Jul 21, 2026 · 51 min · 9 segments
Silicon Valley says move fast and break things. **Shensi Ding** built Merge by doing the opposite: nine months of building with zero users, obsessing over margins while everyone else chased growth at…
Shensi DingGuest
Brennan PothetesHost
And like my experience, like I worked at a bank and a life insurance company and trying to find being a data analyst or trying to find data at those companies, I'd be merging four or five different tables together.
And then you'd usually... pieced together a lot of stuff to then infer a data report.
I remember in life insurance, they'd wheel up a physical cart that would go through a life insurance claim.
And so in that business, there's a ton of human labor that is going and finding that stuff.
And with Merge and with with AI search, essentially, that's all persistent.
How would you describe like AI search to someone that's like not an industry?
It's just like your brain, like you're pulling all the data, like these are my files from like 10 years ago.
These are all the customer interactions I've had in the past, like five years only.
Here's some of like the transcripts, you're pulling all the information that you would want to be able to search across because there might be something that you might need in the future.
after that, you're able to like try to combine the information to find something or you're able to be pretty like niche and like emotional, honestly, about like what you're trying to look for.
So you can be like, I want to find the customers that are most unhappy about this feature.
And it's able to combine data from multiple sources in order to return like a best answer for you.
Another example is like where you want like an output based off of all the information that you then pulled in.
So it's like, I want to be able to create like a report on which customers are most likely to be upsold based off of like historical data that you've seen from other customers.
And like my experience, like I worked at a bank and a life insurance company and trying to find being a data analyst or trying to find data at those companies, I'd be merging four or five different tables together.
And then you'd usually... pieced together a lot of stuff to then infer a data report.
I remember in life insurance, they'd wheel up a physical cart that would go through a life insurance claim.
And so in that business, there's a ton of human labor that is going and finding that stuff.
And with Merge and with with AI search, essentially, that's all persistent.
How would you describe like AI search to someone that's like not an industry?
It's just like your brain, like you're pulling all the data, like these are my files from like 10 years ago.
These are all the customer interactions I've had in the past, like five years only.
Here's some of like the transcripts, you're pulling all the information that you would want to be able to search across because there might be something that you might need in the future.
after that, you're able to like try to combine the information to find something or you're able to be pretty like niche and like emotional, honestly, about like what you're trying to look for.
So you can be like, I want to find the customers that are most unhappy about this feature.
And it's able to combine data from multiple sources in order to return like a best answer for you.
Another example is like where you want like an output based off of all the information that you then pulled in.
So it's like, I want to be able to create like a report on which customers are most likely to be upsold based off of like historical data that you've seen from other customers.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.