Back in the day I made a rookie mistake over and over when launching new business ideas.
I'd spend weeks building something, get excited about testing, and then... share it with my mum.
"Oh darling, this is wonderful!" she'd say.
My friends were just as bad - "Looks great mate!" or "Yeah, cool stuff!"
And there I was, feeling on top of the world, thinking I'd cracked it. Until I'd launch to actual users who - surprise, surprise - weren't nearly as kind as my mum. Weird right? 🤣
Of course they weren’t as positive! They were busy, distracted, sometimes downright hostile. And they found problems. Lots of them.

Here's the thing: your mum is always going to be a fan. Your friends will always try to be supportive.
What you actually need are strangers who couldn't care less about your feelings. People with real problems to solve and limited patience. They're the ones who'll tell you the brutal truth about what you are building.
And today, I'm going to show you exactly how to find these beautifully honest critics - even if you don't have an existing audience or network.
Let’s get started:
Systematic approaches to testing (beyond just hoping it works)
Different types of testing you need to consider
Common failure points and how to catch them early
Before we dive into finding testers, we need to talk about how to get genuine feedback. There's this brilliant book called "The Mom Test" by Rob Fitzpatrick that completely changed how I approach testing.
Obviously should be “The Mum Test” but Rob is Canadian so he gets a pass! 😉
The core idea? If you ask someone "Would you buy this?", pretty much everyone will lie to you to protect your feelings. Especially your mum!
Instead, Fitzpatrick teaches us to focus on concrete past behaviours and real problems.
Don't ask "Would you use an AI tool to write social media posts?" Ask "Tell me about the last time you wrote a post. What was hard about it? What tools did you use?"
This is perfect for AI tool testing because we want to see how people actually interact with our tools, not hear polite opinions about them.
How do we practically ask these questions though?
I suggest 15-minute demo calls. Not lengthy feedback sessions, not surveys - just watching someone use your tool while you keep quiet.
Following The Mom Test principles, we want to observe real behaviour and get concrete facts rather than abstract opinions. That's why I use 15-minute demo calls where we:
Start by learning about their current process ("Tell me about the last time you did this task")
Watch them actually use the tool (without explaining or defending it)
Look for signs of real interest versus polite feedback
Focus on specific problems they've faced, not hypothetical future use
The staying quiet part is crucial. Your job is to observe and take notes, not defend or explain. Let them struggle. Let them get confused. That's where you'll find the insights that matter.
Here's a prompt to help structure your demo calls:
You are an AI expert in user testing, following The Mom Test principles of focusing on past behaviours and concrete facts rather than future intentions.
Create a structured guide for a 15-minute demo call that includes:
Pre-Demo Questions:
- Questions about their current process ("How do you handle this today?")
- Questions about specific recent examples ("When was the last time you did this?")
- Questions about problems and costs ("What's the hardest part about this?")
During Demo Observation:
- Behaviours that indicate real vs polite interest
- Signs of genuine frustration vs politeness
- Moments when they compare to existing solutions
- Comments about their current workflow
Post-Demo Exploration:
- Questions about specific features they actually used
- Questions about how this compares to their current process
- Ways to explore casual comments about problems or needs
- How to dig deeper without leading questions
Include additional guidelines for:
- Staying quiet during testing
- When to probe deeper (and how)
- Red flags that indicate someone is being politely positive
- Notes on capturing concrete facts vs opinionsHere’s what this might look like in practice:
Start with Past Behaviour (5 mins) "Walk me through the last time you created social media content" rather than "Would you use an AI tool for social posts?"
Let Them Drive (7-8 mins) Hand over control and watch them use your tool. Fight the urge to explain or help. Their confusion is your insight.
Look for the Gold (2-3 mins) Pay attention when they:
Mention specific problems ("This part always takes ages...")
Compare to existing solutions ("Oh, this is like X but faster")
Show genuine excitement ("Can I get access to this now?")
Express real frustration ("I don't understand why it's asking for...")
The key is staying quiet and neutral. Don't pitch, don't defend, don't explain unless they're completely stuck. Just watch and learn! It’s hard but with practice you’ll get it.
From a practical POV some extra pointers:
Use Zoom or similar - you want to see their screen and their reactions
Record the session (with permission) - you'll miss things in real-time
Have your prompt-generated questions ready but don't be rigid about following them
Take notes about what they actually do, not what they say they'll do
Pay special attention to moments when they go off-script - that's often where the real insights hide. Listen to the words they use. This is gold.
Now for the tricky part - finding these testers! Having an audience makes this infinitely easier (and if you want to build one, check out our various Audience Playbooks). But what if you need testers right now?
Here's what works:
Direct Outreach: Find potential customers on LinkedIn or Twitter. Look for people following related accounts or discussing similar problems. Be transparent - tell them exactly what you're building and what you need.
Partner Up: Find someone who already has an audience in your space. Maybe they're building something complementary, or they're just interested in what you're doing. Remember, tool building isn't enough - you need distribution too! (We'll dive deeper into this in our upcoming Launch Playbook).
Borrow my audience. I’m running a 30 Day AI Agent Accelerator next year. Part of this will include me helping you find testers and then promoting your product. Waitlist here: https://heyform.net/f/ZCCsfMqx
In both cases make sure to value their time. Offer something meaningful in return:
Free lifetime access to your tool
A strategy consultation in your area of expertise, ie. a coaching call
Direct payment (yes, really - good feedback is worth paying for!)
Whatever else you have that might be valuable to them
What if you can’t find testers?
If you can't find people willing to spend 15 minutes testing your tool, that's valuable feedback in and of itself!
It might mean:
Your target audience isn't right
Your value proposition isn't clear
If this means going back to the drawing board or even scrapping the tool then it’s better to do so now (early!) rather than in 6 months after more time/energy/money has been ploughed into the project. Doesn’t feel that way at the time but it really is!
OK! Now go and actually do this. Most of you won’t. Because it’s hard. But running this exercise is the single most important pre-launch activity you have. Suck it up. Sorry!
In Part 3, we'll tackle what to do with all this feedback - how to sort the useful from the noise, and how to implement changes without getting overwhelmed.
Keep Prompting,
Kyle