Zapier’s Chief Says the Real AI advantage is Culture, Not Technical Sophistication
Discover how Zapier scaled AI across hiring, performance, and daily work by focusing on culture, learning speed, and human potential instead of tools alone.
Brandon Sammut
Chief People and AI Transformation Officer at Zapier
AI-Driven Transformation: True AI transformation is a people challenge, focusing on integrating AI into organizational culture.
Impact on Jobs: AI is reshaping jobs by automating repetitive tasks, allowing more creative and strategic work.
Performance Enhancement: AI improves performance by reducing task time and enhancing team productivity and salaries.
Simplicity Over Complexity: Valuable AI solutions often arise from straightforward use cases solving real human problems.
Cultural Priority: Successful AI transformation requires a culture supporting experimentation and learning, not just technology.
Brandon Sammut is the Chief People and AI Transformation Officer at Zapier. He and the Zapier team have spent the past three years integrating AI into their systems, workflows, and people practices.
We sat down with Brandon to learn why his role went from "People" to "People and AI Transformation." He told us that true AI transformation is a people challenge, not a technical one.
Building Systems That Maximize Human Potential
I'm Brandon Sammut, Chief People & AI Transformation Officer at Zapier. I focus on the future of work at the intersection of talent and technology.
Keep Reading—and Keep Leading Smarter
Create a free account to finish this piece and join a community of forward-thinking leaders unlocking tools, playbooks, and insights for thriving in the age of AI.
Have an account? Log In
My path here is anything but linear. I started in education—first as a residence director at the University of Texas at Arlington—then spent six years at Teach For America across recruiting, finance, ops, and corporate partnerships.
Later, I got an MBA and a Master's in Education at Stanford, interned in venture capital at Owl Ventures, and eventually landed in my first CPO role at LiveRamp (NYSE: RAMP). What those experiences share is a thread that deeply motivates me. How do we build systems that maximize human potential, and how do we scalably match that potential to opportunity?
I joined Zapier in 2021, before the gen AI revolution. In March 2023, our CEO issued a company-wide "AI Code Red"—a call to action that shifted everything. We realized that Zapier’s AI opportunity (and the risk of inaction) was massive.
Our People team and I have spent the last three years implementing AI transformation across the entire employee experience—from how we hire, to how we develop people, to how we redesign work itself.
In late 2025, we formalized this work by expanding my role to explicitly own AI transformation alongside People. The logic was simple: true AI transformation is a people challenge, not just a technical one.
Zapier is an AI orchestration platform with an 800-person, fully remote team across 40+ countries. My scope spans the full People function, from talent acquisition to talent development, people operations, total rewards—plus Zapier's company-wide AI transformation effort. That includes our AI Transformation Center of Excellence, 11 embedded department-level AI transformation pods, and our open-source AI Transformation Framework shared publicly with the broader market.
In short: I'm accountable for both the humans and the AI systems that help those humans do their best work. I also work with our customer-facing teams to help other organizations succeed within their own AI transformations.
True AI transformation is a people challenge, not just a technical one.
How AI Makes Work More Human
Across our company, 97% of employees use AI coding tools in their core work. Eighty-eight percent say it has measurably improved their performance and employee engagement is above 80% in our pulse surveys. This tells me AI isn't eroding the employee experience at the moment, but we say this with humility and remain curious about AI's impact on how people experience their work.
Jobs are becoming more interesting, and that's the qualitative story I'm most excited about. When we automate repetitive tasks, people do more creative, strategic, human work.
Our talent acquisition team routinely gets accolades for candidate experience. Not because they have an unusually large team, but because thoughtful orchestration of automation and AI increases personalization.
New career paths are emerging. People who were doing manual, repetitive work are now building automations, staffing AI transformation roles, and contributing to entirely new functions.
AI Strips Away Low-value Work and Reveals Elite Performers
Our Customer Support team reduced average ticket handle time by 50% through AI-powered summarization. That means they now answer customer questions twice as fast. That same team is now paid at the 90th percentile of the market. That's not a coincidence.
When AI strips away low-value repetition, people do genuinely hard, high-judgment work — and you can pay them accordingly. They're elite performers producing valuable outcomes.
Our Demand Gen team built an AI-powered workflow that automated lead research and generated personalized outreach for AEs in real time. That reclaimed 4.5 FTEs of capacity, saved $15–20K in tooling costs, and generated $1.2M in pipeline. One workflow.
