Andreas Welsch

Right.

So we realized very quickly we need to feed data to these models, augment the prompts, things like retrieval, augmented generation as one mean, or give agents tools, give agents access to specific data sources so we can make that specific.

And data is never as clean, as good, as complete, as fresh, as accurate as you want or need it to be.

Because most of the time it's data observed in the real world through real life interactions of people with your company.

Think of customers or people within the company, within departments and what have you.

So focus on people.

Make sure you bring them along on this transformation journey.

And number two, focus on data so you get real good quality specific results for your business.

Absolutely, right?

Otherwise, you'll get the generic job description, if you will, for a podcast host or for a show writer, unless it's specific to the company or to the show or, again, whatever situation you have.

It's going to be that same bland generic output that is not wrong, but it's also not great, right?

Exactly.

So you see large enterprise software companies make a play and make a push in these directions.

For example, Salesforce recently acquired Informatica.

That's a big data, a huge data play, not just a big data play.

You're seeing others like SAP making similar moves saying, hey, we have applications, we have data, and we bring that to AI.

So funny enough, it goes full circle from AI.

It goes back to data.

How can we bring in data to the environment?

How can you use the data you already have to augment what we want these models to do?