Get your first 10,000 users for an AI app by picking one channel and working it daily for 90 days. Post as the founder, answer questions in the communities where your users already are, and build a reason to share into the product itself. Measure retention before you measure growth. Paid ads come later.
Why the first 10,000 is a different job
The next million users come from systems: paid creative at volume, a blended CAC you can predict, a product that spreads on its own. The first 10,000 come from you. From doing things by hand that do not scale and would embarrass a growth team.
At 10,000 users you do not yet know your best channel, your retention, or which message lands. The job at this stage is to find those three things while adding users. I treat the first 10,000 as research with a growth number attached, not as a small version of scale.
The ZuAI case study picks up at roughly 10,000 users and runs to 2 million. This post is about getting to the starting line.
Pick one channel
The most common mistake I see is trying five channels at once, 2 hours a week on each, and concluding that none of them work. Nothing works at 2 hours a week.
Pick one channel based on where your users already spend time and where you can show up daily for 90 days. The candidates for an AI app are usually:
- Founder-led content on one platform (X, LinkedIn, or YouTube).
- Communities (Reddit, Discord, niche forums).
- Short-form video (TikTok, Reels, Shorts).
- Launch platforms, as a one-off event, not a channel.
Pick one primary. Add a second only when the first one is producing users every week without you thinking about it.
Founder-led content
Nobody can explain why your app exists better than you. At this stage that is your only unfair advantage, so use it.
What works in my experience:
- Build in public. What you shipped this week, what broke, what users said. Real numbers when you have them.
- Show the product doing the thing. Screen recordings with your voice over them. No slides.
- Answer the questions people ask about your category. Not your app, your category. The app comes up naturally.
- Post daily on one platform. Consistency beats quality for the first 90 days.
The goal is not followers. The goal is that when someone in your niche has the problem you solve, they have already seen your face attached to it. I run this as a service at founder-led content, but at this stage you should do it yourself.
Communities: Reddit, Discord, niche forums
Communities are where your first users already are, asking for what you built. They are also where a pushy founder gets banned in a day.
The rules are the same everywhere:
- Be a member for weeks before you are a founder. Answer questions with no link.
- Disclose. Every mention of your app says it is yours.
- Help 10 times for every 1 mention.
- Post a launch thread as yourself. What you built, why, what broke, what you want feedback on.
- Stay in the thread and answer everyone.
Reddit is the biggest of these and the one I have tested most, across 40+ accounts, each with real standing. Discord servers and niche forums work the same way at smaller scale. Full details in my Reddit marketing guide.
Short-form video
Short-form video is the fastest way to show an AI app doing something surprising. A 15 second screen recording with a hook reaches people who have never heard of you, at zero media cost.
How I approach it before there is any ad budget:
- Face plus screen performs better in my campaigns, but screen alone works if you are camera shy.
- One video, one use case. “Watch it do X.” Nothing else.
- Post 1 a day for 30 days. The first 20 will be bad. That is the cost of learning what hooks work.
- Reuse the hooks that got views. Same hook, new use case.
This organic phase is also research for later. The hooks that work organically become the brief for paid TikTok UGC when you get there.
Launch platforms
Product Hunt, Hacker News, and similar sites are a one-day spike, not a channel. They are good for a burst of early adopters who like trying new tools, some backlinks, and a story you can reuse.
- Launch after you have a few hundred users and some retention data, not on day 1.
- Prepare the story: what it does, why you built it, what is different.
- Reply to every comment for 24 hours.
- Expect a spike and a drop. Measure the users who are still around 30 days later, not the launch day number.
Referral inside the product
The cheapest user is one your existing user brings. Build the share moment into the product before you spend a dollar on ads.
- Find the moment of delight. The point where the AI output surprises the user. Put a share button there, not in the settings menu.
- Make the shared thing valuable on its own. A result, an image, a summary the recipient can use. Not a coupon.
- Keep any reward simple. Extra credits or a free month. Count it as acquisition spend in your CAC.
- Track referred users separately and compare their retention.
What to measure at this stage
Do not build a dashboard. Track 5 numbers in a spreadsheet, once a week.
- New users this week, by source, from a signup question if attribution fails.
- Activation rate. The share of new users who reach the core value once. Define it precisely.
- Week 1 retention. The share of new users still active 7 days later.
- Blended CAC. Total acquisition spend, including your time, divided by new users. Even when the spend is small.
- Referral share. The share of new users who came from another user.
Retention comes before growth. If week 1 retention is poor, more users means more churn. Fix the product, then go back to the channel. I keep free calculators for CAC and payback on my tools page.
A weekly cadence
This is the cadence I hand to founders at this stage:
- Monday. Review last week’s 5 numbers. Pick one thing to change.
- Tuesday to Thursday. Daily content on your primary channel. Daily community time, 30 to 60 minutes. Ship the product change from Monday.
- Friday. Talk to 3 new users. Ask what they were trying to do and what almost stopped them.
- Weekend. Record next week’s videos or write next week’s posts in one batch.
Ninety days of this cadence is what I ask founders to commit to. If it does not produce users, it will at least tell you why.
When paid ads make sense
Paid ads make sense when three things are true:
- Week 1 retention is at a level you would be happy to pay for.
- You know which message and use case converts, from organic content.
- You can afford to spend for a month and learn nothing.
Until then, ads buy churn. Once those are true, a small paid test takes the hooks that worked organically and puts budget behind them. That is where the user acquisition side of my work starts, and where the playbook shifts from hand work to testing at volume.
If you are between zero and 10,000 users right now, the Goal Digger Club is being built for this stage. The waitlist is open. And if you want to talk through your app, book a call with me.
FAQ
How long does it take to get 10,000 users for an AI app?
It depends on the use case and the channel, and I do not put a number on apps I have not seen. What I can say from my own work is that founders who show up daily on one channel for 90 days get a clear answer either way. Either the users come, or the retention data shows what to fix first.
Should I spend money on ads to get my first users?
Not as the main channel. A small paid test can help you learn which message converts, but before you have retention data it mostly buys installs that leave. Use organic channels to find the message and the use case first. Then put money behind what already works. Ads are an amplifier, and amplifying nothing is expensive.
Do I need a marketer for the first 10,000 users?
No. You need a founder who will do the unscalable work daily. A marketer becomes useful once there is a channel that repeats and a CAC to manage, which is the stage I usually step in as an AI app growth marketer. Before that, the best money you can spend is on the product and on talking to users.