Semantic layer

Um, the- Uh, to, to be like very specific, I think that, um, th- the, the things that, uh, we will see gravity around, uh, will be, uh, really at all the major layers of the stack.

I think that, uh, we're going to increasingly see, uh, driven by th- these kind of, uh, th- these kind of needs, um, gravity at the iceberg tier, at the, um, ingestion tier, at the transformation tier, the semantic layer tier, at the metadata tier.

Um, I think we're even gonna start seeing this at compute, um, increasingly, like capable open compute, uh, options are available.

Um, and even at the visualization tier.

I'm, I'm, what I'm not s- suggesting is like, well, just tell the agent what you, what business outcome you want and let it make every single decision along the way.

It's like, no, you need an architecture, but, um, as long as the right- ... pinch points are, uh, happen in that architecture, then, uh, you, you free up, uh, agents to make decisions from there.

So here's the, the reason that we failed on semantic layer from '21 to '24 is that we didn't have this agentic capability to integrate systems together.

And so we were like, "Here's this cool piece of technology.

Today's session is part of a series where we have the opportunity to speak with the key voices in the semantic solutions space, bringing together industry thought leaders, experts, and practitioners to explore the latest advancements and best practices in the implementation of semantic layer.

I am Lulit Tesfaye, partner and vice president of Knowledge and Data Management at Enterprise Knowledge, your host for the Semantic Layer series on this podcast.

And over the next few episodes, we are bringing you conversations with speakers from our Semantic Layer Symposium, taking place this fall in Vienna, Austria, October 14th and October 15th.

And as part of this series, we will be getting behind their sessions to hear what's really on their minds, what they are seeing, what they're looking forward to, what's working, what's not, especially in the world of data, AI knowledge, and semantic layers.

So you can consider this a backstage pass, if you will, of the Semantic Layer Symposium, and a chance to get to know our speakers behind their talks before they take the stage in Vienna.

Today, I am joined by Rosanna Stevens.

Rosanna is a senior global knowledge manager at Adobe.

She brings more than 20 years of experience in knowledge management, primarily focusing on knowledge AI readiness and employee self-service strategies at Adobe.

Give me the fix and start from zero, because I'm the exec who's never heard the phrase.

The fix is called a semantic layer.

Think of it as a single dictionary sitting between your raw data and every dashboard, chart, and deck.

Instead of each team writing its own formula for revenue, the definition lives in one place, and everyone pulls from it.

A dictionary? That's the whole revolution?

The revolution is that it's enforced.

Right now, active customer means whatever the analyst who built the report decided at 11 p.m. A semantic layer makes the definition a shared organizational fact.

You can't quietly redefine it in your corner.