Growth & Strategy
Personas
SaaS & Startup
Time to ROI
Short-term (< 3 months)
Last year, a potential client approached me with an exciting opportunity: build a two-sided marketplace platform with a substantial budget. The technical challenge was interesting, and it would have been one of my biggest projects to date.
I said no.
Not because I couldn't build it, and not because the money wasn't good. I turned it down because they had the classic startup blindness that AI and no-code tools have made worse, not better. They thought because they could build faster and cheaper with AI, they should skip the hard work of finding product-market fit.
Here's what I told them instead—and why this conversation happens almost weekly now with AI-excited founders:
Why modern AI tools create a dangerous PMF illusion
The one-day MVP framework that actually tests demand
How to use AI for validation, not just building
Real examples of what worked (and what catastrophically failed)
The framework I now use with every potential client before we build anything
This isn't anti-AI advice. This is pro-business reality. SaaS founders need to understand that AI has made building easier, but it hasn't made finding the right thing to build any less critical.
Industry Reality
What every startup founder believes about AI and PMF
Walk into any startup accelerator, scroll through Twitter, or attend a founder meetup, and you'll hear the same excitement: "AI changes everything about product development." The conventional wisdom has crystallized into a few core beliefs:
"Build fast, iterate faster" - Since AI can help you build MVPs in weeks instead of months, the thinking goes that you should ship quickly and let the market tell you what works.
"No-code + AI = instant validation" - Founders believe that because tools like Bubble, Framer, and various AI platforms let them build without technical constraints, they can test more ideas faster.
"The constraint isn't building anymore" - The popular narrative is that since building is now "solved," the only thing stopping success is having enough ideas to test.
"AI can analyze user feedback to find PMF" - Many founders think AI tools can process user interviews, survey data, and usage analytics to automatically identify product-market fit signals.
"Speed to market is the ultimate competitive advantage" - There's a belief that whoever ships fastest with AI wins, regardless of whether they've validated the core problem.
This conventional wisdom exists because it feels empowering. After decades of technical barriers preventing founders from testing ideas, AI genuinely has removed many constraints. The tools work. The speed is real. But here's what the industry gets wrong: they're optimizing for the wrong constraint.
The bottleneck was never really the building. It was always knowing what to build and for whom. AI made the easy part easier while leaving the hard part—finding genuine demand—exactly as difficult as before.
Consider me as your business complice.
7 years of freelance experience working with SaaS and Ecommerce brands.
The client came to me with a two-sided marketplace concept. They'd done their homework on the technical side—researched no-code platforms, explored AI integrations, even sketched out the user flows. What they didn't have was a single customer on either side of their marketplace.
"We want to see if our idea works," they told me. "With these new AI tools, we can build it cheap and fast, then iterate based on what users tell us."
This is where I stopped them. Not because their idea was bad—it actually had potential. But because their approach was backwards, and I've seen this movie too many times.
Six months earlier, I'd worked with a different client who took this exact approach. They spent three months building a beautiful AI-powered productivity app with every feature they could think of. The no-code tools worked perfectly. The AI integrations were smooth. The app looked professional.
Zero paying customers after launch.
The problem wasn't the execution—it was that they'd built a solution for a problem people didn't know they had, targeting users who were happy with their current tools. All the AI in the world couldn't fix the fundamental disconnect between what they built and what the market wanted.
But here's what really bothered me about the marketplace client: they were making the same mistake, but with higher stakes. A marketplace needs both suppliers and buyers to succeed. If you build it and neither side shows up, you've just created an expensive empty platform.
That's when I realized the real issue. AI and no-code tools haven't just made building easier—they've made the wrong kind of building more tempting. Because you can build faster and cheaper, there's less pressure to validate first. It feels low-risk to "just build it and see." But that's exactly backward.
Here's my playbook
What I ended up doing and the results.
Here's exactly what I told that marketplace client, and the framework I now use with every founder who comes to me excited about AI-powered MVPs:
The One-Day MVP Rule: If you're truly testing market demand, your MVP should take one day to build, not three months. Yes, even with AI and no-code tools, building a functional two-sided platform takes significant time. But your first MVP shouldn't be a product at all.