Revenue Operations now runs 25 AI workflows and agents, saving the equivalent of 34 full-time work weeks every month. Finance reduced month-end close by 25% despite increased complexity — with an 8-person team for a business exceeding $5 billion in annual revenue.
Where it's been hard is that the pace of change creates fatigue. Not everyone moves at the same speed, and that's okay. But managing that unevenness across 800 people takes intentionality and some AI-powered workflows needed multiple iterations before producing outputs we were genuinely happy with.
I don't call those failures. That's the nature of the work. The intelligence keeps growing. Something that's not quite there today could be a completely different story in a month.
Why Simplicity Trumps Complexity in AI Initiatives
Early on, my team and the company reflexively believed that the most valuable AI use cases would be technically sophisticated. This led to over-engineering. In reality, some of our highest-impact work has been fairly straightforward: automating candidate communications personalized to each role, building a simple AI coach for goal-setting, and summarizing long support tickets into actionable insights.
If I could go back, I would tell myself to start with the simplest build that solves a real human pain point. Don't confuse technical complexity with business value. The question isn't "How impressive is this?" It's "Does this genuinely improve someone's work?"
How an AI Fluency Assessment Raises the Bar for New Hires
We now evaluate AI fluency across 100% of candidates, for every department and every role. This is the single biggest operational change we have made because of AI. AI is becoming table stakes for how work gets done.
We built an AI Fluency Framework with four levels: Unacceptable, Capable, Adoptive, and Transformative.
- "Unacceptable" means someone is actively resistant to AI or has never used it in a purposeful way — that's a non-starter for us.
- "Capable" means you've repeatedly used AI with purpose and show forward momentum in learning — we look for real examples they can describe, even simple ones.
- "Adoptive" means you've moved beyond casual use into regular, purposeful application with real impact — we look for clear ROI stories.
- "Transformative" means AI has fundamentally changed how you work, and that new way of working influences your team, projects, or organization — not just your own productivity.
Across systems, we embedded AI fluency assessment into our structured interviews and built a Bar Raiser Interview Agent — an AI-powered tool that generates tailored interview questions based on the role, level, and our rubric. It explains why it chose each question, helping interviewers calibrate consistently across our 800-person organization with 10 VPs.
We trained our recruiting team to evaluate fluency and calibrated what "good" looks like at each level by role. Behaviorally, hiring managers now have a shared language for discussing AI readiness.
For outcomes, the quality of experimentation and peer-to-peer AI learning accelerated noticeably once every new hire arrived with at least baseline fluency. It raised the floor for the whole company.
How AI Enhances 1:1 Meeting Effectiveness
I'll walk through our One-on-One Meeting Coach because it's our most sophisticated People-team build and illustrates how AI can augment (not replace) the human side of management.
Problem: Every people leader knows that the quality of one-on-one conversations between managers and their direct reports is one of the strongest predictors of engagement, retention, and performance. But most managers get little real-time coaching on how they're showing up in those conversations. Are they asking enough questions? Are they doing too much of the talking? Are they creating space for the other person?
AI Solution: We created an AI-powered one-on-one meeting coach. After a one-on-one meeting (with full transparency to both parties — this is opt-in and visible), the system analyzes the conversation and generates coaching feedback for the manager.
End-to-end flow:
- Manager and direct report hold their regular one-on-one (nothing changes about the meeting itself).
- Post-meeting, the AI analyzes the conversation structure — talk-to-listen ratios, types of questions asked, and whether the manager created space for the direct report's priorities.
- The coach generates personalized feedback, such as “You spoke 70% of the time in this meeting — consider asking more open-ended questions,” or “You asked great clarifying questions, but you missed an opportunity to coach on the problem your report raised.” The coach delivers feedback only to the manager — private, constructive, and immediately actionable.
- Over time, the manager can see patterns across multiple one-on-ones and track their growth,
Why it works: It solves the "HR Business Partner scaling" problem that every company faces. We can't put an HRBP in every one-on-one meeting. But we can provide every manager with a private coach that helps them get better at the conversations that matter most—without requiring them to ask anyone for help or to admit they're struggling.
The trust architecture is everything: What data is used? Where is it stored? Who sees it? We're radically transparent about all of this. It's an "Iron Man suit" for managers — it enhances their capability — rather than an AI overlord watching over their shoulder. That distinction matters enormously for adoption.
I'll also share a simpler but equally powerful workflow: our AI Feedback Coach, which we use when preparing to give constructive feedback. It's trained on our internal feedback framework and exemplars of excellent constructive feedback.
