After two weeks on the road, I came home knowing I needed to write this down.
Friends had been asking what I planned to do next. For weeks, I had been asking myself the same question.
Crossing the AI Threshold
The urge—and the itch—to build something in AI began early this year and kept getting stronger.
Tools such as OpenClaw began attracting attention in January. After trying them myself, I started pushing the people around me to do the same. I brought up AI in meeting after meeting, collected tools and learning resources, and encouraged colleagues to experiment rather than wait for someone else to explain everything first.
I also picked up the nickname “Big Bro JZ”—courtesy of an OpenClaw agent.
My motivation was simple, perhaps even a little blunt: I wanted the people around me to remain relevant—and able to earn a living—in the AI era.
After what probably felt like a borderline evangelical campaign, something began to change. Colleagues who had initially watched with uncertainty started sharing their own breakthroughs with me. They built product prototypes with AI. Some of those prototypes even began to look like the foundations of real businesses.
They had crossed the line from watching AI to using it. More importantly, they had done it themselves.
That experience stayed with me. Change does not always begin with a sophisticated technical breakthrough. Sometimes it begins when starting becomes easier than waiting.
Where SMBs Get Stuck
Later, as my time with my former employer was drawing to a close, I finally had the space to think more seriously about what I might build in AI.
One question kept returning: if individuals need a path and a first win to cross the AI threshold, what would it take for businesses—especially SMBs—to do the same?
How do they move beyond awareness, close the gap between AI experimentation and business adoption, and find where AI can create real value?
Many SMBs lack the knowledge, experience, and resources to answer those questions with confidence. They know AI matters. But every week brings another wave of models, concepts, and tools. They are unsure where to begin, which use case to choose, which technology to trust, or how to tell whether an experiment has made any meaningful difference to the business.
A common mistake is to pick up an AI hammer first and then go looking for a nail—choosing a model or system first, then searching for a problem it might solve.
The better starting point is the business as it already exists—its opportunities, pain points, constraints, and risks.
What is worth doing? Why is it worth doing? What should be different if it works? Only then should we decide how to build it.
If a use case has no clearly defined problem, owner, operating context, or measure of success, a faster build will not make it part of everyday work.
How FFDE.ai Began
Around the same time, Forward Deployed Engineering (FDE) was moving back into the AI conversation. It gave me an important starting point.
A Forward Deployed Engineer works inside a client’s real operating environment, defines the problem with users, builds the system, and takes a capability to an operable, testable, and handoff-ready state.
What makes FDE valuable is that it does not separate business problem-solving from engineering execution. Someone has to enter the real workflow, understand what is actually happening, and stay with the work until the result is usable.
But for many SMBs, engineering delivery is only part of the gap. They may also need help clarifying the business problem, choosing the right opportunity, understanding the risks, and connecting strategic intent with technical execution. Without those pieces, a company can remain stuck between “What should we do?” and “How do we do it?”
That led me to FFDE.ai: Fortified Forward Deployed Engineering.
“Fortified” is not decoration. It adapts the FDE model for SMBs by strengthening the parts that often determine whether a pilot survives: use-case selection, data and knowledge, workflow design, governance, training, and handoff.
In plain terms: decide what deserves to be built. Then build it so the client team can run it, evaluate it, and keep improving it.
Success is not a company that keeps calling outside experts. It is a client team that can identify the next opportunity, learn a new method, apply AI to more of its work, and continue improving after the external team has stepped away.
That is the organizational muscle I care about: the ability to keep choosing, learning, and applying AI. I have a simpler name for it—AI freedom.
For an SMB, that matters far more than another impressive proof of concept.
Time to Build
Over the past few months, I have met many people who are also trying to find their own way to ride the AI wave. Those conversations have made the scale of the change—and the opportunity—much clearer to me.
At the same time, I have been pressure-testing these ideas through the work itself: securing the domain names, building the website, and completing the necessary compliance work in China. I have also been independently building and testing knowledge-base applications, enterprise applications, LLM fine-tuning tools, content-production skills, AI-enabled social media operations, and workflow platforms.
The work has not been neat. Early results can be unstable. Tools change. Assumptions break. But that is exactly the point. You understand AI by putting it to work on a real problem, getting your hands dirty, and staying with it until something useful holds up in practice.
I hope FFDE.ai becomes a platform for sharing, learning, and collaboration—a place where people bring their perspectives, examine real problems together, turn ideas into small experiments, and use the results to decide what should happen next.
I am also exploring a separate, consumer-facing AI direction. That is another story for another day.
I do not know exactly where this path will lead. But I know what comes next: keep learning, keep building, and stay close to the real work.
Play.
Stay curious.
Let’s build something together.