Hello all,
In the past few years, we've seen a wave of SaaS applications rushing to add LLMs to every feature they can find. Most of these additions feel rushed, driven by the fear of missing the hype train rather than a desire to solve a problem. As a result, the UX suffers. The default approach seems to be a generic chatbot tucked in the corner of the UI or a million annoying pop-ups that try to beat you into submission.
About a year and a half into working on RelayGrid, I decided it was time to add AI capabilities to the platform. I waited for two reasons. Firstly, I did not feel like it needed it; secondly, I felt the core features needed to be polished enough that adding AI would be a legitimate feature, not a bandaid.
To give some context: RelayGrid is finally at a point where I can "dogfood" its own features. I am now sending notifications to my own users using RelayGrid. The first step in that process is creating a Notification Template.
They say that to build a great product, you should solve a problem you personally face and I've always had a hard time writing these types of message bodies. So I built an AI Assistant specifically to help me write template subject and bodies.
It isn't an autonomous agent. It is specialized tool designed to generate content. By treating the LLM as a "helper function" rather than a primary interface, I was able to keep the UI focused on what actually matters: the reliability an delivery of your messages.
We are officially moving out of beta.
You can sign up and start exploring for free at relaygrid.dev, I'd love to hear your thoughts on the new template assistance (with 14-day free trial) specifically if you think the "helper function" approach actually works for your workflow.
Email me at ali@relaygrid.dev or on our support page with any questions or issues.