Skip to content

The500

A members-only matchmaking app and the admin system behind it.

Conversations list with Your concierge pinned at the top, above two conversations with matches
Introductions tab with a new profile card for Margaux, 29, a painter in Silver Lake, shown on a soft blue gradient instead of a photo
Profile review screen headed Claire refined your draft, with three written prompts and a voice prompt drafted from the member's call
Role
Founding Product Manager
Timeline
December 2023 to present
Platform
iOS, Android, web admin system
Tools
Figma, Plane, Claude, Claude Code, Supabase
Problem
A matchmaking firm with 30 years of clients had no app. Everything ran on phone calls, email, and spreadsheets.
My role
Product manager. I own the roadmap, the requirements, and the design direction for the member app and the admin system.
What I built
The member app, and the admin system concierges and recruiters use to run the business. I wrote the spec for the whole admin side.
Where it is
Built and not live yet. The company is getting ready to launch it properly.
What I'd change
Map every member-to-admin dependency before designing a single screen.

Overview

The500 is an invitation-only dating app for The Standard Agency, a matchmaking firm that has been pairing people up for 30 years. Members pay. Every Sunday they get 4 profiles their concierge picked for them, and 3 more on Wednesday. To like someone, you have to send a message. If it's mutual, it's a match.

There are really two products here. Members see the app. Concierges and recruiters work in the admin system, and what they do there decides what a member sees. One doesn't work without the other, so I've owned both.

Problem

The firm had the expertise and the clients but no software. That works when it's small. It doesn't scale, and it doesn't feel premium to someone paying for a premium service.

The app had to keep the part people pay for, which is a real person with good judgment picking for you. It also had to fix everything that breaks when you do that by hand. And it had to feel nothing like a swipe app. Dating apps have taught people to expect fake profiles and low effort, so it needed to be obvious from the first screen that a person chose these profiles.

What I did

I wrote the requirements for every part of it: onboarding, invitations, matching, messaging, reviews, blocking and reporting, billing, and the full admin system. I designed the two sides together. When I designed the drops for members, I designed the screen concierges use to build it at the same time.

The admin system is the piece I'm proudest of. I wrote the spec for all of it, from how applications get reviewed and waitlisted to how a new member gets assigned a concierge. I also own the business rules that tie it together, like how long a recommendation stays open and how many invites a member gets.

Rebuilding it with AI

In 2026 we redesigned both products and changed how we build them. The team went from three developers, two designers, and a project manager to one designer, two developers who build with AI, and a project manager.

That changed my job. I prototyped the admin system in code first, with Claude Code, and we moved the finished prototype into Figma as the design source. I used AI to turn meetings and decisions into versioned requirements, and to break those requirements into a full ticket backlog in Plane. The documents still exist. I just spend a lot less time typing them and a lot more time deciding what goes in them.

The original design felt dated. For a product priced like The500, it also didn't have the wow factor a member should feel the first time they open it. So we redesigned both products from scratch and used the chance to add a few new features too.

Old home screen with a blurred profile photo, the name Margo, 34, and pass, expand and like buttons
Before
New home screen. Introductions tab with a new profile card for Margaux, 29, a painter in Silver Lake, shown on a soft blue gradient instead of a photo
After

What we built

  • The drops. 4 profiles on Sunday and 3 on Wednesday, picked by your concierge. To like someone you write them a message. Profiles lead with photos, a voice greeting, and what the person is looking for.
  • Weekly check-in. At the start of each week the app asks what you're looking for right now. It goes straight to your concierge for the next drop.
  • Video prompts. A question and 15 seconds to answer it. The point is to get something unrehearsed.
  • Your concierge in your inbox. Pinned at the top of your messages from day one.
  • Invitations. Members can invite people they know. The invite is a single-use link they send from their own phone.
  • The admin system. Where recruiters review applications and manage the waitlist, and where concierges manage their members and build each drop. It filters the singles database on 20+ criteria.
Sunday's drop as a grid of four new profile cards, then two earlier recommendations with 36 and 50 days left
The drops
Like sheet for Margaux's profile with a required message, a send button, and an option to request a warm introduction
A like comes with a message
Preferences screen titled Who you'd like to meet, a ranked list starting with relationship intent, family plans and age
Preferences
Video prompt setup for the prompt A day in my life, with buttons to film now or choose from the camera roll
Video prompts
Conversations list with Your concierge pinned at the top, above two conversations with matches
Concierge pinned in the inbox
Invitation landing page at the500.org reading The most wanted room in the city, with a Begin your application button
Invitation landing page
Admin All Clients table listing clients with their status, concierge, member since date, last touch and open tasks
Client list
Admin Create Rec screen where a concierge builds a client's next deck from cards of singles, with three open slots on the right
Deck building
Admin applicant record with identity details, a family plans flag to confirm, and a quality score panel for the recruiter
Application review

Where it stands

The app is built and not live yet. One thing I've learned here is that a product doesn't work just because you switch it on. It needs real support behind it to launch well, and the company is setting that up now.