Uploaded February 2026 | Updated September 2026, 1 week ago
The hardest part of building AI products is deciding which workflow you’re willing to own.
On this episode of Builders, Aatish Nayak (Harvey) and Sachi Shah (Sierra) share how they’ve made those calls across legal and customer experience, and what they learned building systems customers rely on every day.
Highlights from the discussion:
“Traditional product management says, pick a user, pick their use case, solve for their use case, make it super narrow. With AI, you can actually broaden out the number of use cases.” - Aatish Nayak, Harvey
“You have to constantly re-earn product market fit all the time. Customer expectations change.” - Aatish Nayak, Harvey
“We had this rule of thumb where every PM should talk to at least five customers a week. So I still try and stick with that personally, where there’s nothing that substitutes going directly to the customer.” - Sachi Shah, Sierra
“SRE teams work with things like error budgets ’cause we’ve understood that a hundred percent uptime goal or reliability goal is not really a goal. And I think the same is true with agents, where you don’t wanna let perfect be the enemy of good.” - Sachi Shah, Sierra
Guest: Aatish Nayak, VP of Product at Harvey and Sachi Shah, Product Manager at Sierra
Connect with Aatish Nayak
X: https://x.com/nayakkayak
LinkedIn: linkedin.com/in/aatishn
Connect with Sachi Shah
X: https://x.com/SachiAShah
LinkedIn: linkedin.com/in/sachi-shah-26920336
Connect with Josh Coyne
X: https://x.com/josh_coyne
LinkedIn: linkedin.com/in/joshuacoyne
Connect with Leigh Marie Braswell
X: https://x.com/LM_Braswell
LinkedIn: linkedin.com/in/leigh-marie-braswell
Learn more about Kleiner Perkins: kleinerperkins.com
00:00 Trailer
01:03 AI Applications in Enterprises
02:32 Transactional Work and AI
04:07 Building Effective AI Agents
11:56 Forward Deployed Engineers (FDEs)
20:13 Setting Roadmaps in AI Companies
25:34 Customer Feedback and Adaptation
29:58 Optimizing AI for Customer Experience
36:56 Evaluating AI Models
39:29 Trust and Transparency in AI
45:34 Exciting AI Application Categories
49:06 Competitive Advantage in AI
The hardest part of building AI products is deciding which workflow you’re willing to own.
On this episode of Builders, Aatish Nayak (Harvey) and Sachi Shah (Sierra) share how they’ve made those calls across legal and customer experience, and what they learned building systems customers rely on every day.
Highlights from the discussion:
“Traditional product management says, pick a user, pick their use case, solve for their use case, make it super narrow. With AI, you can actually broaden out the number of use cases.” - Aatish Nayak, Harvey
“You have to constantly re-earn product market fit all the time. Customer expectations change.” - Aatish Nayak, Harvey
“We had this rule of thumb where every PM should talk to at least five customers a week. So I still try and stick with that personally, where there’s nothing that substitutes going directly to the customer.” - Sachi Shah, Sierra
“SRE teams work with things like error budgets ’cause we’ve understood that a hundred percent uptime goal or reliability goal is not really a goal. And I think the same is true with agents, where you don’t wanna let perfect be the enemy of good.” - Sachi Shah, Sierra
Guest: Aatish Nayak, VP of Product at Harvey and Sachi Shah, Product Manager at Sierra
Connect with Aatish Nayak
X: https://x.com/nayakkayak
LinkedIn: linkedin.com/in/aatishn
Connect with Sachi Shah
X: https://x.com/SachiAShah
LinkedIn: linkedin.com/in/sachi-shah-26920336
Connect with Josh Coyne
X: https://x.com/josh_coyne
LinkedIn: linkedin.com/in/joshuacoyne
Connect with Leigh Marie Braswell
X: https://x.com/LM_Braswell
LinkedIn: linkedin.com/in/leigh-marie-braswell
Learn more about Kleiner Perkins: kleinerperkins.com
00:00 Trailer
01:03 AI Applications in Enterprises
02:32 Transactional Work and AI
04:07 Building Effective AI Agents
11:56 Forward Deployed Engineers (FDEs)
20:13 Setting Roadmaps in AI Companies
25:34 Customer Feedback and Adaptation
29:58 Optimizing AI for Customer Experience
36:56 Evaluating AI Models
39:29 Trust and Transparency in AI
45:34 Exciting AI Application Categories
49:06 Competitive Advantage in AI
