Turning novelty into a repeatable capability.

This work explored where GenAI adds real leverage in early concept development: speed, breadth, and visual thinking before translating those learnings into a lightweight framework that teams could adopt immediately.

The challenge

Engineering teams often need to explore form, configuration, and user-facing direction before a design function is availableor before committing resourcing. The challenge was to test whether GenAI could expand the concept space while maintaining quality and decision-making clarity, and to package it as a usable method rather than a one-off demo.

My role

As program lead, I designed and facilitated the experiments, built the workflow, generated and curated the concept sets, and guided the team through critique and iteration. I also documented the process as a transferable framework the team could reuse.

What I delivered

Evidence

• Rapid experiments mapping capability limits and best-use moments

• A library of prompts, parameters, and example outputs for reuse

• A critique approach for selecting “signal” from high-volume output

Direction

• A structured GenAI concept workflow (brief - prompt - generate - curate - iterate)

• Guidelines for quality control, bias avoidance, and consistency

• Integration points for engineering constraints and feasibility checks

Commitment

• A repeatable framework enabling engineers to run concept sprints

• A large concept set (1000+ images) demonstrating breadth and iteration velocity

• Presentation-ready artifacts for internal communication and adoption

Scale

4-day sprint / 1000+ concept images generated

Stakeholders: KTM Technologies engineering team
Pro bono / PhD exploratory work

Credits

GenAI workflow design, facilitation, concept generation, curation, and framework documentation by Thom Luke / Threshold Design in collaboration with KTM Technologies.