An AI-forward startup with 200 team members had just brought on their first Chief People Officer. And one of the first things she quickly realized is their compensation program needed to be fixed.

She knew what was coming — more hiring, more complexity, more scrutiny from employees who know exactly what they're worth in this market — and she wanted to get ahead of it.

So they came to us. What followed was one of those engagements where everything just clicked.

What They Were Walking In With

Their pay program at that point, was held together with a mix of free salary data, old benchmarks projected forward with a finger in the air, and candidates' salary expectations used as data points. Many startups move fast — you make it work and figure it out as you go. That's not a criticism. That's just how it goes.

The cracks were already there.  They had pay bands, but there wasn’t really a clear compensation philosophy behind them  — not even a sense of what market percentile they were roughly targeting.

Still, things didn’t feel quite right.

Roles that should have been valued similarly were showing up with surprisingly wide pay differences. There was also growing uncertainty about whether their sales comp was actually strong enough to compete with other fast moving AI startups hiring similar talent skillsets.  

And the foundation underneath all of it wasn’t good. The pay bands had been built from a compensation survey someone downloaded three years ago and kept alive in a spreadsheet that, as it turned out, had a few formula issues of its own.

They had a job architecture, but only within Engineering. And even there, the Staff Engineer title was handed out because a manager was scared to lose someone — because titles were free. They created three sublevels for every real market level, which was an unnecessarily complicated leveling framework. Two engineers doing essentially the same work with wildly different pay — one had negotiated salary harder, so they got lower equity. When too much flexibility is given with offers, perceived unfairness spreads fast when people eventually compare notes.

None of it was intentional. Nobody sat down and designed it this way. It's just what happens when nobody steps back to build a thoughtful framework — and the new CPO felt it immediately. Not just from her own observations, but from employees raising it at all-hands, managers venting, and even executives flagging that something was off with pay.

It’s the kind of setup that works for a while, until you’re trying to compete for some of the hardest-to-find talent in the market. Then, the gaps start showing up in hiring and pay decisions.

What We Built Together

First: A compensation philosophy they could actually articulate — for both cash and equity.

This sounds abstract, but it's one of the most valuable things a People leader can do early. When someone asks "why am I paid what I'm paid" or "how do we think about equity versus cash," the answer shouldn't be "it depends." It should be a clear, documented point of view that everyone from the Board and CEO to a hiring manager can reference and stand behind.

We worked with their CPO to define that philosophy — how they think about market positioning, what role cash plays versus equity, how they differentiate by role type and level. It became the backbone for everything that followed.

Second: A job architecture that actually aligned with their business strategy — no faux levels

We built out a full job architecture — job families, levels, and the criteria that define each one. Not a generic framework borrowed from Levels.fyi (no offense), which reflects Big Tech realities and doesn’t fit a 200-person startup’s unique growth path. Instead, we designed it around how this company actually operates and what realistic career progression looks like for their employees.

When levels are clear and consistent, everything downstream gets easier: compensation decisions, offer conversations, promotion discussions. And managers can actually explain a pay decision without dreading the follow-up questions — which is good for everyone's blood pressure.

Third: Pay bands built on real-time data and actual role value 

Out with the three-year-old survey. In with validated, methodology-backed data cut specifically for their talent market. AI-forward companies competing for AI engineers need to know what AI engineers are actually getting paid — not what generic software engineers make, and not national averages applied across the board. We used location-specific data too — Bengaluru market data for their team members based there, not a guessed percentage discount off US rates.

Fourth: An equity program with a proven structure behind it.

This is one we're especially proud of.

Equity is table stakes in AI right now. Anyone who's tried to hire a principal AI engineer in the last two years knows cash alone isn't closing those offers. But throwing equity at people reactively without a thoughtful program behind it is expensive and inconsistent — and employees notice when they compare notes.

We designed an equity program built on our proprietary Equity Model and market equity data — actual benchmarks for equity grants by level, job category, and location. We helped them think through the full picture: new hire grant philosophy, refresh cadence, eligibility criteria, how equity scales with performance, and how to use it as a real retention lever. We even helped them build the case to their board for additional option pool shares — and yes, we have a proven deck for that conversation.

The result was an equity program that meshed with their overall comp philosophy — one that makes recruiters look sharp when walking a candidate through an offer, because they can explain exactly where the numbers came from and why.

What They Came Away With

A compensation philosophy their CPO could walk their board through — which landed so well the board asked their other portfolio companies to do the same thing. A framework managers could use to have real conversations with employees about pay decisions. A job architecture that gave everyone a shared language for levels and career growth — and with that clarity came something unexpected: employees felt more motivated, because they could actually see a path forward within the company.

Pay bands built on real-time, defensible data. And an equity program designed to actually compete for the talent they needed to hit their business goals.

Their CPO told us it was the first time she'd walked into a leadership meeting to present their comp strategy and felt confidently prepared. Honestly, that meant everything to us. Our job is to help People leaders be the hero in that room — and moments like that are exactly why we do this work.

Takeaways

If you're a People leader at a startup and any of this sounds familiar, a few things worth thinking about:

• Your first comp philosophy conversation is the most important one. Before you touch a single pay band, get alignment from leadership on the foundational questions. How do you want to position against the market? What's the role of equity versus cash? What does "competitive" actually mean here? Everything downstream gets easier once those questions have real answers.

• Free data has a shelf life. It might get you through your first 50 hires. By 200 people, it's a liability. The cost of getting comp wrong — losing a key hire, overpaying in certain countries while underpaying in others, an employee who feels misled is always higher than the cost of getting real data.

• Job architecture isn't bureaucracy. It feels like a big-company thing until you're trying to make a promotion decision and realize nobody can agree on what the next level actually requires. Build the framework before you need it to resolve a conflict.

• Equity needs a program, not just a budget. In AI and ML, equity is one of your most powerful retention tools — but only if it's deployed consistently and explained clearly. The same equity budget goes so much further when there's a thoughtful program behind it versus just handing out grants whenever someone pushes back on an offer.

The best time to do this was six months ago. The second best time is now. Comp debt compounds. The longer inconsistencies sit, the harder they are to unwind — and the more noise they create as your team grows and people start comparing notes.

Growing fast is exciting. But at some point, the things you hacked together to get here will become the things holding you back.

This company had the self-awareness to recognize that moment — and a People leader with the conviction to act before it became a hot mess fire drill.

What they built isn't just a comp program. It's the infrastructure for the next stage of growth.