A pattern that shows up repeatedly in software is the closed agentic platform versus the open composable tool. The closed platform offers a complete experience: the agent does the whole job, the workflow is opinionated, the integrations are pre-built. The open tool offers a single capability that the user can connect to anything else they use. Both can serve the same job. They appeal to very different buyers.

The closed platform’s pitch is leverage. The user buys an agent that already knows how the work should flow, and they get the entire workflow for the price of one product. The downside is the user has to accept the workflow the platform was built around. Customization happens within the platform’s vocabulary. Anything the platform doesn’t natively do, the user either lives without or builds outside the platform with no integration help.

The open tool’s pitch is composability. The user keeps their own workflow, their own document storage, their own analytical environment, and plugs in a capability that does one thing. The downside is the user has to assemble the workflow themselves. The upside is that the workflow is theirs — they can rearrange it as their needs change, and the tool doesn’t have a vote in what they build around it.

For a certain kind of buyer, the closed platform wins. The customer who wants an end-to-end solution, who values not having to think about integration, who is comfortable adopting the platform’s opinion of how the work should be done. The cost of the platform is justified by the cognitive load the platform absorbs.

For another kind of buyer — usually the more technically capable or workflow-particular professional — the open tool wins. This buyer already has a workflow they like. They use their own preferred document tools, their own analytical environment, their own way of organizing the work. They don’t want a platform that asks them to move into a new home. They want a capability they can wire into the home they already live in.

The open standard is a specific variant of the open tool. When the capability is delivered through a standard protocol that other tools also speak, the composability multiplies. The user is not just plugging in one tool — they are participating in an ecosystem of tools that all interoperate. The closed platform cannot replicate this even if it wants to, because participating in the ecosystem would mean giving up the workflow control that defines the closed platform’s value proposition.

The strategic question for a small tool building against a well-funded closed platform is whether the open standard wedge is genuinely defensible. Usually it is, for a specific reason: the closed platform’s reluctance to open up is not arbitrary, it is structural. Opening the platform would erode the workflow opinion that the platform is monetizing. The platform’s customers chose it precisely because they wanted the opinion. Trying to serve both audiences — those who want the opinion and those who want composability — usually means serving neither well.

The open standard tool can therefore stay focused on the composability audience without much fear of the closed platform copying the position. The tool wins not because it has more resources, but because the position it occupies is one the closed platform cannot move into without becoming a different product.

+++