The ability for users to bring their own ChatGPT token allowance makes previously unprofitable, hyper-niche AI applications financially viable. Entrepreneurs can now build specialized tools for small markets (e.g., architectural file cleanup) that are too niche for major platforms, monetizing via upsold workflows instead of token usage.
As AI agents handle digital tasks, a major opportunity exists in bridging the gap to physical, real-world services. Businesses can create platforms that act as APIs, connecting requests from agents like OpenAI's Dots to vetted networks of human specialists (e.g., permit expediters, customs brokers) for last-mile task completion.
As thousands of developers build plugins for agent platforms like OpenAI's Dots, getting discovered will become a major challenge. A new industry, analogous to SEO for web search, will emerge to provide analytics and optimization tools that help companies get their agents selected by users, creating a large B2B SaaS opportunity.
Developers can now offer core AI applications for free by leveraging a user's existing ChatGPT token allowance. This lowers customer acquisition barriers, allowing businesses to focus on upselling proprietary workflows or enterprise features rather than charging for basic usage.
In an agentic world where AI decision-making is commoditized, a strong business framework is Trigger, Decision, Action, Feedback. The most defensible companies will own the workflow components surrounding the AI: the initial trigger (e.g., an invoice is due), the resulting action, and the feedback loop (e.g., payment confirmed).