The Alpha Team · Los Angeles
The internet already got corrupted once.
Ask AI what to buy and it answers in a heartbeat, with total confidence and no accountability. Ask it to prove the answer and it can't.
It will only get worse as AI chat platforms start selling ads. We are not going to watch it happen to AI.
So we are building the verification layer for commerce: the part that checks whether a claim about a product actually holds up.
The Alpha Team is a group of people who get the product before the public does, and who argue with the people building it about whether the answers hold up.
Who this is for
Three kinds of people get the most out of the Alpha Team.
Most members are one of them. A few are two.
01
You are an expert in your field, and you can catch the AI getting it wrong.
You can read an AI answer about your own subject and see exactly where it went wrong, and why. You want what reaches people in your field to be true.
02
You use AI every day, and you want a real say in what it becomes for shopping.
You still check every recommendation before you spend money. You want a version that shows its evidence, so you can decide with confidence, including deciding not to buy.
03
You have your own research system for checking product claims before you spend money.
A spreadsheet, a custom GPT, the patience to read the actual studies. You built it because you did not believe the recommendations you were given.
What you get
What you get the day you join
01 · First eyes
The product before the public, and the people who built it to argue with
You use it while it is still being made, and you take your objections straight to the team building this product. The version everyone else ends up with is the one you pushed on.
Every answer opens up: the facts it leaned on, how many independent sources back each one, when each was last checked, and what it could not verify at all. Every one of those facts has a button next to it, so you can contest any single fact and say why.
you asked: is the salomon sense ride worth the money?
A confident pick
Get the Salomon Sense Ride.
Fits narrow feet, built for everyday trail miles, and $140 is enough.
VERIFIED · WHY IT WINS · 3 TRUTHS · CHECKED SEP 2026
The category's narrow-foot pick
RE-VERIFY · The category's narrow-foot pick
We'll re-check this against its source. If it doesn't hold, we fix it for everyone.
Light at 10.1 oz, cushioned for daily miles
33 cents a mile, the going rate for trail shoes
Test Team members open every receipt and can contest any verdict.
you asked: is the lumenpro sculpting wand worth $199?
No verdict yet
We can't call the LumenPro yet.
Every claim we found traces back to the brand itself, so we won't guess. This answer updates when independent evidence lands.
WHAT WE CHECKED
Every source we found is the brand's own
Nine sources carried a claim about this device, and every one traces back to the brand or a page it sponsors. Brand-owned evidence can describe a product; it cannot settle whether the product is worth $199, because nothing independent is testing the claim.
NO TRUTHS HELD YET · CHECKED SEP 2026
Test Team members read every receipt, and can challenge any answer.
02 · The method, pointed at your question
Your hardest question, forged
The patent-pending method behind our own tools, our company and our brand is now yours, pointed at the hardest question in your field.
You hand us the question
Five frontier models research it in parallel
We apply our patent-pending verification method to evaluate the claims
We share a full-length report plus all the verified claims (we call them axiomatic truths: claims that survived every attempt to knock them down) you can analyze with your favorite AI
Today a person here runs it for you.
What counts as a hardest question
It is about how something works: the physics, the mechanics or the psychology underneath. It is not a fact you could look up. You are genuinely stuck on it, and the right answer would unblock you and show you what to do next.
- Which trust signals still work in my field?
- How do people actually use music to change how they feel, and when does that help or make it worse?
- When two designs work equally well, what decides which one people prefer?
Three examples of the shape. One question at a time, and we run a limited number in each round, so there is a wait between yours and the next.
03 · The room
Working sessions with the people building Product.ai
Golden Hour is our public event, running the first Wednesday of every month in West LA. It is free, and it needs no application. Most members came to one before they ever applied. And people leave with more than they came with: somebody they met, or something they learned.
You learn from us, and from the group.
You learn from each other.
We learn from you.
A few times a quarter, members are invited to something closer in. Six or eight people at a table, working through the categories you know best or care most about, deciding what goes into the product and what stays out.
