Aug 24, 2026 · 32 min · 13 segments
**Manu Mehra** is Head of AMER Industries Strategic Deal Pricing at Databricks, with more than 12 years of experience across pricing, product, cloud, and AI, including Google Cloud and Thermo Fisher…
Manu MehraGuest
Mark StivingHost
I think the biggest change that is happening is if we think about the evolution of pricing, we had the subscription pricing and then we went to license-based pricing and the consumption-based pricing.


A good example to think about, if I'm serving 10,000 customers, Every customer has a very different outcome they're trying to achieve.

And that's why it is so different from how we have been doing pricing historically.

Because none of these traditional models, even though they can help evolve the model, but none of these exactly fit the puzzle.

A good example is, let's say, if I'm trying to automate some of my Jira tickets, for example, right? And I want to like do everything through an agent, right? Now, the idea is how many tickets is the agent solving on an everyday? What's the amount of rework the agent is doing? Is the agent really hallucinating? Can it be better done by a human? So I think that's where like, to answer your question, Mark, like that's why the AI is so different.

pricing now the tricky thing with outcome is it's very hard to standardize it because again going back to the same thing if i'm serving 10 000 customers every customer perceives a very different outcome so how do you standardize it because your billing systems need to be robust to standardize stuff like otherwise if we are building a custom model or a custom outcome for every customer it's going to be like so hard for our billing system so i think that's where the evolution needs to happen I don't know what the right answer is today because it's still a lot of work needs to be done in this area.

But I think that's where I see the difference between pricing AI versus doing the traditional and the more legacy methods.

Okay, I have to say that you redeemed yourself a little bit at the end of that answer because I was going to disagree with you vehemently.

And so there are so few examples in the world today where we can say, yes, I can attribute an agent did this.

I think the biggest change that is happening is if we think about the evolution of pricing, we had the subscription pricing and then we went to license-based pricing and the consumption-based pricing.


A good example to think about, if I'm serving 10,000 customers, Every customer has a very different outcome they're trying to achieve.

And that's why it is so different from how we have been doing pricing historically.

Because none of these traditional models, even though they can help evolve the model, but none of these exactly fit the puzzle.

A good example is, let's say, if I'm trying to automate some of my Jira tickets, for example, right? And I want to like do everything through an agent, right? Now, the idea is how many tickets is the agent solving on an everyday? What's the amount of rework the agent is doing? Is the agent really hallucinating? Can it be better done by a human? So I think that's where like, to answer your question, Mark, like that's why the AI is so different.

pricing now the tricky thing with outcome is it's very hard to standardize it because again going back to the same thing if i'm serving 10 000 customers every customer perceives a very different outcome so how do you standardize it because your billing systems need to be robust to standardize stuff like otherwise if we are building a custom model or a custom outcome for every customer it's going to be like so hard for our billing system so i think that's where the evolution needs to happen I don't know what the right answer is today because it's still a lot of work needs to be done in this area.

But I think that's where I see the difference between pricing AI versus doing the traditional and the more legacy methods.

Okay, I have to say that you redeemed yourself a little bit at the end of that answer because I was going to disagree with you vehemently.

And so there are so few examples in the world today where we can say, yes, I can attribute an agent did this.
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.