Lots of teams I've talked to build one or two internal tools for general use cases. The future of software is re-defining this, and increasingly, I think people are realizing the potential of doing the opposite.
The previous roadblock to creating tools that drive insight was the engineering challenge. It doesn't make sense to put effort into creating more internal software because they're not your core product. But what we miss is the importance of measuring how well things work, what happens inside the product, and what you can learn from these observations.
It almost feels like muscle memory that we limit the number of tools we build. We end up with a mindset where we build tools that are absolutely critical and ignore the rest. For a long time, the cost of building those outweighed the benefit. Even if you had a good idea, it just wasn’t worth asking engineering to build something custom.
Now we have LLMs.
The engineering roadblock is gone. When you build enough, you learn that random ideas that cross your mind no longer need prioritization. They can just happen.
We're shifting our messaging for Olive (YC W25) to highlight the importance of creativity in building internal software. When you can just ask in a prompt, and the building/schema reasoning/deployment/authentication are handled, you end up with a new challenge: what do we build?
We tried figuring this out by giving users suggestions based on their database schema. We still do, but I think there is more work to be done on creating comfort with the practice of building more tools, more often.
That could be a table to see your users. A dashboard with some charts you're curious about. A tool that will let you take notes on your users. The important part is that it's something and you can run an experiment to see how helpful it is. We're still learning the best tools to create ourselves.
Many great ideas for tools don't come from the roadmap.
They come from being able to try without friction.
We want to remove as much friction as possible while still giving customers creative control.