Which has more entrepreneurial potential: ChatGPT Plugins, GPTs, or Agents?

As the wave of artificial intelligence technology sweeps the globe, OpenAI's ChatGPT and its ecosystem have sparked an unprecedented entrepreneurial boom. According to CB Insights, the AI sector attracted over $42 billion in venture capital in 2023, with generative AI startups accounting for a significant share. In this thriving market, entrepreneurs face a key question: Should they invest their energy and resources into ChatGPT plugins, custom GPTs, or more complex AI Agents? This article will delve into the business potential, market status, and future prospects of these three technology paths to help entrepreneurs make informed decisions.

Definition and Differences of Three Technology Paths

Before discussing entrepreneurial potential, let's first clarify the definitions and core differences between the three:

ChatGPT Plugins

ChatGPT plugins are extensions developed for the ChatGPT platform, enabling it to access third-party services, real-time data, and professional tools. Plugins connect via APIs, giving ChatGPT the ability to perform specific tasks, such as querying real-time data, controlling external applications, etc. OpenAI first launched the plugin feature in March 2023, and there are currently hundreds of officially certified plugins available in the store.

Custom GPTs

GPTs (i.e., customized GPT models) are a feature launched by OpenAI in November 2023, allowing users to create ChatGPT variants specifically for certain domains or tasks without programming knowledge. These custom models can have specific "personalities," knowledge bases, and behavioral patterns, and can be shared or sold to other users through the GPT Store.

AI Agents

AI Agents (intelligent agents) are more complex autonomous AI systems that can understand user goals, plan execution paths, and proactively complete a series of tasks. Unlike plugins and GPTs, Agents typically have stronger autonomy, planning capabilities, and memory functions, and can make decisions and execute complex processes on their own. Typical examples include Anthropic's Claude Opus, AutoGPT, and Adept's ACT-1, among others.

Market Status Analysis

ChatGPT Plugin Market

Since its launch in early 2023, the ChatGPT plugin ecosystem has expanded rapidly. According to official OpenAI data, as of early 2024, its plugin store has over 1,000 active plugins with more than 15 million monthly active users. However, from an entrepreneurial perspective, the plugin market exhibits several distinct characteristics:

1. Highly Concentrated Usage Patterns: Data shows that the top 20 plugins account for over 70% of usage, with tool-based plugins such as WebPilot, Wolfram Alpha, and Zapier dominating.
2. Unclear Profit Models: With the exception of a few paid plugins, most plugins have not yet formed a clear profit model and primarily serve as traffic channels for existing products.
3. Strong Platform Dependency: Plugin developers are completely dependent on OpenAI's policies and technology roadmap, facing high platform risks.

For example, while the Wolfram plugin has over 2 million monthly active users, its founder, Stephen Wolfram, stated in an interview that the plugin primarily serves as a promotional tool for the Wolfram Alpha main product, with limited direct revenue.

Custom GPTs Market

As a relatively new technology path, the GPTs market is still in its early stages of development. According to OpenAI's disclosure in early 2024, over 3 million custom GPTs have been created within three months of the GPT Store's launch, with approximately 20,000 publicly available. It is worth noting that:

1. Low Barrier to Entry, High Participation: Creation requires no programming knowledge, leading to a large influx of amateur creators, but high-quality GPTs are relatively scarce.
2. Emerging Profit Models: OpenAI has launched a revenue-sharing program, allocating revenue to GPT creators based on user usage. According to unofficial statistics, the top 100 GPT creators earn monthly incomes ranging from a few hundred to several thousand dollars.
3. Opportunities Emerging in Vertical Fields: In professional areas such as legal advice, copywriting, and educational tutoring, some custom GPTs have shown commercial potential.

For example, the "Legal Advisor GPT" created by a team of lawyers has attracted over 50,000 users within two months of its launch. According to the founder, it has indirectly generated approximately $120,000 in business revenue by directing traffic to its legal service platform.

AI Agents Market

Compared to the former two, AI Agents represent a direction with higher technical barriers and greater imagination. The market is currently in the technology validation and early application stage:

1. High Entrepreneurial Investment Heat: According to PitchBook data, startups focused on Agent technology attracted approximately $2.8 billion in venture capital in 2023, with generally high valuation multiples.
2. Enterprise-level Applications Lead the Way: Successful cases have emerged in enterprise scenarios such as customer service automation, data analysis, and software testing. For example, Adept's ACT-1 has partnered with several Fortune 500 companies.
3. High Technical Threshold: Developing full-featured Agents requires a deep AI research background and engineering capabilities, forming a high barrier to entry.

A notable case is Cognition AI, a startup focused on code generation Agents that raised $45 million in 2023. Its product, Devin, is known as the "first AI software engineer" and can independently complete the entire process from requirement understanding to code writing, testing, and deployment.

Comparative Analysis of Entrepreneurial Potential

Based on the current market status, we can compare the entrepreneurial potential of the three technology paths from multiple dimensions:

Entry Barriers and Competitive Landscape

Plugin development requires some programming ability but does not require AI expertise, custom GPTs have almost no technical barrier, while developing fully functional Agents requires strong AI research and engineering teams. From a competitive landscape perspective, GPTs have the most intense homogenization competition due to the lowest barrier.

Business Models and Monetization Capabilities

The plugin ecosystem has not yet formed a clear monetization model and mostly serves as a traffic channel for existing products. Custom GPTs have launched a revenue-sharing mechanism, but individual GPT revenue is limited. AI Agents have demonstrated strong willingness to pay in the enterprise market, with AutoGPT enterprise edition pricing reaching hundreds of dollars per user per month.

