The Rise of the AI Agent
Jack From Web3.com Ventures
Market Overview
With the recent boom in the artificial intelligence industry, AI Agent has been repeatedly mentioned in various media as the main application method of AI. The AI Agent industry has experienced significant growth over the past decade and has now become an important part of the AI market. According to the latest McKinsey Global Survey, by 2023, the AI industry will receive special attention due to the explosive growth of generative artificial intelligence tools, and this surge highlights the gradual increase in the adoption rate of AI technology in various fields. .
Looking at the overall market, the overall AI market size will be $428.0 billion in 2022, and is expected to grow to $515.31 billion in 2023. By 2030, it is expected to reach $2025.12 billion, with a CAGR of 21.6%. The global AI market will be worth $136.55 billion in 2022, and the CAGR from 2023 to 2030 is expected to be 37.3%. This growth is driven by an increasing number of companies and collaborators adopting AI technology, indicating that AI applications in various industries will integrate more and more functions to achieve broader applications in the next few years.
In the next section, we will start with the definition of AI Agent and dive deeper into this emerging technology. Get a glimpse of a vibrant and rapidly growing AI industry through some influential players in the field of AI Agents.
Understanding AI Agents
AI Agent refers to systems and software programs that can perform tasks, make decisions, or take actions on behalf of humans or other programs. These agents are designed to observe their environment (which can be digital, physical, or both), interpret data, and act based on that data to achieve specific goals.
The overall development of AI Agent shows two major trends, Autonomous Agent and Generative Agent. Autonomous Agent, represented by AutoGPT, has strong tool attributes and is designed to assist humans in performing tasks. On the other hand, Generative Agent has certain human-like characteristics and has a stronger ability to perform tasks autonomously, represented by Smallville Town proposed by Stanford researchers.
Today’s AI Agent field has not yet seen a dominant application, and a large number of companies are still exploring the commercialization of AI Agent. The AI Agents developed by these three companies have great influence in their respective fields.
- Perplexity AI: Founded in 2022, Perplexity created the “first” conversational search engine by combining AI chatbots with traditional search engines such as Google. The platform allows users to perform searches in a conversational manner, leveraging advanced LLM to solve and interpret follow-up questions. Its working mode is to disassemble user questions and search for keywords, arrange the answers logically, and provide feedback and conclusions to users based on the answers. Note that unlike ChatGPT, Perplexity’s answers will have source citations, rather than “original” text produced by machines. Perplexity provides beginners with a simple and effective tool to help them quickly become experts in a certain field.
- Hugging Face: Hugging Face stands out for its open source collaborative AI community, which is known as the Github of AI because it provides necessary tools for AI developers. In the past eight months, the number of pre-trained models on Hugging Face has grown from 100,000 to more than 300,000, the number of datasets has grown from 10,000 to 58,000, and the valuation has soared from $2 billion to $4.5 billion, making it the undisputed center of open source AI development and the world’s largest LLM library. In August, Hugging Face completed a new round of financing of $235 million, with investors including Salesforce, Google, IBM, Amazon, Nvidia, Intel and other famous technology companies.
- VoiceFlow: An AI Agent building platform that has now attracted more than 150,000 users from influential technology companies. VoiceFlow just completed their new round of financing in August, with the current valuation reaching $105M, a 50% increase from the previous round. Investors include Amazon, OpenView, Google and other institutions. As an AI Agent production platform, VoiceFlow is characterized by ease of use and adaptability to team work. Creating an AI Agent in Voiceflow does not require any coding skills. Beginners can quickly build a voice customer service AI Agent by simply dragging and dropping the basic template block. For more professional developers, they can use advanced mode to add logic and variables to the Agent in the development process, and integrate enterprise APIs into the AI Agent process. This modular development model enables visualization of the overall process, which can help designers improve product logic links. After the recent introduction of its Generative AI function, Agents developed based on VoiceFlow can give more reasonable responses by reading customers’ past interaction records.
Challenges and prospects
Although companies above have shown great development potential in the field of AI Agent, on the road of higher adoption of AI Agent products still faces various obstacles, ranging from technology to regulations.
The first is that the transition from independent products to integrated functions currently lacks a clear and replicable path. The current overall trend in the AI Agent industry is shifting from independent AI Agents to being integrated into larger products as part of the functionality. For example, the GPTs community platform just launched last week allows users of ChatGPT to customize their own GPTs through chat and share them with other users. With GPTs, users can easily access AI Agents with different characteristics to assist in performing tasks.
The second issue is the reliability and performance of AI Agent. In order to be adopted by wider enterprises, AI Agents need to demonstrate a high level of ability in the corresponding field while maintaining the reliability of performing tasks. Current factors that limit the performance of AI agents include issues such as latency, evaluation, debugging, and privacy issues, which hinder their reliability and performance in enterprise environments.
Finally, there is the establishment of relevant standards in the field of AI Agent. The AI Agent community is discussing establishing a common framework and standards that meet current industry needs, benchmarks for evaluating Agent performance, and incorporating safety and ethical considerations.
Despite these challenges, the outlook from capitals for the AI agent industry remains positive. The market is valued at $4.8 billion in 2023 and is expected to reach $64.69 billion by 2030, with a CAGR of 45% during the period. This sharp growth trajectory demonstrates the industry’s rapid expansion and its potential for more significant developments in the coming years.
