David Turetsky interviews Gabriella Garcia, COO and Chief People Officer at Warp, about how AI is reshaping the build versus buy decision in HR and people operations. They dig into why this question is changing now, what guardrails matter for sensitive data, and why people leaders need to take a more strategic role in technology decisions.
Gabby shares how her engineering background informs her view of modern people tech, while David adds real-world context from years of hacking together internal systems. Together, they explore how AI lowers the barrier to building, but also raises new questions around governance, cost, and organizational design.
Key topics
In this episode, Gabby Garcia explains Warp’s mission as an AI native employee management platform handling payroll, benefits, state tax compliance, and HR systems for more than 1,200 customers.
David and Gabby compare today’s AI-enabled build culture with older eras of custom tooling, including early internal dashboards and Access database work.
Gabby says the build versus buy equation has changed because non-engineers are now more technical and because tools like Claude Code and Lovable reduce the activation energy for prototyping.
The conversation centers on the risk of rogue projects, especially when people teams handle PII, PHI, SSNs, birthdays, payroll data, and other sensitive information.
Gabby points to Shopify as an example of a company encouraging broad internal experimentation, while also surfacing the downside of AI slop, access issues, and repetitive low-value tools.
She argues for cultural guardrails, approval rules, and manager review processes rather than blanket bureaucracy.
The discussion shifts to the engineering function, where Gabby says engineers are increasingly becoming prompt-driven problem solvers and thought partners rather than manual code producers.
David and Gabby agree that this changes organizational DNA, especially when non-engineering teams can build internal tools, bots, and workflows that still need governance.
Gabby makes the case that people leaders, not just CIOs or CTOs, should lead AI adoption strategy because they own talent, enablement, growth, and role evolution.
The episode closes on token spend, cost control, and the need for finance and people teams to understand model selection, usage caps, and the accounting implications of AI spend.
Timestamps
00:00 - Introductions and Gabriella Garcia’s background at Warp
01:35 - From Colombia and Miami to MIT and engineering at Facebook, Apple, and Google
02:38 - What Warp does as an AI native employee management platform
03:38 - Fun fact: Gabby’s unusual sense of balance
05:28 - Why build versus buy is changing in HR and people tech
06:10 - More technical people are now building prototypes themselves
07:37 - Low-code and AI tools lower the barrier to entry
09:45 - Why PII, PHI, and compliance make people tech risky
10:11 - Shopify’s “build whatever you want” AI experiment
12:28 - How long old internal tool-building used to take
13:48 - Why innovation needs risk awareness, not reckless experimentation
14:40 - Borrowing engineering-style PR reviews and testing for people tools
16:44 - The future of engineering in an AI-native workplace
17:41 - Engineers, designers, and operators as prompt-based builders
18:10 - AI as a thought partner, chief of staff, and first-pass reviewer
19:38 - Internal tools can be pushed to managers, but customer-facing work still needs engineers
20:22 - How AI changes the DNA of an organization
22:17 - Why the people function should drive adoption strategy
23:11 - OpenAI’s people leadership and the future of every role
24:55 - Why technical depth in a CPO matters for AI transformation
26:14 - People teams have been implementing tech for years, but AI is different
27:39 - Technology adoption takes time, and the AI curve is still early
28:45 - Bad HR data and weak products still block mobile adoption
29:11 - Skipping the messy middle of legacy systems in emerging markets
30:35 - How chat-based tools and image-based workflows improve accessibility
32:14 - Serving small businesses and frontline teams with practical AI
33:40 - AI may make customization accessible beyond large enterprises
34:38 - Token costs, model selection, and the ROI of AI tools
35:55 - One sales rep’s token bill as a cautionary tale
37:18 - Education gaps around model choice, data access, and rollout strategy
38:13 - Token spend as a new cost per employee problem
39:09 - Why accounting and capitalization rules matter for AI spend
40:09 - Tracking utilization and setting token caps
41:00 - Hardware choices, open source models, and reducing model spend
42:11 - A future conversation on tokenomics and people leader skills
43:08 - Moving from experimentation to pragmatic implementation
44:24 - Why every organization needs its own AI point of view
Notable quotes
“The people function is in service of the people.”
“There are ways in which we can absorb and implement a lot of the learning from other organizations.”
“My recommendations are likely not a great fit for your business.”
Powered by the WRKdefined Podcast Network.
Sponsors
