AI is taking over the world. We are going to see 2025 as the Year of AI Agents. This is going to be explosive.
The foundation is already laid, especially in DeFi. This will result in an onslaught of AI agents being created.
It is something that is getting attention. My guess is this becomes commonplace throughout the year.
Like most things in crypto, a fancy acronym is required. Enter DeFAI (DeFi + AI).
For many, this will be their first foray into a world of automation. It is something that major financial institutions have utilized for years. Proprietary software that cost hundreds of millions of dollars was the norm for major Wall Street firms. We are now starting to see the democratization of that.
Crypto is the area to build these. Not only is this the future, we are looking at an area where oppressive regulation is not present. The fact that assets are accessible outside major exchanges is a help.
DeFAI Starting To Ramp Up In Crypto
To most users, AI is really nothing more than a prompting system. People head over to a chat bot and prompt it, seeking to receive some information (or image). The key to this is the process and automation of the output.
DeFAI is similar. The difference is the output is some type of financial activity. Coupled with decentralized assets, it is an instruction to complete some transaction.
Another way of thinking about this is through the lens of tasks. AI agents are capable of handling useful tasks for people. What those are depends upon the design of the agent. That said, it becomes a force multipier, utilizing the massive ability to process information to complete the prescribed activities.
When it comes to finance, this has huge ramifications.
Here is an example of an agent:
With DeFAI projects like Orbit, this can all be accomplished with just a simple prompt! Users simply ask the agent “Can you deposit some ETH into Aave for highest yield on any network?” and the agent will be able to bridge your ETH if needed to the correct network to deposit your funds without you having to search for it. Orbit offers a variety of protocol-specialized agents that new users can interact with and browse what all these different agents offer. They can also just type into the Orbit agent a prompt like “Where can I deposit my USDC to earn yield?”, which Orbit will return a list of protocols that users can deposit their USDC into.
Consider how long it might take to reach this conclusion if a human did it. How many options exist for earning yield on ETH staking? Perhaps there are lists that people put together, coupled with oracles. However, there is a time investment. This is all part of the Coase theorem and how AI might allow that to become a reality. (More on that in future articles)
As we can see, this agent is a massive time saver. Due to the fact that it deals with finance, we can see a trangible result. If an individuals is able to get .25% or .5% more on a return, that could be significant depending upon what is staked.
AI Agents Will Scour Everything
We are heading towards the Agentic Internet.
As mentioned, major corporations are going to churn these out by the millions. Salesforce is already using it to assist in coding. This resulted in the company increasing productivty by 30% without hiring any more people.
For this reason, it is a positive to see this emerging in the crypto world. A lot of data is starting to find its way on-chain. Utilizing it in different capacities is what makes it valuable.
Ultimately, this will benefit developers:
Not only are these DeFAI projects able to assist in completing actions onchain, but they can also help out with heavy lifting on the research side. For many veteran users, scrolling through docs, white papers, and GitHub repositories are a necessary tasks before touching any project at all.
DeFAI teams, like Clustr Labs, are creating AI agents to scrape through project documentation and give an output on the most relevant and important information, saving users hours of research.
In other words, there will be nothing that is not read by agents.
Personalization
The major benefit of this advancement is personalization.
We saw early attempts at this with things such as recommend engines. Platforms tried to generate a personalized experience by recommending videos, songs or products based upon past activities. This is a rudimentary way of thinking about the concept of taking in data and providing an output.
Over the next few years, this will be enhanced. If we look at the financial world, the benefits become clear.
One of the major developments will be personalized financial services. For simplicity sake, think of this as one's own personal automated financial advisor. The process could start with the determination of goal, then create a strategy to achieve it. Agents will be used to automate the investing process, while also monitoring what is taking place.
For example, if there is a new option for ETH staking that significantly exceeds the present choice, the funds could be moved. Of course, the agent will research whether it is legitimate or not, checking to see if it is a scam.
AI is bringing the tools to average individuals. With DeFAI, we can see how the potential for personalized finance is getting closer. This will no longer be the realm of the wealthy.