Live
Persona-Based Usability Testing: testing every journey against the right persona, automatically
AI testing agent · Connects to a persona repository · 4 core UX metrics
Task success% of persona-driven tasks completed without assistance
Time on taskDuration to complete each core task
Error rateMissteps or recoveries per task
ConversionCompletion of the full end-to-end journey
Challenge. Usability testing usually stalls on setup, not analysis: picking the right persona for a journey, writing a task script, recruiting or simulating a session — and only then finding out whether the flow works. By the time results come back, the design has often moved on.
Action. An AI agent reads from the validated persona repository design and research already use, identifies which persona a journey is actually designed for, runs a representative task set in that persona's context, and compiles a structured readout the team can act on immediately.
Result. Testing stops being a scheduled, occasional event gated by recruiting. It becomes a continuous check available the moment a flow is ready — and it reflects the audience the journey was designed for, not whichever persona was convenient to recruit.
Live
Heuristic Scoring Tool: turning heuristic evaluation into a repeatable, comparable score
Evaluation framework · HFI NPCI (VIMM) · 20 checks · Score out of 100
Tier 1 — Automated (~30%)Linters, contrast and accessibility audits score directly
Tier 2 — AI-assisted (~45%)A vision-language model drafts ratings; a designer validates every one
Tier 3 — Human (~25%)Checks needing task context are scored by a designer
Always reviewedAutomated scores are drafts until a human accepts or overrides them
Challenge. Heuristic evaluation is one of the oldest tools in UX, but it usually produces an inconsistent one-off audit — different evaluators, different criteria, and no way to compare a score to last quarter's.
Action. Built on Human Factors International's VIMM model, the framework scores four focal points — Navigation, Presentation, Content and Interaction — with 5 checks each, rated 1–5. The subtotals sum to a clean composite out of 100, and every check rated 1 or 2 automatically generates a prioritised issue with evidence attached.
Result. The same 20 checks and the same scale, every time — a score that can be tracked release over release, and never just a number without a reason.
In development · Internal alpha
Craftline: making design-system governance an operating system, not a vibe check
Governance platform · Integrates with Figma
Figma file auditA rule engine rates violations Critical, High, Medium or Low
Broken component detectionDetached instances, deprecated sources, local overrides
Component lifecycleIntake → Design Authority → Engineering → Release
Exception managementEvery exception has an owner, evidence and an expiry date
Challenge. Large organisations don't lose design-system consistency because designers don't care. Dozens of product teams, multiple Figma libraries, local exceptions and regulated accessibility requirements make drift inevitable without a system to catch it.
Action. Craftline scans live Figma files, flags what's actually broken, and routes new and changed components through four gated review stages. Designer scorecards combine adoption, repeat violations, remediation time and exception ratio into one explainable score.
Principle. Coaching, not blame — designers see remediation guidance, not just penalties, and the design-system team can always override, annotate and audit every decision the system makes.
