Chat interfaces have been getting a lot of heat recently.
The main criticism of chat interfaces it that they are considered nerdy. Developers are in love with their terminals, and are trying to impose a similar input experience on the general public. Isn’t it time we remember the world has transitioned to graphical interfaces?
The followup argument is chat gives you extreme freedom: is it too much? You can type in whatever you want: while this sounds great, users can feel confused about what they should ask for.
To be fair, they do have some valid points. Chat is be the first user interface we’ve used to interact with intelligent systems, but probably won’t be the last.
I think they are missing a key element, though.
The continuing relevance of terminals
Nobody1 is installing a GUI on a remote server: you just ssh into it, because it’s faster and requires no setup.
There are still valid use cases for a terminal: chat will follow the same trajectory.
How are terminals used? Off the top of my head, they are a good fit when you:
- just need a flexible way of interacting with a system under development
- only serve technical users - for example, many of our internal tools2 just have a command line interface
- need to provide a flexible admin-level access.
This is the lens we’re looking through when considering chat interfaces.
Chat interfaces will be superusers’ tool of choice to interact with any intelligent system.
One more thing
Interested in upcoming product features integrating this approach with our AI Engine? We’ll be announcing an addition to Connhex in a few weeks: make sure to subscribe to our newsletter to get notified once it’s out!
