x+ the clout.