AI that ships into your production stack.
For fintech and B2B SaaS companies under SOC 2, PCI, HIPAA, or GDPR. Discovery, build, handoff. Audit-ready documentation as a deliverable, not a follow-up.
See the offerWhat I'm building.
Every product I'm currently working on, with status and context. Real problems. Built in public. Updated as things ship or change.
Advisory
Sound AI Advisors.
Production AI for fintech and B2B SaaS. The consulting side of the practice, separate from the indie SaaS. 4-8 weeks. $25K-$50K. Built on 20 years of fintech security experience.
soundaiadvisors.com
The dedicated home for the advisory practice. Discovery call booking, sample workflows, the AI Readiness Assessment, and early access to the AI Agent Implementation Playbook.
Visit the site
Indie SaaS
What's in the indie lineup.
Each product lives as its own page so the purpose, status, and related writing stay in one place. Progress gets documented in public, including what is working, what is stuck, and what is coming next, so anyone following along can trace how an idea actually moves.
ReplyWisely
A Chrome extension that helps small accounts grow on X by scoring replies for visibility potential and tracking what's actually working.
Explore product pageTwitMix
Turn bookmarked posts into fresh, original content. Saved inspiration becomes something you actually publish.
Explore product pageIkigaiNiche
Discover aligned business opportunities at the intersection of your strengths, interests, and real market demand.
Explore product pageFinSec Scorecard
A diagnostic for small fintech companies assessing PCI DSS and SOC 2 readiness of their AWS infrastructure.
Explore product pageDomainNameNow
Find a brand-ready, actually-available domain name in minutes. Built for founders and indie builders who refuse to let naming delay a launch.
Explore product page
What didn't work
Killed products. Lessons kept.
Build in public means showing the cuts, not just the launches. Each card here is a real product I shipped or built and then shut down, plus one idea I never built. The postmortems explain why and what's worth keeping for the next builder.
CampfireX
Private circles for business owners. 8 days, 200 clicks, zero customers. Network-effects math doesn't work at a 1,200-follower audience.
Read postmortemCompliQuiz
A framework-discovery quiz for small businesses. Killed because awareness problems get solved for free. Execution is where founders actually pay.
Read postmortemHardChoice
A structured decision tool for couples. Real product, wrong audience for this brand. I wasn't willing to run a parallel B2C consumer brand for one product.
Read postmortemDearMeWiser
A reflection app: ask a question, get an answer from your wiser future self. Killed because a single ChatGPT prompt did the same job. If a prompt replaces your product, you don't have one.
Read postmortemNeverMissAgain AI
An idea: AI text-back for missed calls at small businesses. Written down, never shipped. The lesson: ideas you don't build are usually ideas you don't believe in enough to grind through. Honesty about that beats a half-built product page.
Read the build log
Build timeline
Order of operations. What shipped, what got cut.
Every product, in the order I worked on it, with one honest line on what happened. Useful if you're a builder trying to read the shape of a portfolio rather than just its winners.
-
Mar 2026
First batch shipped: ReplyWisely, TwitMix, IkigaiNiche, FinSec Scorecard, CompliQuiz
Five products in one wave to test which problems people actually wanted solved. Four survived. CompliQuiz did not — the quiz answered an awareness question that founders solve for free.
Lesson: ship in batches when you don't know which problem is real. The market votes with attention faster than any pre-launch survey will tell you.
-
Mar – Apr 2026
HardChoice and DearMeWiser: built, shipped, cut
HardChoice was a real product for couples. DearMeWiser was a reflection-app learning project. Both came off the lineup in April. HardChoice was killed because it was the wrong audience for the brand. DearMeWiser was killed because a single ChatGPT prompt replaced it.
Lesson: niche discipline is what you take off the page. A wrapper around an LLM is not a product. Both rules are easy to read and hard to apply when it's your work.
-
Apr 2026
DomainNameNow shipped
Built to compress naming and domain search into one working session, because slow domain hunts kill indie launches that were otherwise ready.
Lesson: the products that earn their place are the ones that remove a small, sticky step every builder hits. Boring is fine. Useful is the bar.
-
May 2026
CampfireX: 8 days from idea to kill
Built around private circles. Ran the indie-hacker playbook end to end: features, landing page, demo videos, pre-pay validation. 200 clicks, zero customers, killed on day eight. Network products need a busy room before any single visitor stays. At 1,200 followers, the math doesn't work yet.
Lesson: run the audience math before the first commit. If "build it and they will come" is your distribution plan for a network product, you do not have one.
-
May 2026
CompliQuiz killed; pivot toward execution-layer compliance
The CompliQuiz domain was redirected here. The product page came down. The new bet is the harder, more valuable problem: helping small fintech companies actually achieve compliance, via FinSec Scorecard and a fintech compliance handbook in progress.
Lesson: awareness problems are usually solved for free. Execution problems get paid. If a smart user can answer your tool's question with twenty minutes and a search bar, you have a quiz, not a product.
-
May 2026
Sound AI Advisors launched
The consulting side of the practice. After CampfireX, the question business owners kept asking was not about communities. It was: how do we put AI into production without breaking what already works? Sound AI Advisors is the answer, scoped to fintech and B2B SaaS. 4-8 week embedded engagements, audit-ready by the time we hand it off.
Lesson: listen to the question that keeps coming up, not the one you were expecting. Twenty years of experience is leverage when you point it at a problem people are already paying to solve.
-
Idea only
NeverMissAgain AI: written down, never built
An AI text-back tool for missed calls at small businesses. The idea has been sitting on the list. It has not been shipped. Keeping the idea visible without pretending it's a product is the more useful entry on this timeline.
Lesson: the gap between idea and ship is usually conviction, not capacity. Ideas you don't build are usually ideas you don't believe in enough to grind through. An honest "not yet" beats a misleading "shipping."
Follow the build
New products ship here first.
The journal captures what I'm learning along the way. X is where launches, small wins, and mid-build updates land.