Monetization potential from high to low: AI Agents > ChatGPT Plugins > Custom GPTs

Platform Dependence and Risk Control

The three paths all have varying degrees of platform dependence risk:

• Plugins are completely dependent on OpenAI platform policies and traffic distribution
• Custom GPTs can hardly exist independently from the OpenAI ecosystem
• AI Agents, while potentially dependent on basic models, have the potential to build independent brands and user relationships

Independence from high to low: AI Agents > ChatGPT Plugins > Custom GPTs

Technical Barriers and Moats

From the perspective of establishing technical barriers, AI Agents, due to their complexity and professionalism, are the easiest to form a technical moat. Plugins can form certain barriers through proprietary data or algorithms, while custom GPTs have almost no technical barriers to speak of.

Moat potential: AI Agents > ChatGPT Plugins > Custom GPTs

Industry Case Analysis

Plugin Entrepreneurial Success Story: Canva Plugin

Canva's developed design plugin allows ChatGPT users to directly create and edit design works. Although the plugin itself does not directly generate profits, according to Canva's internal report, the new user conversion rate through plugin traffic is about 40% higher than traditional marketing channels, and customer acquisition costs are reduced by about 35%. This indicates that the value of plugins as a traffic-generating tool is significant.

GPTs Entrepreneurial Case: Financial Advisor GPT

The Financial Advisor GPT, created by a former Wall Street analyst, focuses on personal financial advice and portfolio analysis. The creator achieved monetization through a three-step strategy: first, provide basic consultation for free to establish a user base, then launch a $29.99 monthly premium version, and finally guide high-net-worth users to purchase one-on-one consultation services. According to the founder, it has accumulated 50,000 users within six months, of which about 3% have converted into paying users, with monthly revenue exceeding $40,000.

AI Agent Entrepreneurial Case: Adept

Adept's developed ACT-1 is an intelligent agent for enterprises that can automate complex workflows. The company completed a $400 million Series B financing in 2023, with a valuation of $2.5 billion. Its business model is based on enterprise subscriptions, charging based on the number of seats and usage scenarios. It has currently partnered with several financial, medical, and retail companies, and annual recurring revenue is reportedly over $20 million.

Entrepreneurial Opportunities and Strategy Recommendations

Based on the above analysis, we can provide targeted advice for entrepreneurs with different backgrounds:

For AI Agents with High Technical Thresholds

Best Suited Team: Technical teams with AI research backgrounds and sufficient funding to support long-term R&D

Opportunity Areas:

• Enterprise process automation Agents
• Vertical industry-specific Agents (finance, healthcare, law)
• Agent orchestration and management platforms

Strategy Recommendations:

1. Focus on solving specific industry pain points, rather than general-purpose Agents
2. Prioritize the enterprise market and establish a results-based pricing model
3. Build a data moat and optimize Agent performance through proprietary datasets

For ChatGPT Plugins

Best Suited Team: Companies with existing mature products, or entrepreneurial teams with unique data/APIs

Opportunity Areas:

• Professional tool integration (design, programming, data analysis)
• Real-time data services (finance, weather, transportation)
• Access to professional knowledge bases

Strategy Recommendations:

1. Treat plugins as customer acquisition channels rather than core products
2. Design a "free trial + paid premium features" conversion funnel
3. Seek cooperation with AI platforms other than OpenAI to reduce platform dependence risks

For Custom GPTs

Best Suited Team: Domain experts, content creators, individual entrepreneurs

Opportunity Areas:

• Professional consultant GPTs (law, tax, mental health)
• Educational training GPTs (language learning, exam preparation)
• Creative assistant GPTs (writing, design, music)

Strategy Recommendations:

1. Combine unique professional knowledge with customized instructions to create differentiated GPTs
2. Build a GPTs matrix to cover different stages of the user journey
3. Use GPTs as a traffic-generating front end to connect to higher-value back-end services

Future Development Trend Predictions

From the perspective of technological evolution and market maturity, the development trajectory of the three paths may be as follows:

1. Short-term (1-2 years): The plugin ecosystem continues to flourish, the GPTs market gradually stratifies, and a few leading GPTs form a brand effect; AI Agents mainly serve the enterprise market
2. Medium-term (3-5 years): The boundaries between plugins and GPTs blur, possibly merging into a unified ecosystem; AI Agents begin to enter the consumer market, with "personal digital assistant" products appearing
3. Long-term (5-10 years): Agent technology matures, and Agent-based entrepreneurship becomes mainstream, forming a new computing platform

It is worth noting that as technology converges, the boundaries between the three paths may gradually blur. The most promising entrepreneurial opportunities in the future may lie in the intersection of multiple technologies, such as professional Agent systems controlled by GPTs, or intelligent workflow platforms that can automatically call multiple plugins.

Conclusion: Entrepreneurial Potential Ranking and Final Recommendation

Comprehensive analysis of all dimensions, the entrepreneurial potential ranking of the three technology paths is:

1. AI Agents: Despite the high technical threshold and long development cycle, it has the largest technical barrier and commercial value space, suitable for entrepreneurial teams with technical backgrounds and sufficient funding.
2. ChatGPT Plugins: Medium technical threshold, can be combined with existing products to form synergy, suitable for companies or entrepreneurial teams with existing core products.
3. Custom GPTs: Lowest barrier but high differentiation difficulty, suitable for individual entrepreneurs or professional field experts as an entry point, but need to be combined with a more complete business model.

For entrepreneurs, they should make choices based on their own technical capabilities, resource endowments, and risk preferences. Regardless of which path you choose, the key is to find the combination of technology and real market needs, rather than simply following technological trends. As OpenAI Chief Scientist Ilya Sutskever said: "The truly successful AI products are not those with the most advanced technology, but those that solve real pain points for users."

In this emerging field full of possibilities, the real entrepreneurial opportunity lies not in choosing which technology path, but in how to transform technology into an effective solution to solve user problems.