Ai study: Designing faster without compromising craft

Project overview

Eugene Labs offers at-home genetic testing across two categories — reproductive carrier screening and preventative health. Despite strong product-market fit, behavioural data showed the homepage was failing to convert: high drop-off rates, frequent back-navigation, and mis-clicks all pointed to the same problem. Users couldn't identify which product was relevant to them, and left without exploring further.

The redesign had two objectives. Structurally, rebuild the page's information architecture to create clear pathways into each product line. Strategically, evolve the brand from a startup aesthetic toward the trusted, clinical-grade positioning the founders were targeting — a shift that, in genetic health, is as commercial as it is visual.

Stage 1 — Research, analytics and session data

The process began with a thorough audit of behavioural data using Amplitude and session recording tools. Rather than relying on assumptions, the goal was to let the data surface where the experience was breaking down.

Key signals included elevated bounce rates on the homepage, high rates of back-navigation shortly after arriving, and scroll depth patterns that suggested users were scanning without finding a meaningful anchor point. Click maps revealed mis-clicks concentrated in the hero area — evidence that users were trying to self-select into a product category but couldn't find a clear mechanism to do so.

This analysis formed the brief: the page needed to do less, more clearly. Two user intents — reproductive health and preventative health — needed to be treated as distinct entry points rather than competing messages sharing the same real estate.

Stage 2 — Wireframe validation

With the structural problems defined, the next step was to test solutions before committing to visual design. Low-fidelity wireframes were built in Figma, these lo-fi frames were imported into Claude and used to generate high-fidelity wireframes.

This accelerated the feedback loop with the founders considerably, allowing visual direction, content hierarchy, and brand tone to be validated in a single round of review rather than across multiple design iterations.

Stage 3 — Image and content production

With the structure approved, attention turned to the visual layer. The existing image library leaned heavily on stock photography — a mismatch for a brand positioning itself as a trusted, clinical-grade product.

Generative AI tools were used to expand the existing asset library. The result is imagery that feels specific to Eugene rather than generic to the category — warmer, more considered, and more aligned with the brand's tone of "medical-grade genetics, human-grade care."