JUst wanted to drop in and say hello

Greetings everyone. I am very excited to get in here and learn as much as I can.

I have a small software and AI integration business, less than a year old, that I started with a former software engineering student of mine. We started out building training apps for civil service tests at the NYC MTA, taking bus drivers from driving to passing the dispatcher exam. That involved a data tool I built that can take virtually any kind of data, whether a PDF, a database, or a scan of a page in a magazine. It feeds the data into a series of highly specialized, focused AI agents. Each one is trained to identify a certain pattern or separation point. After many, many hours of work and feeding it different scenarios, the AI became very good at recognizing even the most subtle and minute patterns. The result is a quiz, scenario, and study guide builder: you plug in whatever subject you want, and it builds customized, tailored study materials for whatever you are trying to study.

We then met our pilot client, whom I can finally name because it’s in the signed contract. I built an intake automation, SEO, and web hosting replacement for the setup they currently run through Scorpion Software. They signed a 60-day paid trial and are ready to sign a $45k deal for the rest of the year.

My cofounder, Haron Wilson, and I built FutureEng LLC from nothing. We paid for the LLC, the business liability insurance, and our LLM subscriptions, but that’s it. For seven months we had no pay, nothing, just 1,000-plus hours of grinding and learning how to build an intake automation system for a law firm that was safe, fast, and affordable, at a lower price than what they were already paying.

We built something with minimal configuration that can be installed in any business, company, or office that has client buildup online. Picture a client who emails an estate planning firm at 6 p.m. on a Thursday to set up a will and trust because his father is in the hospital. I conducted anonymous tests on about 100 law firms of our target size across the country. Only 8% got back to me in a decent amount of time, and even then it was over a day before I could get a consult scheduled. Crain and Wooley was not one of them. Any time, day or night, weekend or holiday, our system doesn’t funnel people into a queue. It walks them all the way through to paying the fee and scheduling the consult.

We just went live, so we don’t have many actual numbers yet, but we hope to see virtually zero missed opportunities and no clients lost in the shuffle.

At FutureEng, we ship fast, we ship right, and we ship once. Safety in AI is a number one priority, not an option, ever. If DeWalt can’t release a table saw without an emergency off button, why are so many developers willing to vibe-code and put out just about anything with either no way to interrupt and control the system, or one that takes too much involvement or time to activate? There must be a human in the loop to evaluate anything an AI generates before it ever reaches a decision-making point. An off switch is nice, but kill switches are phenomenal. We need to be able to instantly resume control over the system we built, since we really have no idea what it does inside or why. We only know that data goes in and we get close to what we want out. Sometimes.