AI Product Build SaaS Behavioural Science MLP

From Behavioural Science
Thesis to Buyable SaaS Product

We bridge the gap between your AI thesis and a product customers can buy.

AskPhi platform — AI-powered audience intelligence
3 months prototype → production
2 active revenue streams
Full MLP with 4 core workflows
Live with paying clients
Engagement Snapshot
ClientAskPhi (Predictive Human Insights Ltd)
FounderJason Nguyen, Co-Founder & CEO
SectorAI-powered audience intelligence
Engagement3 months
Service lineAI Product Build
OutcomePrototype → live SaaS product

Key Highlights

3 months — Prototype to production-grade MLP
Full MLP — Personas, focus groups, concept testing, querying
2 revenue streams — Done-For-You sprint + self-serve platform
Live — Early clients inc. Run It Once, Teleya, Slippery Saints
Context & Challenge

Jason Nguyen spent 11 years as a professional poker player — making decisions with incomplete information and reading human behaviour under pressure. He formalised that intuition through behavioural science and economics, building Bloom Story, his behavioural marketing consultancy.

The thesis was sharp: most companies profile their audience by demographics. The best companies understand why people buy, leave, and never considered them in the first place. Demographic data tells you who someone is. Behavioural intelligence tells you what they'll do next.

Jason had built a working prototype using Claude Code to automate and productise that thesis — an AI-powered audience intelligence tool. The prototype proved the concept. But it had the problems all vibe-coded builds have: no production-grade architecture, brittle in edge cases, no multi-user support or proper data layer, no purchasable product structure, and not something a paying customer could use independently.

Jason needed a technical partner who could look at what he'd built, understand what it was trying to do, and make it real — without stripping out the thinking that made it interesting.

Approach & Solution

Understanding the idea before
touching the code.

Hypothesis
Stage 01 — Hypothesis

Shape the vision

Ferrous came in at the point where the idea was proven but the product didn't yet exist. The first task was understanding the prototype — not just the code, but the behavioural science logic underneath it. What was Jason actually doing when he ran an audience analysis? What decisions did the tool need to support? What would make a marketer or insights lead reach for this over traditional research?

From that, we identified the core workflows that needed to be real, reliable, and sellable: Persona Generation, Virtual Focus Groups, Concept Testing, and Audience Querying. We also helped shape the two-mode product structure: a Done-For-You sprint service and a self-serve platform.

Experiment
Stage 02 — Experiment

Validate with real buyers

Ferrous built around and on top of what Jason had created — preserving the behavioural logic while engineering it properly. We built a minimum viable version and tested it with Jason's existing consulting clients to validate the concept with real users and real data.

Formulation
Stage 03 — Formulation

Build the minimum loveable product

Using the experimental data, we built a production-grade MLP:

A production-grade application architecture
Multi-user platform with proper auth and data separation
Full AI pipeline — persona generation, synthetic interview engine, concept testing, audience querying
A purchasable product structure with bookable demo flow
Execution
Stage 04 — Execution

Launch and grow

The MLP launched as AskPhi with a live website, demo booking, and the Done-For-You sprint already in market. Early sprints with clients including Run It Once, Teleya, and Slippery Saints validated both the product and the methodology.

Results & Impact

What changed.

Before Ferrous Labs After Ferrous Labs
Claude Code prototype — unpresentable to customers Production-grade MLP — live and bookable
No way to sell or scale the methodology Two active revenue streams: sprint + self-serve platform
Jason required in every engagement Platform operates independently of founder time
Concept with no commercial shape Visitors can book a demo directly from the website
Jason Nguyen
I had the thesis, I had the prototype, and I had no idea how to turn it into something a customer could actually buy. Ferrous took what I'd built, understood why it worked, and made it real. Three months later we had a product.
Jason Nguyen, Co-Founder & CEO, AskPhi
What's Next

Still going.

AskPhi is live, taking sprint clients, and building out its self-serve platform. The behavioural intelligence that Jason spent over a decade developing is now a scalable product — not a consultancy deliverable, but a platform.

Got a prototype that's not ready to sell?

We bridge the gap.

If you've built something with Claude Code, Cursor, or Replit that proves your concept but can't face a paying customer yet — we bridge the gap. Book a discovery call and let's see what you've got.