How Red Hat cleared IT debt for scalable AI ft. Marco Bill

Yeah, I would say before we go to AI, really the data has been a trick.

I mentioned in the complexity, data has been super complex.

And some is just in the past we had...

You have the master data, and then somebody wants to do some reports.

What happened is they would actually take the master data, copy it, be it a spreadsheet, be it a dashboard.

They will copy it, do their little work, and provide those numbers.

Let's pick the sales team, for example.

You have master data.

Somebody copies, manipulates it.

sends this to our sales leader, there will be different numbers, right? And this has been, I think, the last 20 years.

We always heard, oh, can't trust that data.

You got to take that data.

So we've done a massive cleanup.

And, you know, like in some cases, I think we had 73 copies in spreadsheets, every one a little bit different modified.

There's no way you can manage it.

The reason I'm saying this in the context of AI is If your data is not clean, forget it, forget it.

Garbage in, garbage out.

And so that was, I think we lucked out that two or three years ago, we really started on that data transformation, not really knowing that AI is... kick-starting exponentially the last two years, right? But I think this really helped us to get a bit the legs up.

I would do it again.

I would do it again, kind of clean the data.

And now I think the AI transformation, I think every IT company is, or every company, It's kind of experimenting, right? And in some ways, experimenting for the masses, right? As in all like the associates, right? Last year, we introduced a platform, the AI platform.

that everybody's using now, right? Never seen something adopted as fast as that 12 months ago, right? That's in literally 90% adoption in weeks because it gave such a value for associate productivity.

The next level for me is like in the AI transformation is really some experiments on some business transformations, right? How can you really use AI to make something go really fast and really accurate, right? And now it's my push is really like, how do we get the real ROI to really make a difference?