An employee can engage privately, describe the situation, and get help structuring their feedback before a live conversation. It solves the human problem of: "I know I need to give this feedback, but I don't know how to say it skillfully, and I don't want to talk to another human about it yet because the context is too sensitive." The AI removes that barrier entirely.
It solves the human problem of: ‘I know I need to give this feedback, but I don’t know how to say it skillfully, and I don’t want to talk to another human about it yet because the context is too sensitive.’ The AI removes that barrier entirely.
Why AI Integration Challenges are Bigger Than Tech Failures
To be perfectly open, the places where AI has underdelivered have been less because the technology failed and more because of the integration challenge. The biggest gap has been transitioning from individual to institutional AI adoption.
While 97% of our team uses AI, fundamentally reshaping core business processes end-to-end with AI is a different mountain to climb.
We picked the low-hanging fruit fast. The harder, higher-value work — redesigning workflows from first principles and getting AI-powered tools to talk to each other in strategic orchestration — is slower and messier than I anticipated.
Internally, we call this the shift from "individual AI" to "institutional AI at scale," and it's our primary focus for 2026. It requires shared authentication infrastructure, cross-functional data foundations, and fundamentally rethinking how teams hand off processes. That's architectural work, not just experimentation.
One more honest admission: AI vendor overload is real. The number of tools claiming to solve AI transformation is overwhelming. Another lesson we learned the hard way is that consolidating into a small number of core platforms is essential, rather than accumulating ten-plus point solutions. We now track usage, reliability, and ROI for each tool and actively sunset those that do not integrate.
And the last drawback of AI: There has been an erosion of accountability because of the convenience AI creates. When AI generates outputs quickly and easily, leaders may ship things without genuine human review. The quality bar can drop for speed if leaders are not careful. If AI contributes to the work, humans remain accountable for its accuracy, tone, and impact.
How AI Changes the Need for a Formal Curriculum
I held the assumption that capability development requires a formal curriculum, but I was wrong. For years, the standard playbook in People/L&D has been to identify a skill gap, design a training program, deliver it, measure completion.
AI broke that model for us. The technology shifts so fast that by the time we could write a traditional curriculum, it would already be outdated.
We replaced it with a learning philosophy anchored in rate of learning over stock of knowledge. We don't have a formal AI curriculum at Zapier. Instead, we integrate AI learning and experimentation into as many of Zapier's existing practices as possible — hack weeks became AI hack weeks, onboarding became an AI fluency onramp, peer-to-peer Q&A channels with assigned AI power users became the support layer.
The deeper insight here is how much we know at any given point is less important than how fast we're learning. And the ingredients that maximize rate of learning — psychological safety, experimentation culture, sharing what works and what doesn't — are organizational culture ingredients that not every company has in place.
Why AI Transformation is a Cultural Challenge
AI transformation is a people challenge, not a technical one. Every company is talking about AI transformation, but few are doing it well because they focus on tools first.
True transformation starts with people, culture, and leadership. If your culture doesn't support experimentation — if people don't feel safe trying, failing, and sharing — no amount of AI tooling will get you where you want to go.
The People function has rarely been more important than in the AI era. A lot of things we're trained in — workforce planning, change management, leadership development — are more important now, not less.
The organizations that will learn AI fastest are the ones where people feel safe trying things that might not work, sharing those experiments openly, and learning from failure publicly. If your culture punishes failure or hoards knowledge, AI tools won't save you.
We've spent as much time ensuring those cultural components are in place as we have on traditional skilling. The tools are table stakes — the culture is the differentiator.
Why Performance Management Needs a Redesign
Performance management is one area that should be actively redesigned for an AI-augmented future. The traditional performance review cycle — annual or semi-annual, backward-looking, heavily dependent on a manager's memory and writing skill — needs to go.
At Zapier, we've already started redesigning this across multiple layers:
- Layer 1 – Evidence collection (the "Blob Notes" workflow): We built a dead-simple Slack integration. When something noteworthy happens — a great piece of work, a moment of leadership, a growth area — a manager or team member reacts to the relevant Slack message with a specific emoji (we call it "blob notes": 📝). That reaction triggers a Zapier workflow that captures the message, summarizes it using AI, and logs a clean performance evidence entry into a structured document. No forms, no context-switching, no relying on memory months later. We capture evidence in the moment, with a single emoji click.
- Layer 2 – Review drafting (the AI Performance Review Coach): When review time comes, you provide the name, job level, and accumulated performance evidence. The AI drafts a review aligned to our Impact Behaviors — our performance standards differentiated by job level — using concrete examples of excellent reviews as reference. This results in clearer, more evidence-based, and more actionable reviews, dramatically reducing the time tax on managers.
