Conversational AI marketplaces
Expert advice, one conversation away
01 Overview
One technology, its first marketplace
Expert advice costs a lot because most of an expert’s time goes into finding out what the client needs. Scalez hands that part to a conversational AI agent: experts only step in to answer, and clients pay a fraction of the price.
-
My role
Platform and product designer
I set the visual direction for Scalez and designed its website, then designed Savvy for both sides: two websites and a mobile app with one experience for clients and another for stylists.
-
The problem
A new model nobody had seen yet
Expert advice was expensive because intake ate expert time, Scalez had no mental model to borrow, and nobody would type styling details into a form — the marketplace had to earn them through play.
Three chapters, one structure
Scalez is the platform, Savvy its first use case — so the story starts with the technology and moves to the marketplace built on it.
02 Scalez website
Explaining a marketplace nobody had seen yet
Challenge
Scalez sells an idea before a product: an AI agent that interviews clients, briefs experts and relays their answers. “Conversational commerce” and “NLP” explain the engine, not why anyone should care.
Outcome
A bright, illustrated site that states the belief in two sentences, walks through the model in four steps, and points to Savvy as proof.
Four directions before one felt right
A comic style, a dark crowd theme and a dark faces theme all landed too loud or too serious. The winner: a light page with one character per idea and a friendly robot guiding every step.
-
Comic
Hand-drawn characters, black and yellow
-
Crowd
Silhouettes and dark chat cards
-
Faces
Dark theme with real portraits
Final home page
The business model as a four-step story
How it Works keeps one cast: a client, the robot agent, a group of experts. Whoever isn’t in the step fades back, so the eye follows the conversation from client to agent to experts and back.
The agent interviews and summarizes, so experts spend their time only on the answer.
What it achieved
A three-way AI-matchmaking model that’s hard to explain in words became a diagram and four illustrations anyone could follow — visitors understand how Scalez works before they finish scrolling the home page.
03 Savvy design system
Structure before the screens
Challenge
Two apps and two websites, one developer to build them all. Anything left open cost an hour of deciding, then a round of QA when the next screen answered it differently.
Outcome
What repeats, built once: two color ramps, one type scale, a four-pixel spacing rule, and components that carry their own states.
Header, navbar, button, input, card, avatar, chat bubble all went onto one Components page before any real screen. Screens got assembled from named parts, and a rethink cost half an hour instead of two days.
Savvy UI kit: two five-step color ramps with their tokens; primary, secondary and text buttons in primary, hover, deactivated and enabled states, with icon and size variants; pill input fields in placeholder, filled, error and success states plus a dropdown; the one navbar component shown on both platforms, UP for clients with Home, Trends, Tips and My Closet and SP for stylists with Inbox, Chat, Insights and Profile; avatars in four user types and four sizes; tags, category chips and a rating; chat bubbles from Savvy and the client; and a StyleUp swipe card with a style match badge.
04 Savvy websites
One brand, two front doors
Challenge
A marketplace has to win both sides at once. Clients want confidence without endless shopping; influencers want to turn followers’ style questions into income. One page can’t pitch both.
Outcome
Two sites in the same blush and navy voice: one sells a personal stylist, the other a styling business — each leading into its own side of the app.
Same look, a different pitch
Same header, footer, chat bubbles and fashion-magazine serif. Different promise at the top, different sign-up at the bottom.
-
For clients
A personal stylist in your pocket
-
For stylists
Turn followers into paying clients
-
The AI
For clients: Asks the questions, so the stylist knows you
For stylists: Interviews followers, so you skip to styling
-
Leads to
For clients: The client app and a free style quiz
For stylists: Early access to the stylist app
05 Savvy app for clients
Getting styled should feel like play
The client side plays like a game: quick taps, swipeable cards, visible progress, a small reward at every step — each flow short enough for a coffee break.
Every swipe and answer is a question the stylist doesn’t have to ask.
-
Feels like
Client app: A game: swipe, like and level up
Stylist app: A workspace: inbox, client profile and outfit builder
-
Starts with
Client app: A chat with Savvy and a stylist match
Stylist app: A personal styling page to share with followers
-
Shares
Client app: Taps, swipes, photos and short answers
Stylist app: Full outfits, advice and a video message
-
Gets back
Client app: A styled look and a closet of saved items
Stylist app: A client profile, ratings and income insights
Chat, match and level up
A short chat interview, then Savvy matches her with a stylist rated highly by women like her — five more to choose from.
Home is a journey, not a form: the first week builds a style profile in short sessions on body shape, colors and style, each unlocking the next, each saying what comes after and how long it takes.
-
Last question
-
Stylist match
-
Top 6 stylists
-
First week
-
What’s next
-
Style advice
StyleUps and Trends that learn
StyleUps are short swipe sessions: pick up to three categories, read a tip, swipe stylist picks — heart to keep, cross to skip. An empty state keeps items coming when tips run out, and each session ends with a one-tap rating.
Trends is the same idea as a feed: stylist-curated looks where every heart tunes the next ones, explained by a Style Match badge.
-
Today’s StyleUps
-
Pick categories
-
Style tip
-
Swipe items
-
Out of tips
-
Rate it
-
Saved to closet
-
Like a look
-
Style match
06 Savvy app for stylists
Everything a stylist needs before the first look
The stylist side is calm and practical. Stylists work through several clients from a phone, so every screen answers one thing: who’s next, what do I know about her, what do I send back?
From profile to styled look
The inbox shows who’s waiting and what she wants. Each client opens into everything she shared while playing — one summary sentence, photos, body details — so styling starts without an interview.
Outfits follow a stylist’s structure — basic, interest, completer, accessories. Each slot opens a searchable catalog with brand filters and Savvy suggestions, sent with a note explaining the pick.
-
Inbox
-
Her photos
-
Body
-
Empty outfit
-
Pick a category
-
Full outfit
-
Talk to the client
-
Sent StyleUp
Run it like a business
The rest of the app helps stylists grow: the inbox nudges them to finish the oldest StyleUp and share their link, separate lists for trial clients and subscribers, and Insights tracking sessions, reviews, profile views and trial-to-subscription conversion.
-
Invite clients
-
Subscribers
-
Client thread
-
Insights
-
Reviews
-
Profile
07 Takeaways
What this work taught me
-
Design both sides of a marketplace
Clients and stylists want different things from the same product. The client app plays, the stylist app works — both run on the same data.
-
Turn the questionnaire into a game
Swipes, levels and ratings collected more than a form would, and every answer landed where the stylist needed it.
-
Show the AI, but lead with people
A friendly robot made the technology visible, then handed over to a human. The benefit comes first.
My impact
I gave an idea nobody had seen a visual language people got in seconds, then built Savvy on top of it end to end — two websites, a client app that plays like a game, and a stylist workspace where every brief arrives ready-made. It shipped as a live two-sided marketplace.
