Business / 2026
Mainstay
Good businesses shouldn’t lose customers to bad digital experiences.
Why I built it
Mainstay began with a practical observation: many strong local service businesses lose demand at visible, fixable points—often before a real conversation ever begins.
The product
Making the premise tangible.
The starting offer focuses on rescuing weak websites, with a broader view of the customer journey: lead response, missed-call recovery, follow-up, review systems, reactivation and the infrastructure around growth.
How it evolved
The thinking behind the build.
Narrow the capability
‘AI-powered growth’ was too broad to be useful. The offer became concrete only when we focused on a visible problem: good local service businesses losing demand through weak websites and slow response systems.
Choose a credible wedge
Website Rescue became the first offer because the problem is public, diagnosable and finite. That gave the broader customer-journey thesis a clear place to start without creating a menu of vague service tiers.
Build Product Zero first
Mainstay’s own site and operating system are the first proof of quality. Research and drafting can be accelerated by AI, while outreach and other reputational actions remain supervised and approval-gated.
The product / In context
Key decisions
Where product judgement showed up.
- 01Lead with customer outcomes rather than AI
- 02Start where problems are public and diagnosable
- 03Keep the initial offer simple and understandable
- 04Separate the commercial brand from my personal site
My role / AI’s role
I founded and shaped the offer, positioning and operating boundaries. The experiment also tests supervised autonomy: AI can research, audit, build and draft, while consequential external actions remain approval-gated.
Iteration
The first build was the beginning.
Mainstay is being built as a real operating experiment. Its first proof point is its own digital experience; external prospecting follows only after the offer and execution standard are credible.
What I learned
- A capability is not a business until it solves a problem someone will pay to remove.
- Autonomy needs explicit boundaries, especially when reputation is involved.
- A narrow wedge can support a broader operating thesis.