Lightcast · Senior PM · 2025
The AI Product We Shelved — and the Architecture It Left Behind
I was one of three PMs on an AI workforce-planning product for Fortune 500 executives. When the market shifted, I helped surface the signal to sunset it. Months later, the architecture we'd built became the foundation for the next product I led.
TL;DR
We invested upfront in scalable architecture instead of shipping a brittle MVP. By the time the prototype was ready, the macro had flipped — the Great Resignation had turned into a layoff cycle — and our core use case had no buyers. I helped surface that signal and bring the team to a sunset call. Months later, the architecture came forward as the foundation for the next product I led.
1 of 3
PMs
Led discovery, strategy, prioritization, frameworks & AI feature design
Fortune 500
CHROs & business heads
as the target buyer
40%
Roadmap acceleration
on the next product the architecture carried forward into
01
What I Owned
Three PMs split the work. My scope:
- Customer discovery — Fortune 500 executive interviews and synthesis.
- Product frameworks — buyer journey, the "one-screen" answer model, decision flows.
- Prioritization — what made the MVP and what didn't, based on customer evidence.
- Strategy — how Data Stories fit into the Lightcast portfolio.
- New AI feature design (AI-leveraged Insights) — the AI layer that translated dense workforce data into executive-readable answers.
02
The Opportunity, and the Architecture Call
Lightcast's flagship Analyst was the gold standard for researchers — but the executives buying it weren't using it. They wanted answers, not search-and-filter.
That was the gap Data Stories targeted, positioned as "Strategic workforce planning at the speed of business." Inside that umbrella, the MVP was opinionated about one use case: competitive hiring velocity.
We had a choice. Ship a thin executive UI on the existing data stack in 3 months — brittle, every report hard-coded, wouldn't scale past the second customer. Or invest 6+ extra months building two foundations: a redesigned data layer for executive rollups (the legacy star schema was built for granular research queries, not high-level summaries), and a configurable smart object engine that could generate reports dynamically against each customer's specific logic.
The PM team aligned on the foundational path. I was a strong voice for it. Engineering agreed. We built foundation-first.
03
The Market Shifted Under Us
By the time the prototype was tested and ready, the macro had reversed. The CHROs we'd been interviewing weren't asking "how do I hire faster than my competitor?" anymore. They were asking "how do I keep the people I have without overspending?"
The broad workforce-planning positioning still applied. The specific question our MVP was built around — competitive hiring velocity — wasn't being asked.
04
How I Started Raising the Concern
Most PMs default to looking for a pivot when the market shifts. I was one of the first on the team to start asking the harder question out loud — should we still be doing this? — and did the work to back it up:
Re-ran customer discovery.
I went back to the Fortune 500 leaders who had shown buying intent six months earlier. Their priorities had moved from hiring velocity to retention. The Data Stories we'd built wouldn't make their budget cut in the new climate.
Worked closely with Sales to confirm customer intent.
Pipeline conversations and account team feedback confirmed the same signal independently — the pre-shift demand wasn't converting to deals.
Pushed the framing question into the room.
"If we were starting today, knowing what we know now, would we build this product?" Once that question was honestly on the table, the answer became hard to avoid.
The conversation took weeks across PMs, engineering, GTM, and leadership. The signal kept getting stronger. Over a few months the team converged on the sunset call together, and leadership made it formal.
"I wasn't the decision-maker. I was the early voice. At Staff level, that's most of the job. Influence, not authority."
What we kept
A product that didn't ship still left durable value.
We didn't run a salvage operation the day the product was shelved. Months later, when I was leading a separate work intelligence product, the architecture from Data Stories came forward naturally as its foundation — the configurable smart object engine for dynamic reporting, the data layer optimizations for high-level queries, and the discovery learnings that shaped the new positioning. The new prototype's roadmap accelerated by ~40% because of that head start. Not because we scrambled to salvage it, but because we'd built foundations general enough to come forward into the next thing.
05
Lessons
01
The hardest Staff-level skill is raising the signal early.
Shipping a product into a hostile market because you've already spent the budget isn't strategy. It's sunk-cost. Surfacing the truth before sunk-cost takes over is the skill.
02
Staff PMs work through influence, not authority.
Decisions at this scope get made by bringing the right data into the room, mapping honest options, and keeping the right question alive — not by announcing calls.
03
Foundation-first investments compound at the edges.
Good architecture isn't valuable because you can salvage it. It's valuable because it makes the next thing easier.
Next case study →
Randstad RiseSmart: Identity at 110-country scale
Want to talk?
Open to Staff PM, Head of Product, and Fractional CPO conversations.