← Back

From 0→1:
Defining, Designing,

and Building the MVP

Turning a complex, multi-stakeholder, multi-stage process into a centralized, intelligent, and efficient operating system.

My Role

Product Designer

led the UX/UI, information architecture, flows, interaction design, and high-fidelity prototyping.

Team

Product Lead, Product Manager, UX/UI Designer, Lead Developer,three software engineers, and four representatives from the client organization.

THE PROBLEM

A complex, multi-team process was run across WhatsApp groups, email threads and a shared Excel file, with no defined stages and no way to see the bigger picture.

WhatsApp groupsShared Excel filesEmail threads

THE GOAL

Turn a complex investigation process into one clear, centralized workflow for managing, tracking, and moving each investigation forward from start to finish.

Background

What We Started With...

A small unit reporting directly to the senior leadership, pulling OSINT, intelligence and operational material into an investigation, one factual account of what happened.That investigation had to hold up under external legal scrutiny.

No system to redesign. Just people, handing work to each other. So I mapped the handoffs ↓

The RESEARCH challenge

Limited access

We couldn't see the data or the work we were designing for.

Many roles, many stages

Sub-stages ran in parallel — the process rarely followed one path.

The central user was leaving

Everything ran through one key person who was about to leave.

Previous tools hadn't held

Earlier solutions failed to fit the way people actually worked.

They speak their own language

Each stage had its own terms and shorthand.

It had to scale to other teams later

The model would need to become modular for additional teams.

The RESEARCH Process

Simplifying complexity in a multi-stakeholder operational environment.

Research was an important part of this project, so we approached it from several angles: one-on-one stakeholder sessions, a full-day kick-off with every role in the process, a working session that walked one complete investigation end to end, review of completed assessments and existing tools. We mapped the core personas around their goals, needs and permissions. tested 2–3 UI directions with users on language, sequence, information load and edge cases.

Key insights

01

Nobody could say where anything stood

Bottlenecks stayed invisible until someone chased them down.

Time-stamped statuses, with alerts when something stalls.

02

An event and an investigation are not the same

Same record, entirely different actions — one is triaged, the other is worked.

Two dedicated pages, each with its own table and actions.

03

The process isn't linear

Stages run in parallel at different speeds.

Stages and tasks that stay editable, tracked by a non-linear progress bar.

04

It stalled at the handoffs

Information passed between roles through disconnected tools and files.

One shared workspace that keeps the investigation moving.

05

Previous tools had eroded trust

Past solutions didn't fit the real workflow and often worked in isolation.

The MVP had to fit existing work first — and connect what was already there.

06

Existing tools had to come before advanced features

Users first needed the tools they already relied on to work properly in one place.

Cover the essential workflow first — then introduce advanced features, automation and AI.

Two roles, two perspectives

One process seen from two distances — one tracks the whole organization, the other drives a single investigation from start to finish.

Operations

Operations

Managing the organizational picture

Responsible for prioritization, team assignment, and tracking investigations across the organization.

Workload

Oversight of thousands of events in parallel.

Permissions

Unrestricted: sees every investigation across the organization.

Needs to know

What needs attention now? · Where is work stuck? · Who owns what?

MVP response

A high-level view of status, time, teams, and exceptions.

Investigation Lead

Investigation Lead

Managing the individual investigations

Responsible for gathering materials, managing tasks, writing, and approvals through final delivery.

Workload

Managing 5–10 investigations in parallel.

Permissions

Restricted — only the investigations and material assigned to them.

Needs to know

What information already exists? · What is still missing? · What's the next step?

MVP response

One focused workspace for materials, tasks, progress, and approvals.

What the MVP covered

In the MVP
Events & investigations page — filtering, sorting, status & prioritization
Central investigation workspace — materials, roles & tasks
Investigation writing within the system
Role-based permissions
Next phase
Full permissions management & team assignment
Management dashboards & analytics
Query system, version control & archive
AI for prioritization, duplicate detection & writing
Modular workflow engine for additional teams

Design

Events Table

Thousands of triaged events in one view, split into all events and active investigations. Clickable dashboards, filters, and a status chip with days elapsed make bottlenecks visible. Each row expands to quickly view and edit its data.

Investigation Workspace

A single investigation with details and a progress bar that expands into non-linear tasks. Each task opens with its own tailored features. Tabs hold tasks and documents, event details, queries, and the approval round.

Report Editor

A structured editor that opens with the event's metadata and the report's fixed chapter list. Rich text, autosave, and export to PDF or to the knowledge base.

01
Design the process before designing the interface
We didn't treat the existing workflow as something to simply digitize. We mapped it end to end, identified duplication, dependencies, and friction points, and only then decided how the work should happen inside the product.
02
Understand the user's deeper need
Instead of taking user requests at face value, we looked for the goal behind them — what they were really trying to achieve, what was getting in their way, and what solution would address that need most simply and precisely.
03
A good system should work with the user — and earn their trust
After previous tools failed to fit the workflow, it was important that the new system connect to existing processes and tools rather than add another layer of work. Early collaboration with development helped ensure the solution was both feasible and trustworthy.

Let's take it further