Argilla's ReadMe

I worked with Argilla earlier this year, and I was extremely impressed with what they did with their ReadMe after we worked together. In fact, this is part of why I decided that going forward I’m going to work with companies on updating their ReadMe during the work we do together — because most of my clients don’t do as good of a job as Argilla and they definitely don’t act as quickly to get their ReadMe updated.

Why do I love this ReadMe so much? Read on.

The ReadMe starts with a really clear description of what Argilla is, and what sort of outcome a user can expect from using it. It’s worth pointing out that not all AI engineers care about high-quality outputs (though most probably won’t say that publicly) or full data ownership. If you clean data by paying an offshore team $2 an hour, this is not the platform for you.

Oh, and if you don’t really need to be sold on Argilla and just want to get started? That’s easy and there’s an easy link to do so. But…

This the section where this ReadMe is really exceptional, in my opinion. Because for a lot of users, you need to understand the ‘why’ before you care about the ‘how.’ This is where you outline your differentiated value, which is exactly what Argilla does here. If you care about these your AI output quality, control of data and efficiency, Argilla is for you and you should proceed to the ‘how’ section. If those things are not really that important to you…. maybe you can skip to the next project that does address whatever it is you care about.

The last section that I really want to pull out, which is examples of what others have done with Argilla. This falls into the category of ‘proof’ in a positioning canvas — they outlined the differentiated value the project provides in the ‘Why use Argilla’ section, and now they are giving concrete example to prove that you will get those outcomes. They are also doubling down on the idea that this is a project for people who care about high-quality outputs and who need to incorporate human feedback into the data cleaning process.

After this section, the ReadMe goes into talking about how to actually get started. That’s standard stuff, and it’s obviously necessary. The reason I was so impressed with this ReadMe is that it clearly articulates Argilla’s differentiated values, the outcomes a user can expect and when Argilla is most appropriate.

It wouldn’t be possible to write a ReadMe like this if you don’t have complete clarity on your differentiate value and your positioning. And if you’re struggling with that, you might want to work with me to get there.