Summarize with AI
A waitlist landing page for an AI startup should do three jobs: say plainly what the product does, show one real example of it working, and ask each person who joins one question about what they would use it for first. A page that only collects email addresses gives you a list, but at launch most of that list never opens the email and almost nobody on it becomes a user.
This matters more for an AI product than for most. Before launch, the waitlist is the only proof of demand you have, and it often ends up in the pitch deck as one big number. If the page never explained the product, that number counts curiosity. The people who joined could not tell what they were signing up for, so you cannot tell who they are either.
A grey page, one email field, and 3,000 addresses
Picture an AI tool for wedding photographers. After a wedding, a photographer comes home with thousands of frames and spends a weekend picking the few hundred worth editing. This tool reads the whole shoot, groups the frames by moment, and picks the best one in each group. The product is made up for this example, and so is its page.
The waitlist page has a dark grey background and a blurred photo of a camera lens. The headline reads The autopilot for photographers. Under it, in light grey: Something new is coming. Be the first to know. Then a single email field with the placeholder name@studio.com and a white button labelled Join the waitlist. Below the button sits a small counter: 2,814 people have joined. No sample shoot, no before and after, and no word about weddings, frames or culling.
After someone joins, the button turns into a line of text: Thanks! We will be in touch. Nobody is asked anything. Over five months the counter climbs past 3,000.
On launch day the whole list gets one email. The subject line is We are live, and the body is one paragraph and a login link. Most of the list never opens it. A small group signs up, uploads nothing, and leaves. The few who upload a full wedding and finish a cull are mostly photographers the founder already knew before the page went up.
What the founder did with the counter
He checked the waitlist count most mornings for five months. Each time it passed another five hundred, he posted a screenshot in the team channel. The number went into the seed deck, on the traction slide, under the line Over 3,000 on the waitlist. When an investor asked how many of those people shoot weddings for money, he did not know, because the page had never asked.
After launch he blamed the email. He rewrote the subject line and sent a second announcement, then a third with a discount code. Each one was opened by fewer people than the one before.
Every address on that list was paid for, in ads, launch posts and his own evenings on LinkedIn. The list was also a promise to investors. If a round is pitched on a slide that says 3,000 and the launch produces a few dozen active photographers, the next conversation starts from a number that shrank. The seed bar for product-led companies is now roughly 300 to 500 thousand in ARR, and the median gap from seed to Series A is around 616 days across all sectors. And the funnel keeps leaking after signup: product-led benchmarks, measured on SaaS companies, put activation at 20 to 40 percent of signups, and that is for people who knew what they signed up for.
Why minimal waitlist pages fail AI products
Most waitlist advice says keep it minimal. A headline, an email field, a button, and nothing that could pull attention away from the signup. That advice serves one goal, the biggest possible list. For an AI product it is the wrong goal, because a big list of people who do not know what you do is worth less than a small list of people who do.
The page above fails a stranger in the first four seconds. The autopilot for photographers could describe a dozen products. It borrows the X for Y shape of a famous tagline and skips the one word that matters, which is culling. A photographer cannot tell if this is an editing app, a booking tool or a website builder. So the people who join are the ones who join anything: other founders, tool collectors, hobbyists who like a new app. The working photographer with Saturday's wedding still unsorted on her laptop reads it, does not see her problem, and closes the tab.
The founder cannot see any of this. He knows exactly what the product does, so every vague line on the page gets finished in his head. That is the whole problem in one page. He knows his product and nobody else does, and the page assumes they do.
Write the page as if the product were already live
The fix is to write the waitlist page the way you would write the launch page, then swap the signup button for a waitlist form. Here is the same page, rewritten.
Headline: Cull a wedding in minutes, not a weekend. Subhead: Upload the full shoot. The AI picks the best frame from every moment, flags closed eyes and missed focus, and sends your picks straight to Lightroom. Opening to photographers in waves from January.
