5 Things I'd Do Differently If I Started My SaaS Again

After two years of building B2B products and navigating strategic pivots, I've collected enough battle scars to fill a book. These are the five business development lessons I wish someone had told me before I started.

The Hard Lessons

Two years ago I started building a voice AI product. Nobody asked for it. No market research, no customer interviews, no waitlist. Just a gut feeling that typing was dying and voice was the future.

I was right about the direction. Dead wrong about everything else.

1. I'd validate with money, not compliments

Everyone told me the idea was great. Friends, family, LinkedIn connections. 'Oh that's cool!' is not validation. Someone handing you £5 is validation. I wasted 4 months building features nobody wanted because I confused encouragement with demand.

Now I follow one rule: if people won't pay before it exists, they won't pay after.

2. I'd launch ugly and fast

My first version was embarrassingly polished. Took forever. Meanwhile, competitors shipped ugly MVPs and got real feedback. By the time I launched, they'd iterated three times.

Ship something that makes you cringe. If users still love it, you've got something.

3. I'd focus on one use case, not twenty

Genie 007 started trying to do everything — emails, browsing, documents, social media, all by voice. Turns out nobody wants a tool that does 20 things at 60%. They want one that does 1 thing at 100%.

We eventually found our niche: voice-to-action on any website. That focus changed everything.

4. I'd charge from day one

Free users give terrible feedback. They'll tolerate bugs, never report issues, and vanish quietly. Paying users tell you exactly what's broken because they expect it to work. That feedback is gold.

5. I'd build distribution before product

The biggest mistake founders make: building in a vacuum. I should have been writing about voice AI, engaging in communities, building an audience 6 months before launch. By the time we shipped, we had zero distribution.

Now I'm fixing that. Building in public, engaging genuinely, letting people watch the journey. It's harder than coding, honestly.

The Bottom Line

Starting a SaaS is like learning to drive. You think you know how until you're actually behind the wheel. These five mistakes cost me a year and thousands of hours. If sharing them saves you even one month, this article was worth writing.

What would YOU do differently? I'm genuinely curious — the best lessons come from other founders' scars.

— Bill Kiani

I built Genie 007 — a voice AI app for any website, 140+ languages, £40 one-time.

Comments

Post a Comment

Popular posts from this blog

Your Pricing IS Your Marketing — A Founder Lesson

I learned something counterintuitive about pricing: it speaks louder than any ad campaign. The £699 Problem Dragon NaturallySpeaking charged £699/year for voice dictation. For 20 years, that told the market: voice AI is premium. Enterprise only. Not for regular people. Then they got discontinued. Millions of users — especially people with RSI and accessibility needs — were stranded. The £40 Statement When we priced Genie 007 at £40 one-time, it was not just a business decision. It was a statement: accessibility should not be a luxury. Voice AI should be for everyone. That price communicates more than any marketing copy: we are for real people, not enterprise budgets. What Founders Get Wrong Most founders price based on what competitors charge. That is backwards. Price based on who you want to serve. If you want to serve students, freelancers, and people with disabilities — price for them. The volume play wins long-term. 1,000 customers at £40 beats 50 customers at £699 — and those 1...
Read more

Why I Bet Everything on Voice AI — And Why You Should Pay Attention

There is a moment every founder remembers — the moment you stop hedging your bets and go all in. For me, that moment came with voice AI. The Gap Nobody Was Filling For over two decades, Dragon NaturallySpeaking dominated voice dictation at £699/year. It was enterprise-grade, expensive, and inaccessible to most people. When Nuance discontinued the consumer version, millions of users — especially those with RSI, disabilities, and accessibility needs — were left stranded. I saw that gap and I could not look away. Building Genie 007 I built Genie 007 as a solo founder with zero funding. No team, no investors, no connections in tech. Just a clear problem and the stubbornness to solve it. Genie 007 is not just voice dictation — it is voice-to-action. You speak, and it acts. On any website, in any app, across Windows, Mac, and Chrome. It works in 140+ languages. And it costs £40 one-time — not £699/year. Why Voice AI Is the Future We are at an inflection point. AI is moving from text-...
Read more