@tjrobertson52

Right now, every serious business is trying to figure out how to incorporate AI into their workflow.

The most important part of that is how you structure your company's knowledge. As of today, June 12, we have the very first standard for how that knowledge should be structured. This is the moment we've been waiting for, and as I guessed, it's Google who's putting forth this standard.

No surprise there.

They just put out a blog post announcing what they're calling the Open Knowledge Format, or OKF. I'm so excited about this. I already recorded a video, and I had to record another one, even though it's 100 degrees outside right now.

If Google is successful in getting businesses to adopt this standard, and I think they will, they've had a lot of success with Universal Commerce Protocol and Web MCP. Getting your company's knowledge base set up using this standard is going to be the most important step for you to take before integrating AI.

If you watch my videos at all, you know how bullish I am on setting up a knowledge base like this for your company. We build this out for all of our clients. We call it their brand ambassador, and me and my team are already talking about how we might adapt our current brand ambassadors to this new standard.

So what makes this standard so great? The standard itself is pretty simple. I mean, it makes sense.

I like it. But what actually makes it great is just that there's a standard at all.

If we can all agree how your knowledge should be structured, then every AI agent or harness or prompt or skill can assume knowledge is structured in that way, and then we move from having to create custom processes in isolation and into a world where we can all share and iterate on each other's processes.

This absolutely has to happen for us to make quick progress. We still don't know yet if Google Standard will be adopted, but I mean, come on, it's Google. There are no real competing standards right now, and it does make sense.

Alright, TJ, just tell us what it is already.

First of all, it's just files. There's no SDK or compression schema that you have to worry about. It's just a series of markdown files in a nested hierarchy.

Within each folder, you can have an optional index file that describes all the other files in that folder, and at the top of each markdown file, there's a mandatory YAML front matter.

If that sounds complex at all, trust me, it's not. It's just some text at the top of each file defining fields like title, description, tags, type, resource, and timestamp. Only a few of these are mandatory.

That just lets the agent quickly look up relevant files, and then you can add as many custom fields as you want.

1 of the standard. It'll probably change some, but I think it makes sense in its current form. I think the biggest changes will come from people actually implementing it and seeing what best practices emerge, and then over time, those best practices will likely be integrated into the standard.

Where will these files live? Well, ultimately, you want all this knowledge to be platform agnostic, and you want each file to live in a single location. The way things are heading, I'm guessing most businesses will set this up inside of Github.

That's what we're planning to do ourselves.

It's a huge rush to adapt to this standard. However, it's so close to the best practices that already exist, and it makes so much sense given where AI is headed, that I think you'd be safe adopting it now.

And if you haven't already started collecting this knowledge base, I highly recommend you start doing that immediately.