Chat and messaging is possibly the most important piece of software/technology in all of our lives. However, it is also the vertical with the slowest innovation.
Traditionally, this is fine. Chat apps are difficult to use - whatever. But now, with agents capable of higher volume messaging and more automations, traditional messaging breaks down in the face of spam and innovation. Phone number identities are not built for agents, agents need a new messaging interface, and the fragmented protocol layer makes it difficult for them to navigate. A drastic innovation is necessary in this space.
Many people and startups are trying to map out the future of messaging; at Chert, we constantly think about that as well. Here are my thoughts on this.
We can broadly separate messaging into three layers:
- Identity layer: Who am I talking to? Currently, this is often a phone number, email, X handle, or WhatsApp ID.
- Protocol layer: How are the messages sent? This could include SMS, iMessage, RCS, Slack, Matrix, Discord, and more.
- Interface layer: How are the messages sent and read? Historically, these are the chat app interfaces. Increasingly, they are becoming APIs that agents can use.
Under this framework, we can see a few possible futures.
First, the future of messaging could originate at the interface layer. The idea is that chat apps are currently fragmented and difficult to use. They lack even basic features such as native searching or classifying within email. Since the interface is closed, you also can't build software, such as email clients or automations, on top of it.
Startups such as Beeper, OpenPhone, or iMessage infrastructure startups are all betting on the interface layer. Beeper wants to create a better chat app with a customizable interface that developers can build upon. OpenPhone specifically customized the interface for SMS to better suit the needs of businesses. iMessage infrastructure startups want to expose the iMessage API, which is also an interface, for people to build it into software.
I think that the future of messaging will come from the protocol layer. Instead of communicating through different channels with different people, agents gives us the opportunity to converse with anyone in the interface we are already most familiar with. The idea is that we can just text our own personal agent with what we want to say, and the agent will communicate with the person on the other end using their protocol or, even better, through an agent-native protocol between the two agents. On top of simple messaging, this can also involve payments or scheduling. Nothing has to be static anymore.
Agent identity is always a big problem. Maybe there could be a unification of identities? I don't know.
The question now is how this transition will happen. Would it happen from the consumers, or would it happen from the end of the businesses? For businesses, I would imagine it to look something like this: consumers/customers can scan a business QR code from any chat app they choose and communicate with a business/intermediary agent who will handle everything. This agent should also have context on both the customer/user and the business so that everything is familiar and personalized on both ends.
Just some thoughts, not too polished yet.