AI “Mom Test” Interview Script Creator
Try out this AI "Mom Test" interview script creator I made for my workshop attendees to start talking to customers faster.
It reduces the friction of finding and interviewing customers.
Enter a few fields into a form, the magic happens and an interview script customized to your product idea is created and sent.
AI "Mom Test" interview script creator (includes ICP):
In my Customer Interviewing workshop, the attendees learn different types of customer interviewing techniques, participate in role plays and practice their new skills by interviewing each other.
I made this AI "Mom Test" interview script creator to make it easier for them to go from inspiration to action (scroll down for the gory details of how I made the AI script generator).
Scroll down for the gory details of how I made it.
Details → How I created the Mom Test interview script generator
I used Make.com to string together several actions. This is the agentic approach to getting stuff done.
As a software engineer by training, I kept thinking it would be easier to just program this myself. Another Product Coach told me it probably would have been easier to just do this with Replit (a vibe coding tool). I like no code because I assume that the no code folks will update their code when APIs change and whatnot. But I hate no code providers when I can’t do simple things. It’s a tradeoff. I didn’t hit any fatal flaws with Make.com for this project.
The creation and parsing of the JSON was the most frustrating. I have been co-working with a client who is creating their own Make.com agent and we’ve been sharing tips.
This person is not an engineer by training and never touched the raw JSON. They just kept prompting chatGPT to modify and then output the updated JSON.
My sense is that every AI experience maker (code, no code, or vibe code) will need to be proficient in JSON. At least its not XML or CSV.
First, I had to figure out how to get chatGPT to return the same JSON keys each time (use a system prompt with a JSON “template”). Then I had to get the right combo and setup of Make.com modules to parse JSON lists into newline separated items which then turned into bullet points when finally inserted into the Google Doc.
This is the Google Doc template that gets copied, filled in and shared by this agent.
This agent took about 10 hours to make and about 5 hours to adjust after the first round of feedback.
These are the basic steps (Make.com diagram below):
User inputs a few fields into a Google Form
Make.com reads a row of data in the Google Spreadsheet output by the Form
Make.com turns the spreadsheet cells into variables. It formats those variables into a massive user + system prompt and sends that to chatGPT 4.1 API
Make.com then parses the JSON output with its JSON Parser/Iterator/Text Aggregator modules
Make.com creates a Google Doc from my template and then shares that new document with the requestor
It also send an email via Gmail (and it copies me letting me that someone has used the script generator)
Please give it a try and send me feedback:
Screenshot of the Make.com agent flow
The AI-Enhanced Product Manager
Jim coaches Product Management organizations in startups, growth stage companies and Fortune 100s.
He's a Silicon Valley founder with over two decades of experience including an IPO ($450 million) and a buyout ($168 million). These days, he coaches Product leaders and teams to find product-market fit and accelerate growth across a variety of industries and business models.
Jim graduated from Stanford University with a BS in Computer Science and currently lectures at University of California, Berkeley in Product Management.
