Product Builder · 15+ Years · Enterprise to SMB

I find the friction between what users need and what a product gives them, and I close the gap.

Built from scratch. Grown at scale. Now building for the AI era.

I'm Nidhi Shreya. For the past few years I've led AI-powered, AI-native product portfolios and a team of product managers, with one test for AI: does it collapse the distance between what a user wants and what the product delivers? Anything else is decoration. My path into product wasn't a straight line — which is exactly why I'm good at it. The shortest version of me is three Ps: products, people, and poetry.

Read the case studies Get in touch

the gap is usually smaller than it looks, if you know where to press.

Products

Building them is how I think. Sixteen years of platforms, portfolios, and the occasional brave sunset.

People

Understanding them is where every good product decision starts. I keep an empty chair in every meeting - for the user we are building for.

Poetry

Writing it keeps me honest. Both crafts are about cutting everything that doesn't need to be there.

A non-traditional path, on purpose

the detours were the education.

Most recently

Product Advisor to early-stage teams

I advise early-stage teams on product strategy, using first-principles thinking to clarify the actual customer problem, validate product-market fit, explore where AI genuinely helps, and make sharper product and team decisions.

Before that

Simpplr — AI-first Employee Experience Platform

At Simpplr, I led an AI-native Employee Experience platform 0→1 — transforming the static home page into a personalized, dynamic and action oriented digital workspace for millions of users, reducing churn and protecting multi-million-dollar ARR. I owned the agentic AI roadmap — LLM-based writing, summarization, and personalization recommendations — improving content creation velocity by ~20%, cutting content consumption time by ~25–30%, and lifting user engagement by ~15–20%. I also owned the employee Feed portfolio, improving content relevance and personalization while evolving people discoverability and attribute-based access controls across diverse enterprise environments. I drove product strategy and executive demos for Gartner and Forrester analyst evaluations, earning a place in the Magic Quadrant while strengthening Simpplr's market positioning, spearheaded the org-wide ABAC (Attribute-Based Access Control) initiative, and operated as a player-coach to a team of PMs across multiple product lines.

The build years

Senior product roles — Employ/Jobvite, Digital Harbor, Ellucian

Enterprise platforms, HR tech, healthcare, and marketplaces. At Employ/Jobvite: built an AI-driven onboarding platform 0→1 and scaled it to 1,000+ customers, grew offer and requisition adoption from ~20% to 45%+, and launched a marketplace of 30+ integrations serving 10,000+ B2B customers while owning P&L. At Digital Harbor: cut healthcare provider-enrollment processing from 180 to 60 days. Along the way, I learned to sunset products as thoughtfully as I launch them, and to make build-vs-partner calls with data instead of ego.

Where it began

Business analysis → founding PM in India

After early years that included a stint teaching marketing and branding to MBA students, I moved into business analysis and then became the first product management hire in India at a US-headquartered company — at a time when "PM" barely existed as a role here. I had to define the job before I could do it. By then I'd already lived in 12 different places across India and abroad for education and work — being the new person in an unfamiliar room was a habit long before it became a professional skill.

Case studies

real problems, real trade-offs — no vanity metrics.

Agentic AIHuman-in-the-loop

Turning content gaps into a supply problem an agent could solve

Employees were searching for things on the intranet that didn't exist yet. Say someone looks for the international travel policy, doesn't find it, and closes the tab. That shows up as a single zero-result search, easy to miss, and hard to know when it's worth acting on.

What problem was hard?

Zero-result searches were invisible one by one. Nobody owned the gap between what employees searched for and what content actually existed.

Why did it matter for both sides?

For employees, not finding an answer meant raising a support ticket or just giving up. For the business, unanswered searches quietly build distrust in the platform, and that eventually shows up as churn.

What trade-offs did I make?

I went back to first principles: the real question wasn't "can AI write content," it was "should this exist and can we trust it." We built an agent that scans zero-result searches, filters one-off queries from repeated ones, and pulls together what's already out there in Google Docs, Confluence, or SharePoint into one clear, updated answer. Before anything went live, we set an accuracy bar for the draft, and if employees kept ignoring or rejecting a piece of content, we treated that as a real signal and pulled it back for review. The agent never published directly. A person approved it first. Slower than full automation, but we didn't publish anything wrong while the system was still earning trust.

AI Personalization0→1Evals

AI-powered landing page experience (personalization)

Every employee saw the same homepage no matter what their role was or what they needed that day. A new hire in sales and a ten-year engineer logged in and saw the exact same links and announcements. Making the page different for each person wasn't the hard part. Knowing whether it actually helped them, or just looked different, was.

What problem was hard?

Personalizing content is easy to fake and hard to prove. The bar wasn't "different," it was "actually more relevant."

Why did it matter for both sides?

If the homepage isn't useful, employees stop opening it. For the business, a drop in usage is usually the first sign a customer might not renew.

What trade-offs did I make?

The tempting move was rolling personalization out to everyone at once, since that gets adoption numbers up fast. I didn't. I held it back, tested it in stages, and set a minimum accuracy bar it had to clear first. Getting a personalized homepage wrong doesn't just fail to help someone, it damages the one thing that made the page worth opening. So I traded a faster rollout and better-looking early numbers for something slower but real. By the time we scaled it, we knew it worked, not just that it shipped.

PlatformCross-functional

Leading an access-control initiative across the whole product line

Enterprise customers needed access control far more granular than simple roles. Two engineers can hold the exact same job title, but one is based in Germany and the other in India, and depending on the customer's policy, they may need to see completely different things, based on role, location, seniority, business unit, and other attributes layered together.

