2026-11-25
Comective was an experimental web application that organized AI work into reusable frameworks instead of relying on individual prompts. It combined structured workflows with parameterized prompt templates, allowing users to approach recurring tasks through a consistent process rather than starting from a blank chat. The project explored how structured thinking could improve the quality, consistency, and efficiency of AI-assisted work.
The platform was designed for professionals who use ChatGPT regularly in their daily work, including consultants, analysts, managers, educators, and technical teams. It was particularly aimed at organizations that wanted multiple people to follow the same reasoning process while still adapting prompts to their own situations. Rather than serving casual AI users, Comective focused on repeatable knowledge work where consistency mattered more than creativity.
Users selected a framework from a structured catalog and entered a shared situation or challenge that was automatically incorporated into every prompt. Prompt collections were stored as external data, allowing new frameworks to be created, updated, and distributed without changing the application itself. The interface generated complete prompts from reusable templates and opened them directly in ChatGPT, allowing users to focus on the workflow rather than prompt construction.
The project treated prompts as implementation details rather than the primary product. Its core idea was that expert knowledge should be captured as structured decision processes that could be reused, improved, and shared across teams. This separation between workflows, prompt templates, and application logic made the platform flexible while keeping the user experience intentionally simple.
Comective was developed during a period when ChatGPT provided only limited support for reusable workflows, persistent instructions, and structured projects. As AI platforms rapidly introduced features such as custom GPTs, Projects, improved memory, and richer workspace capabilities, much of the value provided by a standalone prompt management application became available directly within the platform. Therefore, continuing to develop an independent service became strategically unnecessary, even though the underlying concepts remained relevant.
The project demonstrated that users gained more value from structured thinking than from increasingly sophisticated prompt engineering alone. It also confirmed that separating reusable knowledge from application code made frameworks easier to maintain and evolve over time. Although the business model did not remain viable, many of the design principles anticipated features that later became standard across modern AI platforms.