Brian Stempeck
speaker
431 appearances
5 recordings
1 series
first heard Dec 2025
last heard 19 Dec
Brian Stempeck’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 5 in all, peaking in Dec 2025 with 5.
Appearances
Let's look at the single, more adventurous buyer.
They have their kind of precept personas that they use for media planning and ad buying.
And they're replicating that in our platform.
And so what they're doing is they're creating hundreds of different line items or variations of what response uh does their brand come back with.
So how often, just making this up, how often is Ford versus GM versus Toyota versus Honda being recommended in each of those categories?
Because the models give very different answers based on geography, persona, the matching of the user to the right make and model of the car.
And so with brands that we're working with, they're starting to develop really more of like a heat map.
Where do they have areas where they're quite weak?
And they need do they know that they're not being recommended for family buyers?
Where do they have areas where they're quite strong, where for SUVs, for um single adventurous types living in the city, they're doing well?
That understanding is kind of where the starting point of the models are not because they're probabilistic, because they vary the answers.
The starting point is a baseline of like, well, how well are you doing across geographies, personas, and product lines?
Which again is a very different way from how you would think about search.
Search was a is a bit more universal than this.
This is much more customized.
And I think that's really important for brands to consider, which is.
This is more like dealing with people, the same way that you might run a focus group or marker research.
We've used a lot of those learnings to inform how we sample the models.
That's how you very quickly get to hundreds of thousands or millions of prompts because you're exploring all these sub-variations of what the model might say for very specific use cases.
A ton.
Showing 121–140 of 431 · page 7 of 22
← Previous
Next →