What problem was hard?

Attribute-Based Access Control (ABAC) touched five or six product lines, and none of them had it on their roadmap or felt like it was their job to own.

Why did it matter for both sides?

For customers, this was often a dealbreaker. Without attribute-level control, their own security or procurement teams wouldn't sign off on the purchase. For us, meaningful revenue sat stuck behind a piece of work nobody wanted to lead.

What trade-offs did I make?

First principles again: start with where the money actually is, not with what's easiest to build. We sequenced the rollout by which product lines had the most revenue sitting behind this gap, not by which team could move fastest. Some teams waited longer than they wanted to. But it turned a piece of work nobody wanted to own into something everyone could see the value of.

StrategySunsetting

Sunsetting a legacy module — and knowing when not to build

We had a module with real customers still using it and real revenue coming from it, but it no longer made sense to keep building on it ourselves. A customer in Canada couldn't onboard new hires because our module didn't meet a local regulatory requirement, and building that support in-house would have taken months.

What problem was hard?

It's a harder call to end something people are using than to launch something new. Sentiment said keep it. The data said otherwise.

Why did it matter for both sides?

Customers needed a clear plan so they weren't left stuck mid-way. For us, continuing to support something we shouldn't keep investing in was costing engineering time and slowing down compliance work customers actually needed.

What trade-offs did I make?

I validated the build-versus-partner decision against real cost and speed data, not a gut call. We partnered with a third-party compliance vendor instead of building the fix ourselves, which meant giving up some control over that part of the customer experience. The right call for speed and for actually closing the compliance gap, unlocking 30–40% of enterprise customers blocked on US and Canadian regulatory requirements, but it meant trusting someone outside the company with something we used to own directly.

Healthcare ComplianceWorkflow AutomationEnterprise SaaS

Reduced Provider Enrollment Cycle Time by 50%, Accelerating Access to Qualified Healthcare Professionals

At Digital Harbor, I worked on Know Your Provider (KYP), a platform U.S. state agencies use to verify healthcare providers before they can practice. Enrollment could take up to 180 days. Say a nurse applies to work in California, and her case sits in a queue because the supervisor assigning cases has no easy way to know which reviewer has the right expertise and free time.

What problem was hard?

It wasn't one single bottleneck. Cases were assigned manually, reviewers lost track of where they left off, and managers had to read long reports before approving anything.

Why did it matter for both sides?

For healthcare providers, every extra day delayed starting work, worse during COVID when states urgently needed more people. For the state agency, a slower process also meant slower fraud checks, since speed and compliance usually pull against each other.

What trade-offs did I make?

We automated case assignment and case summaries, but kept the final approval decision with the manager. A deliberate line, not a technical limit. Automating that last step too would have been faster, but in healthcare compliance, keeping a person as the last check was worth the extra time. Cycle time dropped 50%, from 180 to 90 days, within the first four months.

"I've stopped trying to sound like a product leader. It turns out being a person works better."

— a note to myself, kept

How I lead

strategy is a team sport; clarity is the captain.

01Develop PMs, not just products

When I lead product managers, my bar for success is theirs: I give away the interesting problems, coach on judgment rather than answers, and measure myself by how rarely they need me in the room.

02Portfolio over pet projects

I make calls at the portfolio level — what to fund, what to sequence, and what to sunset. Some of my proudest decisions are things we deliberately didn't build.

03Honesty as an operating system

I'd rather surface an uncomfortable number early than defend a comfortable story late. Teams do their best work when the truth is cheap to say out loud — so I make it cheap.

04Influence without the org chart

Platform work taught me to align five or six product lines that don't report to me. Leadership, to me, is getting to a shared "yes" that survives contact with everyone's roadmap.

Beyond the roadmap

the parts of me that don't fit in a resume bullet.

✦Mentor on ADPList

Mentor on ADPList, coaching aspiring and current product professionals through individual and group sessions.

✦Published poet

My poems appear in Rang Zindagi Ke, a published anthology bringing together poets' voices on life's many colours. Writing poetry keeps my product writing honest — both crafts are about cutting everything that doesn't need to be there.

✦Lean In Circle lead

I lead a global Lean In Circle, creating a space where women navigate careers, confidence, and the messy honest middle of both.

✦Women in Product community

Active in the Women in Product community, with a focus on making non-traditional paths into product visible — because mine was one.

✦The library project

I started and run a community library in my apartment complex. Smallest product I've ever launched; possibly the highest CSAT, engagement and retention I ever pulled off.

Writing

musings on products, people, and everything in between.

Interview · On the Dias · Featured

The Project Underneath the Project

Before a single line of code was written, we spent time going back through a full year of customer usage data, support tickets, and feedback. We walked in prepared, not guessing. On day one we spoke to five key customers and watched how they actually moved through the product, where they got stuck, what they considered non-negotiable.

By day three and four, we went back to those same customers and showed them what we'd built. Not to impress them. To pressure test it before we committed. Nobody called that sprint a project. It had no launch, no announcement. But without it, everything that followed would have been built on assumption.

Read the full interview ↗

Substack

Credibility Doesn't Travel. Until You Make It. ↗

My dad spent 35 years in the same nationalised bank, and his reputation reached the new location before he did. I've switched jobs multiple times, and every time, the work stayed behind. On why credibility doesn't travel unless you build it to.