BUILD AND SCALE A NEXT-GENERATION WORKFORCE MANAGEMENT PLATFORM.

NEWCROSS HEALTHCARE - 2025 - 4 MONTHS

Product Manager on a 0→1 neutral vendor workforce platform built in 4 months using AI-driven development

Context

Operra is a neutral vendor workforce management platform designed to help healthcare organisations manage multiple staffing suppliers through a single system.

I took ownership of the product mid-build, inheriting a rapidly developed but fragmented platform created using early AI coding tools. While a large number of features had been built quickly, the product lacked consistency, clear structure, and alignment across its different parts.

A key complexity was that Operra is not a single application, but a system of three interconnected apps serving clients, agencies, and internal teams. The challenge was to turn this into a coherent, scalable platform — ensuring that workflows, business logic, and features worked seamlessly end-to-end across all users, under tight timelines and evolving requirements.

What is it?

Operra is a B2B platform that centralises how healthcare organisations manage their external workforce supply.

It is built as three interconnected applications:

a client app, where hospitals create and manage bookings, define requirements, and control how shifts are distributed
an agency app, where staffing providers manage their workforce, respond to bookings, and submit candidates
an admin app, used internally to configure the system, monitor performance, and manage the overall ecosystem

These applications are connected through shared workflows, including:

the full booking lifecycle across all parties
timesheet validation and sign-off between clients and agencies
compliance and dispute handling processes
multi-level organisational structures (regions, branches, client systems)

Together, they form a complete workforce management system, replacing fragmented coordination with a structured, end-to-end digital platform.

Visit website

0 to 1

in 4 months

x150

Features built

80%

AI Coded

x3

web-app built

What I did

Rebuilt product foundations from a fragmented prototype

Took over a platform initially built quickly using Replit (AI-generated code)
Identified major inconsistencies:
- missing backend logic
- misaligned frontend / business rules
- unclear feature purpose
Audited and mapped ~100 features across:
- frontend / backend / UX / logic / business requirements
Re-defined the MVP:
- removed unnecessary features
- clarified product scope and priorities
Ensured that all features were implemented consistently across the three applications (client, agency, admin), not just in isolation.

Result

Transformed unstable prototype into a coherent product
Reduced complexity and improved product clarity
Established a scalable foundation for delivery

Introduced structured, high-speed delivery model

Set up daily “guerrilla-style” working sessions across:
- engineering, design, finance, customer success, leadership
Created a fast decision-making loop to unblock progress
Balanced speed with validation:
- ensured business rules were reviewed and agreed
Introduced testing and iteration cycles

Result

Maintained rapid delivery pace without losing alignment
Enabled continuous progress under high uncertainty
Improved overall product quality despite tight deadlines

Leveraged AI tools while reinforcing product fundamentals

Transitioned development from Replit to a more controlled setup (Claude Code)
Used AI to accelerate development, not replace product thinking
Ensured every feature was:
- tied to a real user problem
- validated across stakeholders
Reintroduced strong product discipline (UX, logic, business alignment)

Result

Achieved high development speed while maintaining product integrity
Avoided common pitfalls of AI-generated products (lack of structure, inconsistency)
Built a platform that could scale beyond prototype stage

Delivered a full platform in record time

Led a cross-functional team:
- 2 developers
- 2 designers
Worked directly with clients for rapid feedback loops
Delivered complex functionality including:
- booking flows
- supplier management
- invoicing systems

Result

Delivered full VMS platform in ~4 months
Successfully met aggressive business timeline
Created a competitive product offering for enterprise clients

Key Impact

Delivered a 0→1 workforce platform under extreme time pressure
Transformed an unstable AI-generated prototype into a scalable product
Introduced new ways of working combining AI speed with product discipline
Enabled Newcross to offer a new category of workforce solution

Rethinking product development in an AI-driven environment

AI development significantly increased delivery speed — shifting us from shipping tickets to shipping full features, often in a single day.
We adapted by:
- working at the feature level instead of ticket level
- ensuring daily cross-functional validation (design, product, business)
- introducing fast alignment loops to maintain quality