Client: Confidential Global CPG Company
My Role: Lead Product Designer
Duration: 3 months

Designing an AI Innovation Engine for a Global CPG Leader

The Challenge & My Role
AI Innovation Engine: At a Glance
Module 1: Ideation Studio
Module 2: Concept Generator
Module 3: Validation Engine
Outcome & Purpose of MVP
The Challenge

A leading global CPG company needed a faster, more scalable way to explore new product opportunities across categories - not just to stay competitive, but to lead the market. Like many large organizations, progress was slowed by size and bureaucracy. Traditional innovation cycles took months of research synthesis, brainstorming, and manual concept validation, while valuable insights remained siloed across teams. The client partnered with BOI to explore how AI could accelerate decision-making by building an AI-powered innovation engine software to enhance their current innovation process.

My Role

As Lead AI Product Designer, I designed the system architecture, interaction design and helped define scoring frameworks, and reasoning models. The result was an MVP innovation engine that transformed fragmented data into prioritized opportunity spaces, generated structured product concepts, and validated them using transparent scoring - all within a single workflow. By balancing AI capability with trust and usability, the MVP quickly gained executive support and unlocked funding for future development.

AI Innovation Engine: At a Glance

I designed the engine to be built around three connected modules that move ideas from discovery to decision.

MODULE 1

Ideation Studio

Turns data and insights into prioritized opportunity spaces.
MODULE 2

Concept Generator

Produces test-ready concepts with human-guided refinement.
MODULE 3

Validation Engine

Scores concepts desirability with transparent logic.
Tools should fit real workflows: Because innovation teams don’t all start from the same place, the flow needed to support both exploratory discovery and focused investigation, allowing users to begin with a broad category scan or a specific strategic input.
Transparency builds trust: By surfacing the data sources behind each insight, the interface helps users understand the why behind an opportunity space.
Transparency builds trust: Scoring had to be visible so users could quickly scan and understand the prioritization of opportunity spaces. In the detail page, scores are broken down.

Module 1:
Ideation Studio

The Ideation Studio transforms unstructured data (like sales trends, competitive signals, internal research, and social listening) into ranked opportunity spaces using structured reasoning. Once a user selects a path, the engine begins synthesizing this information into clear, data-backed areas for exploration.

I designed the Ideation Studio around two core principles:
‍
1) Transparency builds trust
2) Tools should fit real workflows

Module 2:
Concept Generator

Once opportunity spaces are defined, a user would be able to generate concepts from the selected opportunity space. At this stage, human-in-the-loop collaboration is essential. 

Regeneration
To preserve human expertise, I built regeneration controls that allow users to guide and refine concepts based on their own insight. Instead of accepting AI output as-is, innovation teams could adjust direction, add constraints, or inject strategic nuance, ensuring the system reflected institutional knowledge and brand strategy.

The result was a collaborative creative loop where AI accelerates thinking without replacing it.
Setting the stage for Validation: Unlike random idea generation tools, each concept followed a standardized format: including a working title, description, insight, job to be done, and experience cues. This consistency would be important for comparability and validation later in the process.

Module 3:
Validation Engine

To earn leadership confidence, the system needed not just to generate ideas but to justify them with data.

I designed a Validation Studio where concept scores were explainable, traceable, and strategically grounded. Once refined, concepts could be validated against existing personas, building on familiar frameworks rather than reinventing them.

Each concept received a desirability score with transparent rationale across key criteria: solution fit, differentiation, future relevance, and customer journey alignment. This turned validation into a feedback engine, helping teams see why concepts performed well or poorly and how to improve them.

By combining explainable scoring, strategic personas, and iterative refinement, the module created a continuous improvement loop where ideas evolved toward both user value and business viability.

Outcome & Purpose of MVP

This MVP proved that AI could strengthen human strategy, successfully securing funding for continued development. It cut concept development from weeks to hours, created a structured, repeatable innovation workflow, and increased leadership confidence. Most importantly, it showed that AI isn’t here to replace teams, but to enhance their thinking and accelerate decision-making.

Related case studies

The future of fashion:
Designing an AI autonomous trend-to-product engine

Read Case Study →

Unlocking the value
of digital twins

Read Case Study →

From idea to validation
to launch 🚀

Read Case Study →