- Layer 3 – Goal-setting (the AMP Goal Coach): A chatbot helps all employees set actionable, measurable, and purposeful goals in just a few minutes through guided prompts. Turns messy intentions into clear, trackable outcomes.
But these are still AI adoption — making the existing system work better. The transformation opportunity is much bigger.
What if performance evidence was captured continuously and passively across all work surfaces — not just Slack? What if AI surfaced patterns across an entire team's performance data that a single manager would never see? What if feedback coaching were available in real-time, idling in Slack, not just before a scheduled conversation?
We're heading in that direction. The leaders who redesign this system now — while maintaining human accountability for the actual decisions — will have a significant advantage in talent density and employee experience.
How AI Reshapes Leadership and Staffing Strategies
We added no new headcount for our AI transformation across the entire business, not just HR. We moved no reporting lines. This was a very intentional decision.
Instead, we built a hub-and-spoke staffing model. The hub, our AI Transformation Center of Excellence, is a small team that sets standards, manages shared tools, maintains our AI transformation scorecard, and helps teams scale what works without reinventing it.
The spokes are 11 AI Transformation Managers embedded inside each department (Build, Sales & Success, Marketing, Support, People, Exec Ops, Finance, Data, Ecosystems, BizOps, Legal). Each manager is accountable for shipping at least one AI transformation case study per quarter from their department.
Critically, we staffed our AI automation engineer roles from our existing People team because those folks have existing HR subject matter expertise. This shows the organization something we want to be true: while disruption will occur in the range of jobs we have, we are betting on our current talent.
An HR leader intuitively understands the sharp edges of using the org chart to drive transformation, which is why the People organization's central role in this work matters.
We also raised compensation from the 50th to the 90th percentile — a concept we call "talent density." When AI makes your team materially more productive, share the wealth. Our customer support team is a great example: elite performers producing rare outcomes, paid accordingly.
Why Sharing AI-driven Productivity Gains with Teams Matters
I hope management teams ask themselves: "What should happen to the economic value AI-augmented teams create?" I hope they consider sharing the incremental productivity that teams achieve by thinking creatively and redesigning work through AI.
Our customer support team halved their average handle time. They're now elite performers producing outcomes that are fairly rare in the world of customer support. And they are paid accordingly — top 10th percentile of the market.
This creates a nice alignment. If AI transformation asks more of people — more experimentation, more learning, more reimagining of their own roles — then the upside should flow back to them, not just to the balance sheet. I think the leaders who figure out that alignment will attract and retain the best talent in an AI-augmented world.
Five Ways Leaders Can Navigate AI-driven Transformation Effectively
When you think about AI this way, you stop chasing it for its own sake and start asking: ‘What problem does this solve, and what does ‘better’ look like for the humans involved?’…Without clear boundaries, people will constrain themselves further than necessary. This limits creativity.
Here are five things leaders need to think about when navigating AI transformation (some of which may sound counterintuitive):
- Start with culture, not tools. Do a health check: Do you have a strong sense of psychological safety? Trusted management? A genuine culture of experimentation? Those cultural ingredients matter at least as much as the skills. They will be the wind at your back — or a headwind against you — as you build fluency.
- Think of AI as a tool, not an outcome. This applies to technology in general. Technologies are tools — mediums — ways to get things done. When you think about AI this way, you stop chasing it for its own sake and start asking: "What problem does this solve, and what does 'better' look like for the humans involved?"
- Publish your guardrails early. Without clear boundaries, people will constrain themselves further than necessary. This limits creativity. The first thing we did after our Code Red was to develop AI use guidelines. This unlocked more experimentation, not less.
- Name a single accountable leader. Not a committee. Not a working group that meets monthly. A single individual responsible for AI transformation, with the power to prioritize and allocate resources. At Zapier, that's me — and clear ownership moved us from scattered experimentation to coordinated transformation.
- Enforce the “rule of three.” If your team does something manually three times, they automate it the fourth time. Or, they defend why it cannot or should not be automated. This creates a cultural norm: automation isn't aspirational—it's expected. It also builds the muscle for noticing repeatable tasks, the first step toward redesigning work.
And one bonus: open-source your playbook. Share what you're learning. Sharing progress candidly accelerates the whole field. For AI transformation, open-sourcing helps us all learn faster, together.
Follow along
You can follow along with Brandon Sammut's work on LinkedIn. And check out Zapier's AI Transformation Framework.
More expert interviews to come on People Managing People!
