Friends,

In an ocean of stories about companies doing shitty things, it’s easy to lose sight of the one’s that are actually going about things in a considered, human way.

This weeks edition we’re talking about Zapier, which to me, are the case study of AI adoption in a way that has been thoughtful and effective.

I’m still a bit in awe of the people who say yes to speaking to me when I randomly slide into their DM’s, but I was incredibly fortunate to speak to VP of Global Talent, Tracy St.Dic on all things towards becoming arguably one of the most AI-native organisations in the world.

I took a lot from this chat, so I hope you do, too.

(and if there’s ever any companies you’re in awe of, let me know, and maybe I’ll slide into their DM’s and see if they’ll yes to a chat, too)

Enjoy this week’s edition ✌️

Matt

PS. If it’s you reading this that is the cool company and have a cool story/experience/initiative to share, don’t be shy. Say hi.

PPS. I haven’t talked about TV and movies in a long time but man things started heating up here. I’m writing this the day before I go see the new Spiderman (yay), although I feel like I’m the only person in the world that hasn’t seen The Odyssey yet (anyone else out there…?). I’ve started watching the new series of Ted Lasso (I missed that positivity so much) and one of my most favourite (and most slept on) shows, Lioness, is back, too. I just finished Cape Fear (meh) and the last episode of House of the Dragon is out next week (eeek). Silo is in the backlog and I’m sure there’s much more cool stuff I’ve missed. What are you watching that I need to put on my list??

In case you’re new here (or just missed it) here’s the past three editions of the FNDN Series:

Before AI, HR wrote requirements while vendors or internal engineering teams built the software.

That era is ending, because HR teams can build now.

Here's what I’m seeing:
→ A 4,000 person company built its own performance review tool, org chart, and internal talent marketplace.

→ A 1,500 person company on Workday priced an LMS, then vibe-coded one themselves instead.

Both had working versions in weeks. Engineers later helped productionise and harden them.

Thinking of vibe-coding yourself? Where to start.


You probably don’t build payroll or a HRIS first 😊 Build the employee experience layer instead. Career frameworks, recognition, pulse checks, 1:1 prep, performance reviews, onboarding, manager coaching.

I've built a few examples you can reference, adapt, or simply use for free.

Thank you for supporting our sponsors, who keep this publication free

Know a startup Head of People looking for answers 🙋 why not forward this to them for some instant karma? ✨

This edition is based on the latest episode of the FNDN Series podcast, with Tracy St.Dic.

Tracy St. Dic is Global Head of Talent at Zapier, where she oversees talent acquisition and intelligence, branding, internal mobility, and executive talent planning. She also leads Zapier's global staffing strategy and AI fluency initiatives, helping shape how the company builds and prepares its workforce for an AI-native future. Prior to transitioning to tech, Tracy was the SVP of national recruitment for Teach For America, an education non-profit responsible for the largest, most diverse teacher pipeline in the country. She's on the Talent100 Global Power List.

Moderna made Tracey Franklin its Chief People and Digital Technology Officer, handing her the technology department alongside the People one. 

ServiceNow added AI Enablement to Jacqui Canney's CPO title. 

Microsoft moved Kathleen Hogan, its people chief of more than a decade, into the Chief Strategy and Transformation Officer seat, and is now leading organisational and cultural transformation for AI.

And at Zapier, Brandon Sammut went from Chief People Officer to Chief People and AI Transformation Officer.

Four of the most closely watched companies in the world, all handing the AI agenda to the people function.

This makes the CPO role now one of the most closely watched, too.

If your board hasn't asked who should own AI in your business yet, that question is coming. And the answer increasingly determines whether your company actually changes how it works, or just buys more licences.

I sat down with Tracy St.Dic, Zapier's Global Head of Talent, to understand why Zapier made this call earlier than almost anyone, and what the people team did with the mandate once they had it.

Zapier's shift started with a literal code red. 