Instead, I walked them through what I call the "Manual-First PMF Framework"::
Day 1: Create a simple validation page
Not a product—just a landing page or Notion document explaining the value proposition. I've seen founders use everything from a simple Google Form to a Webflow page. The key is making it feel real enough that people would actually sign up if the product existed.
Week 1: Start manual outreach
This is where most founders want to skip ahead to building. Don't. Spend a full week reaching out to potential users on both sides of your marketplace. LinkedIn messages, Twitter DMs, industry forums—wherever your users already spend time.
Week 2-4: Manually match supply and demand
If people actually want what you're building, you should be able to facilitate connections manually. Use email, WhatsApp, phone calls, Zoom meetings—whatever it takes. This is your real MVP: you as the human middleware.
Month 2: Build automation only after proving demand
Only after you've manually facilitated successful transactions should you start building automated systems. By then, you'll know exactly what people need and how they want to use it.
The marketplace client pushed back hard on this approach. "But with AI, we could build the whole platform and test everything at once," they argued.
That's when I shared the story of a client who did exactly that—and what they learned from manually processing their first 100 matches. The problems users actually had were completely different from what they assumed. The features they thought were essential turned out to be annoying. The workflow they'd planned to automate needed three manual steps they hadn't anticipated.
The AI Integration Point: Here's where AI actually becomes powerful for PMF. Instead of using it to build your product, use it to analyze the manual process data. Feed your customer conversations into AI tools to identify patterns. Use AI to analyze which manual interventions worked best. Let AI help you spot signals in the noise of early customer feedback.
I've started using tools like Perplexity for rapid customer research and Claude for analyzing interview transcripts. The AI helps me identify market signals faster, but only after I've collected real data from manual processes.
Market Research
Use AI to analyze customer interview data and identify demand patterns, not to build unvalidated products
Manual MVP
Start with human-powered processes to prove demand before automating anything with AI
Signal Detection
AI excels at finding patterns in customer feedback once you have real conversations and data
Demand Validation
Build manual workflows first, then use AI to optimize what's already working
The marketplace client took my advice, reluctantly. Instead of spending three months building a platform, they spent three weeks manually connecting buyers and sellers in their industry.
The results were eye-opening. In those three weeks, they facilitated 12 successful matches manually. More importantly, they discovered that their original idea—a complex bidding system—was completely unnecessary. What people actually wanted was a simple way to find pre-vetted partners with transparent pricing.
By month two, they had a waitlist of 200+ users on both sides begging for an automated platform. That's when we started building—but building the right thing.
The contrast with the productivity app client was stark. The productivity app had beautiful AI features but no clear demand signal. The marketplace had ugly manual processes but overwhelming demand. Guess which one is still in business?
This pattern has held across every client I've worked with since. The ones who use AI to accelerate building before validating demand struggle to find customers. The ones who use AI to accelerate validation and analysis build products people actually want.
What I've learned and the mistakes I've made.
Sharing so you don't make them.
Manual processes reveal what AI can't: Customer behavior in the wild is messier than any user research suggests. Manual operations show you the real workflow, not the theoretical one.
AI amplifies existing demand, it doesn't create it: No amount of intelligent features can fix a product nobody wants. Use AI to serve validated demand, not to generate demand.
Speed is valuable, but direction is essential: AI lets you build the wrong thing faster. The real advantage comes from building the right thing at the right speed.
Customer interviews > user analytics: AI can process analytics beautifully, but it can't replace actual conversations with potential customers about their real problems.
Constraint breeds creativity: When building was hard, founders were forced to validate first. Now that building is easy, you need artificial constraints to maintain discipline.
The best AI use case for PMF is analysis, not creation: Let AI help you understand patterns in customer data, not build products before you have customer data.
Marketplace businesses require double validation: You need to prove demand on both sides before building. AI makes this easier to test, not easier to skip.
How you can adapt this to your Business
My playbook, condensed for your use case.
For your SaaS / Startup
For SaaS startups using AI tools for PMF:
Start with manual customer development, then automate
Use AI for customer interview analysis, not product building
Build waitlists before building features
Test demand manually first, scale with AI second
For your Ecommerce store
For ecommerce businesses leveraging AI for market validation:
Use AI to analyze customer feedback patterns, not inventory decisions
Test demand with pre-orders before building automated systems
Manual customer service reveals automation opportunities
AI should accelerate proven processes, not replace validation