Context
Unifize is an enterprise SaaS company building collaborative quality and compliance tools. When I joined as Product Manager for Core Platforms, two products were on the roadmap but had no shipped features: a Document Management System (DMS) and a Quality Management System (QMS).
My mandate: take both products from 0 to 1. Ship them. Generate revenue. Do it in 9 months.
"Build it from scratch, drive adoption, generate revenue. And by the way - we want AI in it." That was the brief on day one.
The challenge
DMS and QMS are not simple products. They operate in regulated industries (pharma, medical devices, manufacturing) where every feature decision has compliance implications. Users are not consumer-level tech adopters - they're quality engineers, compliance managers, and auditors who tolerate no tolerance for errors.
The specific challenges:
- No existing product to iterate on - genuine 0-to-1
- Regulatory constraints (21 CFR Part 11, ISO 9001, GMP audit trails) shaped every feature
- Multiple stakeholders: sales, legal, QA leads, implementation team, C-suite
- AI features were expected from day one - not v2, not "someday"
- Implementation turnaround time (TAT) was a key sales metric - slow onboarding killed deals
What I owned
- Full product roadmap for DMS and QMS
- Backlog prioritisation and sprint planning
- User research and customer discovery calls
- Cross-functional alignment (engineering, sales, customer success, legal)
- AI feature definition and go-to-market
- Release quality and stakeholder sign-offs
- Implementation process redesign (to hit TAT targets)
How I approached it
Phase 1: Rapid discovery (Weeks 1–4)
Before writing a single PRD, I spent four weeks in discovery. I interviewed 12 customers (current users of adjacent Unifize products), 6 prospects, and 3 implementation engineers. The goal: understand what "good" looks like for quality teams, not what "good" looks like for a product manager who's never used QMS software.
Key insight: customers didn't want more features - they wanted faster onboarding. Every competitor took 3–6 months to implement. If we could cut that to 4–6 weeks, we'd win deals on implementation speed alone before a single feature comparison.
This reframed the entire product strategy. The first major deliverable wasn't a feature - it was an implementation framework.
Phase 2: Core DMS (Months 2–4)
Phase 3: QMS core + AI integration (Months 4–7)
With DMS shipped and in customer hands, I moved to QMS. The QMS scope was larger: CAPA (Corrective and Preventive Actions), deviations, change control, risk management, and audit management.
Simultaneously, the AI roadmap went live. Rather than shipping AI as a standalone feature, every AI use case was embedded directly into existing workflows:
- AI document summarisation - auto-generate 3-line summaries for SOP documents, so quality engineers can triage 50 docs in 5 minutes, not 50 minutes
- AI CAPA root cause suggestion - given a deviation description, the AI suggests probable root causes and references similar past CAPAs
- AI risk scoring - auto-rate risk severity and probability based on historical data patterns
- AI audit preparation assistant - generates an audit readiness checklist based on current document status and open CAPAs
- AI change impact analysis - flags documents, processes, and people affected by a proposed change
Two of these became standalone AI agents (CAPA assistant and audit prep) with conversational interfaces - a user could describe a deviation and get a complete CAPA draft in 3 minutes vs. the industry standard of 3 hours.
Phase 4: TAT reduction and revenue acceleration (Months 7–9)
The discovery insight about implementation speed became actionable here. I redesigned the implementation process with a template library, guided onboarding flows, and pre-built configuration for the 6 most common industry setups (pharma, medical devices, food, automotive, aerospace, general manufacturing).
Result: implementation TAT dropped from an industry-standard 3–6 months to 4–6 weeks - a 60% reduction. Sales used this as their #1 differentiator in late-stage deals.
Results
Stakeholder management: the hidden work
The technical execution was the easier half. The harder half was stakeholder alignment across four groups who wanted different things:
- Sales wanted features that would win the next deal, often at the expense of product coherence
- Engineering wanted design clarity and stable priorities - two things that are inherently scarce in a 0-to-1 build
- Legal/Compliance needed every feature to be defensible in a regulatory audit
- C-suite wanted revenue numbers, now
The only way to manage this was radical transparency about tradeoffs. When sales wanted a feature that would take 6 weeks, I showed the full cost: 6 weeks = these 3 other features slip = implementation for existing customers slows = TAT goes back up = we lose the pricing advantage we just created. Making the tradeoff visible - not just saying no - created alignment instead of friction.
What this demonstrates
This engagement was full-spectrum product work: discovery, strategy, roadmap, sprint execution, AI feature design, stakeholder management, and go-to-market. The $1M revenue isn't a vanity metric - it's the output of getting every layer right.
What I brought to Unifize is what I bring to every engagement: the ability to understand a domain quickly, make decisions with incomplete information, move fast without breaking things that matter, and turn a vague mandate ("build a QMS") into shipped product that generates revenue.