In March 2023, about six months after ChatGPT arrived, CEO Wade Foster shut the company down for a hackathon and told everyone to get hands on keyboard.

Tracy is open about day one bringing the full range of reactions: scepticism, fear, excitement, curiosity. All the feelings many companies like yours are going through now, and Zapier wasn't immune.

But what happened after the hackathon is the part most companies skip.

AI moved into the company's rituals. 

It showed up in team meetings and enablement sessions, and employee resource groups began hosting their own training so people could learn in a community they felt safe in.

Leaders were expected to build too. 

The CEO and his executive team would come to meetings and demo what they'd made, struggles included. Tracy says watching executives figure it out in public put everyone on an even playing field, and minds at ease.

And the systems backed it up. Zapier ensured that experimentation and ‘citizen development’ were recognised in performance management and promotion decisions.

Tracy is refreshingly unpretentious about the playbook. 

"This is change management 101," she said. Explain the why, set the direction, be explicit about expectations, then match the demand with real support.

But she shares a warning for companies running only half of this playbook. 

She sees plenty of businesses pushing adoption top-down only, or expecting it to bubble up from below on its own. 

Zapier deliberately did both at once, and she credits that combination for getting the company past the "AI is going to take my job" wave early.

The artefact that makes people-team ownership concrete is Zapier's AI fluency rubric, now in version two and shared publicly.

It was born from a hiring problem. 

Candidates were reaching final interviews and getting knocked back for "not being AI fluent enough". It turned out, nobody could define what that meant. 

So the team codified what Zapier's best performers were actually doing with AI and turned that into the standard. Quite simply, they searched for those that were using AI best, asked them what they did, and then built a rubric that reflected that.

The rubric assesses four components: mindset, strategic acumen, builder skills and, new in version two, accountability. The minimum hiring bar, which Zapier calls "capable", now assumes AI is embedded in your daily work.

My favourite detail is the principle they call slope, not snapshot. Zapier measures how fast someone is learning, including across the four to five weeks of the hiring process itself, rather than where they happen to sit today. Because the technology moves too fast for a point-in-time reading to hold.

Then we got to the question every board eventually asks: is any of this actually paying off?

Zapier settled its answer early, and defined ROI on the things that mattered to them rather than the finance line alone. Every AI workflow gets judged on three measures.

A successful workflow ideally hits all three. But Zapier is honest that some only improve quality while doing nothing for speed, and they will still ship them.

Tracy’s equally honest that the measurement is a work in progress. 

They're only starting to weigh token costs against what they get back, and her words were "we're still kind of in this messy middle of figuring it out."

If the company furthest down this road is still refining the maths, waiting for a perfect ROI model is a poor reason to delay. Measure what you care about and start.

But the economics are easy to put in front of your CFO. Zapier's 45-person people team spends roughly one full-time salary per year on AI tokens, and in exchange all 45 work with AI every day. Tracy calls that trade-off obvious. And while they do have a token leaderboard, it’s more a signal of use than an incentivised metric. Internally the phrase is "valuemaxxing" (vs tokenmaxxing), getting the most from what you use rather than using the most.

Underneath everything Tracy described sits one pattern: Zapier ran AI adoption as a people program, with people-team disciplines. Change management, capability frameworks, reward systems, rituals.

That's why this mandate keeps landing on the CPO's desk, at Zapier and everywhere else.

If it lands on yours, start with the rubric, and give the full conversation with Tracy an hour of your week.

If you enjoyed this post or know someone who may find it useful, please share it with them and encourage them to subscribe.

A roundup of the most interesting stuff from the week:

When you’re ready, here’s three ways I can help you:

1. Tools & resources
Resources and tools that give you what you need to build your own startup compensation practices.

2. Comp consulting
Building startup compensation practices that are clear, fair and competitive.

3. Startup People Summit
A 1-day annual event for People professionals in scaling companies. Creating the playbook for startup people practices. Grab recordings from past events, or subscribe to join the next summit.