First time I could step into an AI company and wholeheartedly say I love what these people are doing.Elias ArjanLongevity operator, has vetted at least 300 brands over eight years · LinkedIn
I am the target audience for sure. All I want is real information about what products actually work and I can trust.Yira VilaroSkincare enthusiast · LinkedIn
04 · Your name on what shipped
The thing you flag gets built.
Tell us what is broken or missing, and see exactly what shipped because you said so, with your name on it.
What doesn't ship comes back as an honest not yet, and here is why. This is what makes building the product together real.
Why it works
Everyone here is trying to make everyone else smarter.
The tools are what we build together. The people are what you keep. It runs in three directions, and none of them involves selling you anything.
You learn from us.
How we actually run a company on frontier models, and what it takes to be found by an AI agent rather than a search engine. You see it live, as we build in public.
We learn from you.
The things your field knows that the AI still gets wrong, and how you react when we put an answer in front of you. No metric can tell us that.
You learn from each other.
Most of what people take home came from the person sitting next to them: a collaborator, or a fix for the thing that was blocking them.
What we ask
What we ask, and what we don’t
We ask for
- Your time. A few hours a quarter, with us and with the other members.
- Your experience. How you actually buy. What you rejected, and what made you doubt an answer you were given.
- Your knowledge. The things your field knows that the AI still gets wrong. It does not have to be deep expertise.
- Your opinion, especially when it is the opposite of everyone else's in the room.
And what we don’t
- Your contribution is never paid, and that is deliberate. Small payments turn people who want to be here into people doing a task. There is no equity for members and no points scheme either.
- There is no membership fee. Membership is not a subscription and there is nothing to renew.
- Nothing is scheduled for you and nothing is assigned. No quota and no minimum.
- You will never be sold to here.
- Paid work with us is a different door, through the Join page, and through Projects.
How you get in
Share two short stories
We ask for stories because what you have already done is the only signal we can trust. What matters is whether you were doing this before you heard of us.
Before you start. If early access is all you want, this will not be a good use of your time. The people who get the most out of it are the ones who push hardest on what we show them.
◆ Application
Thank you.
A person reads every one. If you look like a fit you get an invitation to a session.
A person reads every one. If you look like a fit you get an invitation to a session, where you meet the people building Product.ai and the members already here. That session is the rest of the decision, on both sides.
After that your access opens up, and we hope to see you regularly. If it goes quiet for a while, we will check in and ask whether you still want the seat.
Questions
Things people ask before they apply
Do I have to be a builder?
No. Some members build agents and ship products; others are here because they want to understand how the technology works and they think carefully. Both are useful, and the second kind asks the questions that catch us out.
Can I come to something before I apply?
Yes, and most people do. Golden Hour is public and free, and it lands on the first Wednesday of every month: the calendar is here. Come to one, talk to whoever is standing near you, and decide afterwards.
Do I hear back if the answer is no?
Yes, and the no comes with a reason. A person reads every application, so it takes a little time rather than a little compute.
Is there a fee, or is any of this paid?
No fee, no payment, no equity and no points. Contribution stays unpaid because small payments turn people who want to be here into people doing a task. Paid work happens through the Join page and through Projects.
How do you know people don’t trust AI recommendations for shopping?
We surveyed more than 1,400 US online shoppers about how they are using AI to shop. The headline surprised us: about 43% have used AI to research a purchase in the last 90 days, and among those people, 86% still verify what the AI told them through another source before they buy. Read the report.
Who owns the research report you do for me using your method?
You do. The question is yours, the report comes to you with your name on it, and you can publish it, send it to a client, or keep it. We keep the findings in our own record and we tell you which ones we kept.
Do I have to be in Los Angeles?
Golden Hour is in West LA, so most members are local today. Some working sessions run remote so people can join from other locations, but the in-person experience is always better.
Does this help me get a job at Product.ai?
No, and it is not meant to. Membership is not a hiring channel. If a job is the thing you actually want, apply for a role through the Join page. That is the front door.
Share two short stories
Not ready to apply? Come to a Golden Hour first. Most members did.