Every Business Needs an AI Pirate
I've spent more than thirty years running businesses, and even a modest company contains an astonishing amount of knowledge, inefficiency, opportunity, and outright nonsense. Much of it is invisible to outsiders. Some of it is invisible to the owner.
Consider what might happen if a technically capable, relentlessly curious person spent a few weeks investigating one of these businesses, equipped with the latest AI tools and encouraged to question practically everything.
I think we're going to need a lot of these people. AI pirates.
There are millions of small and midsize businesses filled with experienced employees, accumulated knowledge, peculiar routines, outdated systems, and problems that nobody has had the time or resources to solve. AI is making it possible to accomplish things that would have been prohibitively expensive just a few years ago. Somebody has to discover what those things are.
The pirate's first job is to get curious.
Follow people around. Watch how things are done. Ask why. Look for the spreadsheets that run entire departments, the phone calls that hold complicated operations together, and the experienced employees everyone turns to when something goes wrong. Find the decisions that depend on memory, intuition, or information scattered across five different systems.
Pay particular attention to what people know but have never written down. An experienced mechanic, nurse, estimator, mason, or logistics manager may carry decades of judgment that the company's software knows nothing about. Capturing that expertise could be more valuable than automating a hundred routine tasks.
Some of this work resembles investigative journalism. Some resembles anthropology, industrial engineering, or old-fashioned management consulting. The pirate needs to listen carefully, earn trust, recognize patterns, and distinguish genuine opportunities from interesting distractions.
And then comes the fun part.
Start building things. Connect a few systems. Give employees better access to the information they need. Create an agent that investigates missing materials, prepares a schedule, follows up on unresolved problems, or helps someone make a difficult decision. Test it with the people who actually do the work. Learn what breaks and try again.
Before long, the problems become more sophisticated. The AI needs to understand what the organization is trying to accomplish, which information matters, who is authorized to do what, and when to involve a human. Its decisions need to be tested, its failures understood, and the company's proprietary knowledge protected.
An entire collection of emerging specialties is taking shape around this work: knowledge mining, intent mapping, context engineering, agent development, forward-deployed engineering, evaluation engineering, and others that barely have names.
I expect demand for these capabilities to grow enormously. Some will become established professions. Others will be combined into jobs that don't resemble anything on today's organizational charts.
And the qualifications are fascinating.
Computer science provides an excellent foundation. Understanding programming, databases, software architecture, security, and how complicated systems behave gives a person a real advantage, even as AI takes over more of the coding.
But technical ability is only part of the pirate's equipment.
The person who can interview a veteran employee and uncover knowledge that has never been documented is doing something closer to anthropology than programming. Understanding why people resist change requires psychology. Identifying valuable opportunities involves economics. Translating human intentions into instructions for machines raises questions of philosophy, language, and judgment.
Writing, observation, skepticism, empathy, imagination, and the ability to ask uncomfortable questions all matter. So does domain expertise. Combine software fluency with an understanding of construction, healthcare, manufacturing, agriculture, finance, or logistics, and the possibilities multiply. A computer science graduate who understands building restoration might recognize opportunities that an accomplished AI researcher would walk right past.
I suspect some of the best AI pirates will have backgrounds that never fit comfortably into conventional job descriptions. They may be engineers, teachers, tradespeople, historians, entrepreneurs, or people who have spent years bouncing between interests that suddenly turn out to be extraordinarily complementary.
The opportunities for smaller businesses are particularly interesting. Large corporations can afford AI research departments, specialized consultants, and dedicated engineering teams. A forty-person company may have equally complicated problems and equally valuable expertise, but no practical way to investigate what AI could do for it.
That creates an opening for people who can become genuinely useful without requiring a million-dollar budget or a two-year transformation program.
And here's some advice for the business owners. You may soon need an AI pirate as badly as you need an accountant, a salesperson, or someone keeping the lights on. Somewhere, a competitor is going to figure out how to capture expertise, eliminate expensive routines, improve decisions, and operate with advantages you don't have. That competitor might be smaller, younger, and considerably more aggressive. Find yourself an AI pirate before a competitor's pirate sinks your ship.
I'm involved with one effort in this direction, Building Value Preservation, where we're exploring how AI, software, and accumulated operational expertise can improve the maintenance, repair, and preservation of buildings. The work brings together construction knowledge, logistics, finance, human judgment, and technical engineering. It's exactly the kind of environment where people with unusual combinations of skills could accomplish interesting things.
But the larger opportunity extends across almost every industry. Wherever people are making decisions, moving information, coordinating work, or relying on expertise that exists primarily in someone's head, there's territory worth investigating.
I can imagine young people building careers around individual businesses, entire industries, or collections of clients. Some will join established companies. Others will start their own ventures. A few may discover opportunities that eventually become substantial businesses in their own right.
There's also an appealing possibility for people who already know how to code but are having difficulty finding conventional software jobs. Their education may have prepared them for a much wider range of opportunities than the current job market suggests.
Becoming an AI pirate requires more than becoming good at prompting. It means getting out into the world, understanding how things work, earning people's confidence, and taking responsibility for results. The tools are increasingly accessible, and the opportunities extend into practically every corner of the economy.
There are businesses out there with problems you've never imagined, knowledge they've never captured, and opportunities they don't yet know exist. Someone is going to help them figure it out. It might as well be you.
AI disclosure: I developed the ideas, observations, and direction of this essay through conversations with ChatGPT, which assisted with drafting and editing. The opinions and conclusions are my own.