Marta Mendez

speaker

112 appearances 2 recordings 1 series first heard Jun 2026 last heard 30 Jun

Marta Mendez’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 2 in all, peaking in Jun 2026 with 2.

Appearances

newest first · ▶ plays the moment

Working Smarter · How agentic AI works behind the scenes to find the answers you need · 30 Jun 2026 podcast

Like the text context and figure out which chunks, which paragraphs, for example, are relevant to answer that question.

That was kind of like a pipeline, a workflow, right?

Now with a Gentic RAG,

we give more power to the LLM to make the right decisions.

So, yes, it searches.

We decide what filters it should have access to to try to find the right files faster.

File type filter, connector, people filter.

Yeah, like you decide those parameters.

And then the output is super important because instead of trusting that the search engine is going to give you the best quality results, you allow the LLM to actually look at the metadata of those results to decide which ones are actually useful to answer the query.

So here in a Gentic rag, instead of us like giving everything, the LLM can choose which of the results are interesting to continue with the rest of the pipeline.

So the choice of which tools are available, what are the inputs, and very important, what are the outputs, is how you help the LLM to make the most of the high reasoning capabilities that it has.

Exactly.

Before you would be like, no, I'm going to trust that search ranking is going to give me the top five results are always great.

So I'm always going to read the full document content and then I'm going to do the context selection.

It's kind of you're accumulating error as you go.

Whereas if you give like the full document content to the LLM, it has full fidelity in the answer it can produce.

So typically what you would see is the agent is actually giving you progress updates as to what it's doing, but we are not showing the full reasoning.

We are curating what we show to the user to make it a better user experience.

This is a design choice, I guess, like also like product choice on like how much level of detail do you want to show to the user.

Knowing that the goal is not to see updates.

Showing 61–80 of 112 · page 4 of 6