← Back to Works

Meet

Meet

Vera.

Vera.

An AI stylist that helps users create and edit outfits in Verifyt.

An AI stylist that helps users create and edit outfits in Verifyt.

2.2x

as likely to return after the first week

Role

Product Design Lead

Product Design Lead

Timeline

May-Jul 2026

May-Jul 2026

Platform

Verifyt consumer app

Verifyt consumer app

Overview

Context

Vera was an AI stylist designed within Outfit Builder to turn styling requests into editable outfits.

Challenge

Few users reached outfit creation, while the manual builder required too much effort to get started.

Approach

Vera needed to provide useful guidance without becoming a separate chatbot or taking creative control away from the user.

Impact

2.2x

as likely to return after the first week (outfit creators vs non-creators)

Context

Experience

Key decisions

Impact

Reflection

Context

Outfit Builder let users arrange clothes on a canvas, but they had to find and combine each item themselves. I led the design of Vera, an AI stylist built into the builder, so someone could start from an occasion ("something for a job interview") and get a complete outfit they could still edit piece by piece.

How might we help users turn a styling idea into an outfit they can make their own?

Experience
Users describe what they need, choose an outfit and refine it directly on the canvas.

Start with an idea.

Users can choose a suggested prompt or describe the occasion, style or item they have in mind.

←

View

01

Describe your idea

←

View

02

Generate an outfit

←

View

03

Keep favourites, change the rest

←

View

04

Edit, save or publish

Start with an idea.

Users can choose a suggested prompt or describe the occasion, style or item they have in mind.

View

01

Describe your idea

View

02

Generate an outfit

View

03

Keep favourites, change the rest

View

04

Edit, save or publish

Key decisions
AI gave users a starting point they could keep changing.
01
Keep Vera where outfits are made

The first version opened Vera in a full-screen chat. Without the canvas in view, users saw her as a separate assistant and asked broader questions, such as “What’s my style type?” or “How many body matches do I have?” Vera couldn’t answer these yet, leaving users frustrated by repeated dead ends.

We brought Vera into Outfit Builder, keeping the canvas visible to make her role clearer: helping users create outfits. Her suggestions appeared directly on the canvas, where users could move, swap or remove individual pieces.

  • Dress me for job interview

  • Casual weekend outfit

  • Date-night look

  • Old money aesthetic

  • Wedding guest outfit

02
Help users get started

We started with an “Ask anything” field. In early testing, most people left it blank or typed a single word like “dress”. They weren’t sure what they could ask Vera.

We added suggested prompts, such as an outfit for an interview or a look built around one item, to give users a starting point. They could tap a prompt or write their own request.

03
Let users try Vera first

Vera launched as part of Pro plan to see whether users would pay for AI styling.

After a week, few people had tried it. We couldn’t tell whether they weren’t interested in Vera or didn’t want to subscribe, so we removed the paywall to see whether more people would use it.

01
Keep Vera where outfits are made

The first version opened Vera in a full-screen chat. Without the canvas in view, users saw her as a separate assistant and asked broader questions, such as “What’s my style type?” or “How many body matches do I have?” Vera couldn’t answer these yet, leaving users frustrated by repeated dead ends.

We brought Vera into Outfit Builder, keeping the canvas visible to make her role clearer: helping users create outfits. Her suggestions appeared directly on the canvas, where users could move, swap or remove individual pieces.

03
Let users try Vera first

Vera launched as part of Pro plan to see whether users would pay for AI styling.

After a week, few people had tried it. We couldn’t tell whether they weren’t interested in Vera or didn’t want to subscribe, so we removed the paywall to see whether more people would use it.

  • Dress me for job interview

  • Casual weekend outfit

  • Date-night look

  • Old money aesthetic

  • Wedding guest outfit

02
Help users get started

We started with an “Ask anything” field. In early testing, most people left it blank or typed a single word like “dress”. They weren’t sure what they could ask Vera.

We added suggested prompts, such as an outfit for an interview or a look built around one item, to give users a starting point. They could tap a prompt or write their own request.

Impact
Outfit creation was linked to stronger return, but Vera’s early adoption remained limited.
Outfit creation was linked to stronger return, but Vera’s early adoption remained limited.

2.2x

higher return rate among outfit creators

36.8% of outfit creators returned on Day 7 or later, compared with 16.8% of users who did not create an outfit.

*May–June 2026 campaign · 1,224 US iOS users · Includes manual and AI outfit creation.

36.8% of outfit creators returned, compared with 16.8% of non-creators.

1 week

Brief monetisation test

Vera was introduced within Free, Pro and future Premium plans. Limited adoption showed that users needed access to the value before being asked to pay.

Access before monetisation.
Access before monetisation.

We dropped the paywall a week in and opened Vera to everyone. At that stage, getting people to use it mattered more than what we could charge for it.

Don’t gate a behaviour before people have learned why it is valuable.

Reflection

Keep users in control

Most of the work on Vera wasn't the conversation. It was deciding how much the AI should do and how much to leave to the user. A generated outfit gives people a fast start, but it only helped once they could see what Vera had done. Users could then keep the pieces they liked, replace the ones they didn't, and carry on manually.

Let users experience the value first

Looking back, I would give users more time to try Vera before introducing a subscription.

Further explorations (Not released)

I later explored a guided Vera onboarding experience and opportunities for her to appear at relevant moments throughout the app, rather than waiting for users to discover her inside Outfit Builder. These concepts weren’t released, but show how Vera could grow beyond Outfit Builder and become part of the everyday experience.

Onboarding

Instead of asking users to fill out a preference form, Vera could appear earlier and learn about their needs and style through a simple conversation.

Wardrobe & Search

Vera could suggest outfit combinations when a search was too broad or returned no useful results.

Shop

Vera could use saved items and browsing activity to suggest clothes that match a user’s style.

Onboarding

Instead of asking users to fill out a preference form, Vera could appear earlier and learn about their needs and style through a simple conversation.

Shop

Vera could use saved items and browsing activity to suggest clothes that match a user’s style.

Wardrobe & Search

Vera could suggest outfit combinations when a search was too broad or returned no useful results.

Back to top

Explore other works