Below that sits one worked example from a sample wedding. On the left, the upload: 3,412 frames. On the right, the result: 486 picks across 61 moments. Under one moment, labelled First look, eleven near-identical frames sit in a row. One is marked Pick, two are marked Eyes closed, and one is marked Soft focus. A small link under it reads Browse the sample wedding. Then one honest line: Works with RAW files from Canon, Nikon and Sony. Fujifilm comes later.
This version does three things the first one could not. It names the user and the job in the headline. It shows the product doing that job, which is the only proof an unreleased AI product has, and how to show an AI product on a landing page covers picking that one input and output. And it says what the product does not do yet, so the people who join are the people it will work for on day one.
This is also where the look of the page earns its place. A stranger decides what kind of company this is from the surface, and a careful, specific example reads as a real product with a team behind it. A blurred lens and a counter read as an idea. When Chariot, a speech AI lab, was about to put its models in front of the public for the first time, the site we designed for it had each section show what the model does before the text explains it.
One question after the email, never before it
The biggest single change is a question. Do not put it in front of the email field, where it costs signups. Put it on the step right after someone joins, where the old page said Thanks.
For the culling tool, the question is What do you shoot most? It has four options, one tap each: weddings, portraits and families, events and corporate, or I do not shoot for clients yet. Then one optional field asks how many shoots a month, with three ranges to pick from.
- It sorts the list. On launch day you know which people have the exact problem the product solves today, and you invite them first.
- It tells you what to build next. If many answers say events, that is the next kind of shoot to support well, picked by the people who will use it.
- It lets you write a launch email about each reader's own problem instead of one email about your product.
Founders worry the question will lower the count. It will, a little, and that is the point. Someone who skips a one-tap question about their own work was never going to upload a full wedding later.
Referral mechanics that bring in the right people
The usual referral setup is a place in line. You join, see that you are number 2,815, and move up for every friend you invite. It makes the list grow fast. It also fills the list with people who wanted to win a line, and with their friends, who wanted the same thing.
For an AI product with a real job to do, reward the invite that brings a user. On the culling tool's page, the screen after the question can ask: Do you work with a second shooter? Invite them and you both get in with the first wave. A photographer invites her second shooter. That second shooter works with six other lead photographers through the season. One invite like that is worth more than fifty strangers chasing a spot.
Keep the reward to earlier access for both people. Earlier access only matters to someone who wants the product, so it filters the list while it grows it.
Turning the list into users: what to change this week
If your waitlist page is live, you can fix most of it in a week.
- Rewrite the headline so it names the user and the job. Read it to someone outside the company and ask what the product does. If they guess wrong, rewrite it again.
- Add one worked example with a real input, the real output and the numbers in it. A static image is fine if the product is not stable enough to embed.
- Add one line that says what the product does not do yet, and roughly when you open. A month or a season is enough.
- Add one qualifying question on the step after the email. One tap, four options at most.
- Change the referral reward to earlier access for both people, aimed at the colleague who would actually use it.
- Email the people already on the list before launch and ask them the same question. The ones who answer are your first wave.
- At launch, invite in waves, starting with the answers that match what works today. Send each wave one button that opens the first useful screen, not an empty dashboard.
- Count activated users from the list, not signups. A photographer who finished one cull is a user. An account with no upload is not.
The last step matters for the raise as well. On product-led SaaS benchmarks, a ten point rise in activation typically drives a 15 to 25 percent increase in free-to-paid conversion. A smaller, sorted list that activates tells a stronger story than a big one that goes quiet. For the pages your site needs once the waitlist turns into a launch, see what a launch-ready website for an AI startup includes.
Studio Maydit designs websites for AI founders in the US, UK and Europe, and a waitlist page is often the first thing we rebuild: a headline a stranger can repeat back, one example that shows the product working, and a form that tells you who joined. We build in Framer, Webflow or custom code, usually as a fixed scope of three to four weeks, so the page is doing its job well before the launch email goes out. If you are comparing studios for this, our ranking of landing page design agencies for pre-seed AI startups lays out the options. If your waitlist keeps growing and you cannot say who is on it, book a 30-minute call with us.
