Scott Brinker published a piece on April 14 that a lot of people in martech are going to quote for the rest of the year. It's called From apps to infrastructure: martech's most important migration. Read it in full before you read this. We're going to respond to it directly.
The short version of his thesis: as AI makes custom apps and agents trivial to build, durable value in martech is shifting from the application layer — the vendor-opinionated UIs and workflows — to the infrastructure layer underneath. Data. Identity. Orchestration. Governance. Context. The marketing ops leaders he's talking to don't think of Braze or HubSpot or Marketo or Salesforce as "where I spend the day" anymore. They think of them as back-end services. They build their own apps and agents on top.
Then Brinker closes with a line worth picking up and carrying:
New horizons of opportunity for agencies and service providers who pair creative and strategic imagination with the technical craft to render that vision into hypertailored, competitively differentiated apps and agents for their clients. For those willing to develop that craft, this is springtime.
He's right. And springtime has a shape.
The application layer is not where the agency business lives anymore
For the last fifteen years, the agency model in martech has been: know the platform, configure the platform, sometimes build on the platform, charge for the hours it took. The platform vendor made the meal. The agency reheated it for the client and set the table.
That model is not dying because AI is making agents. It's dying because the marketing ops leaders Brinker is quoting have already figured out what the agents are for. Jason Lemkin's line — "the CRM decision is no longer a CRM decision, it's an AI infrastructure decision" — is the tell. The CRM didn't become more important because it got better. It became more important because the agent fleet needed somewhere to live, and Salesforce happened to be the substrate underneath.
Meaning: the valuable work is no longer "here is how our platform does X." The valuable work is "here is the hypertailored app or agent your team actually needed, built on a foundation that will still be here in three years."
That shift has a second-order consequence nobody is saying clearly enough. If the craft is building hypertailored apps and agents, then the people with the craft are not interchangeable agency labor. They are practitioners. Named. Accountable. Often better than the platform's default at the specific thing their client needs. And they need a different structure than a traditional agency gives them.
Tech Stack Reimagined: The Reality of Building an AI-Native CRM
The Decentralized Agency
The Value-First Team has been building toward this structure for about a year. We call it the Decentralized Agency.
Here's the division:
The practitioners are the chefs, in Brinker's metaphor. They have client relationships. They have craft. They build the hypertailored apps and agents. They deliver the project work. The project revenue flows to them, not through an agency holding company that taxes them for the privilege of having a logo to put on an invoice.
The foundation is what practitioners build on. Shared infrastructure. A Customer Value Platform pattern on top of HubSpot (the substrate). An agent substrate. A methodology that tells you where a customer actually is, instead of where your pipeline report wants them to be. A shared context layer so the practitioner isn't re-inventing the semantic model for every client.
Practitioners buy the butter. They don't churn it. Value-First provides the butter.
That line isn't cute. It's the operating principle. If you're a practitioner in this model, the stuff that used to eat 40% of your week — building and maintaining the platform architecture, the context layer, the agent scaffolding, the governance, the methodology — is already there. You bring the craft. You bring the client. You bring the judgment about what agent or app this specific business actually needs. You keep the project revenue because you are the one who earned it.
That is the shape of the springtime Brinker is pointing at. It is not "agencies get bigger because AI lets them build more stuff." It is "the agency unbundles into practitioners on top of shared infrastructure, because the infrastructure is where the durable value lives and the craft is where the project value lives, and those are two different businesses that have been forced into one wrapper for two decades."
Why the Value-First methodology fits this shift
Three anchors from how we work map directly onto what Brinker is describing.
Configuration over Customization. This has been a core belief in our methodology since before the AI wave. The idea is that the platform is the platform. You don't bend it into the shape of a custom implementation that nobody can maintain six months later. You configure it cleanly, then you build the specific things your business needs on top — as apps, as agents, as interfaces. That is exactly the build-and-buy-at-different-layers move Brinker calls out. Buy the infrastructure. Build the apps and agents. Stop buying somebody else's opinion of the meal.
The AI-Native Shift. Our four-week program exists because the bottleneck on this whole migration is not technology. It is that most teams still think of AI as a feature inside their existing apps, rather than as the substrate the next generation of apps will be built on. The shift is a mental model change before it's a tooling change. Practitioners who have made that shift can look at a client's stack and see the infrastructure underneath. Practitioners who haven't made it see ten SaaS tools and a budget problem.
The Customer Value Platform pattern. This is the architectural answer to "where does the agent fleet live?" For us it's a HubSpot-native pattern — three custom objects (Deliverable, Interest, Investment), a clean data model, the methodology baked in. It is not a product we sell. It is the shape of the foundation we build underneath practitioners so their apps and agents have somewhere coherent to stand.
None of this is speculative. It is the ground the Value-First Team has been working on since well before Brinker wrote his piece. The piece is useful because it names the migration out loud, in a publication the martech industry actually reads.
What we're not claiming
A few things worth being honest about, because the operating pattern around here is "state the architecture, don't oversell the state."
The Decentralized Agency is not a finished structure. The foundation pieces — methodology, platform pattern, agent substrate, shared context, commerce rails — are at different levels of maturity. Some are production. Some are weeks old. Some are still being named. We are building in public, at the same time Brinker is writing his piece, because the shape is real even when the scaffolding isn't all welded yet.
We are not claiming the Decentralized Agency replaces every agency. It replaces the model where an agency holding company mediates between practitioners and clients by extracting margin from project work. If that describes your firm, the next three years will be uncomfortable. If you are a practitioner who has been quietly doing the craft inside one of those firms, the structural answer Brinker is pointing at is the thing worth paying attention to.
We are not claiming this is the only answer. It is the answer we have conviction in, and the one we are staking the foundation on.
Back to springtime
Brinker's closing line is the right one. The craft is real. The opportunity is real. What his piece doesn't say — because it wasn't his piece to write — is that springtime for practitioners doesn't look like a thousand new agencies. It looks like a small number of shared foundations and a much larger number of named practitioners building the hypertailored work on top.
Buy the butter. Build the meal. Keep what you earned for cooking it.
That is the Decentralized Agency. It is the structural answer to the migration Brinker just mapped. And if you're one of the practitioners he's writing about — the ones with the creative and strategic imagination paired with the technical craft — this is a good moment to stop reading about the shift and start looking at the foundations being built underneath it.
We're building one of them. Ready to make the Shift?
— V
Written in response to Scott Brinker, "From apps to infrastructure: martech's most important migration," chiefmartec, April 14, 2026. Read the original on chiefmartec.com.