Every major lab, every well-funded startup, and every enterprise platform vendor is converging on the same opportunity at the same time. OpenAI announced what is essentially an AI services company. Anthropic's Claude Code is reshaping how developers interact with codebases. Cursor, Devin, Factory, and dozens of others are attacking the coding agent layer from different angles. Underneath all of it, every enterprise leader with a budget is asking the same question: how do we actually operationalize this? The answer is not as clean as the pitch decks suggest. The agent space is the NBA of startups — fast, physical, and unforgiving. And almost everyone selling into it is making the same set of strategic mistakes.
I broke this argument into six atomic posts. Read them in any order:
- The Arena of Arenas — There is no winner-take-all platform in agents. You fight labs, incumbents, peers, and internal builds at the same time.
- The Harness Layer Has No Moat — The execution loop is commoditized. The opportunity lives in the abstractions one layer up.
- Catch the Buyers Before They Become Builders — Mature dev orgs build, less mature orgs buy. The platform play is the maturity gradient.
- Background Agents Are the Underbuilt Layer — Coding copilots are crowded. Autonomous, scheduled, reviewable agents are what enterprises actually need.
- Agents Are About to Break Out of Engineering — Non-engineering agents are easier to build but harder to contextualize. The org knowledge layer is the real unsolved problem.
- The Design Partner Discipline — Pluck a feather from each passing goose, but never follow one off a cliff. Provocatypes beat polished mockups.
The labs, the incumbents, and the startups are all converging on the same opportunity at the same time. There is no moat to be found by sitting still — only by moving fast, building well, and staying close to the customer. That is the platform game in this market, and it is the only game worth playing.
— Ry
Sources
Related Essays
The Arena of Arenas: Why There Is No Winner-Take-All in Agents
Enterprise AI deployment is services-shaped, not software-shaped. There is no single board where you win or lose — you are fighting labs, incumbents, peers, and internal builds at the same time.
The Design Partner Discipline
A formal design partner program is how you avoid building in a vacuum without getting pulled in five directions. Pluck a feather from each passing goose, but never follow one off a cliff.
Catch the Buyers Before They Become Builders
Mature dev orgs build their own agent infrastructure. Less mature orgs buy. The platform opportunity is a maturity gradient, and the play is the PostHog playbook applied to agents.
Key takeaways
- There is no winner-take-all platform in agents — the market is an arena of arenas, and you fight on every front at once.
- The agent harness layer is commoditized. The real opportunity is context, memory, orchestration, and the organizational layer where everyone is still building custom.
- Enterprise GTM for agents is services-shaped. Forward-deployed engineering, Palantir-style, is the only model that closes the gap between demo and production.
- Coding agents are a wedge, not a destination. Pivot to a platform that orchestrates a mesh of agents across the organization before commoditization eats the category.
- Mature dev orgs build their own agent infrastructure. Less mature orgs buy. The platform opportunity is catching the buyers before they mature into builders.
- Non-engineering agents are easier to build but harder to contextualize. The organizational knowledge layer — not the execution layer — is the real unsolved problem.
- Design partner programs are how you avoid building in a vacuum. Structured customer collaboration beats roadmap guessing every time.
- Move fast, build well, stay close to the customer. There is no other moat when labs, incumbents, and startups are all converging on the same opportunity.
FAQ
Why is there no winner-take-all platform in agents?
Enterprise AI deployment is services-shaped, not software-shaped. Every organization has different workflows, compliance requirements, and integration surfaces, so the gap between demo and production is closed by people who understand the business — not by a single dominant platform.
What does it mean that the agent harness is commoditized?
The execution loop itself — Claude Code, Open Code, and similar open-source forks — is no longer a differentiator. Companies have churned through homegrown loops, Goose, and Aider before settling on whatever the current best option is. The real value lives in context, memory, orchestration, and the organizational layer above the harness.
Why are coding agents a wedge rather than a destination?
Coding is the most crowded and most quickly commoditizing surface in the agent market. The durable opportunity is orchestrating a mesh of agents across the entire organization, catching less mature buyers before they mature into builders and bring infrastructure in-house.
What is the organizational context problem for non-engineering agents?
Non-engineering agents do not need environment setup, CI/CD, or test suites — but they need a living representation of how the company actually works. Standard operating procedures, product knowledge, customer segments, and even disagreements between stakeholders all need to be surfaced and maintained. This organizational knowledge layer is the hardest unsolved problem in the space.
Why do design partner programs matter for agent platform companies?
Customers in this space are experimenting rapidly and will ditch tools overnight. Without structured design partnerships, platform companies either build in a vacuum or get pulled in a different direction by every customer request. A formal program — with regular syncs, provocatype prototyping, and shared accountability — lets you pluck a feather from each passing goose without following any single goose off a cliff.