Professional Knowledge Mining

The experiment you should run before you graduate

Spend Enough Time Beside the Work

Spend an afternoon with a good mason and you can learn how to lay a brick. Spend a couple of summers beside one and you may begin to see the part that is much harder to put on YouTube.

You notice how the work is sequenced, how materials are staged, how the crew keeps moving, how weather changes the day, how unfinished work is protected, how water is managed behind the finished wall, how body position affects pace, and when the normal routine gets abandoned because something on the job changed.

None of those things alone is the secret. The value is in how they fit together.

What does that mason have in common with a pipe-organ tuner, HVAC technician, antique-clock restorer, heavy-equipment mechanic, stained-glass conservator, or industrial-controls specialist?

They all carry around bodies of working knowledge that are much larger than a set of instructions. Some of it is judgment. Some is timing. Some is physical. Some is logistical. Some only becomes visible after you have worked beside the person through enough ordinary days, bad days, changing conditions, mistakes, and unusual jobs.

Other people would pay to have that knowledge, but surprisingly little of it has been captured well.

I think there may be a profession hiding inside that problem.

Call it Professional Knowledge Mining.

Run the Experiment

If I were in university today, I would seriously consider spending the next two or three summers testing this.

Pick a field you barely understand. Find one or two people who are unusually good at it. Then spend roughly 18 months trying to answer one question:

Can I capture enough of what they know to make someone else measurably better at the work?

Eighteen months may sound slow if you compare it with building an app over a weekend. But if you are trying to build an asset from knowledge that took somebody thirty years to acquire, 18 months starts to sound remarkably fast.

It also creates part of the advantage. Most people will not do this.

Your first attempts will be clumsy. You will record the wrong things, ask weak questions, miss important moments, and organize material in ways you later abandon. By the fiftieth case, you should be noticing things you did not even know to look for in the first five.

That is the experiment working.

The First Problem Is Human

Your first hard problem probably will not be AI. It will be convincing a busy expert that having a college kid follow them around asking “why?” is worth the trouble.

You have to earn your way in.

Maybe you pay them. Maybe you become genuinely useful around the work. Maybe you organize years of old records or create training for younger employees. Maybe the expert cares about preserving the craft. If something commercial emerges, perhaps they participate in the upside.

The important thing is that they win too.

Then get close enough to see what actually happens. Wear smart glasses or another small camera. Carry a phone, microphone, and whatever inexpensive instruments fit the field. Watch the workflow, the routines, the starts and stops, the staging, the handoffs, the physical movements, and the moments when the plan changes.

Ask what the expert noticed, what they think is happening, how sure they are, and what would change their mind.

One question is especially useful:

What did you notice that I didn't?

Keep bringing the material back and let the expert correct you. A particularly good sign is when, after a while, they call you and say, “You need to come see this one.”

Now you have a collaborator.

Build Something More Useful Than a Video Library

The internet is already full of how-to videos. Another pile of recordings is not the opportunity.

Turn what you capture into something that can be retrieved, compared, tested, and reused. Preserve not only the interesting decisions, but the operating patterns around them: what happens first, what gets staged in advance, what changes with weather or workload, how the expert moves, where delays occur, how the crew coordinates, and what happens when the normal routine breaks.

Connect decisions to outcomes.

“The expert thought X” is interesting. “The expert thought X, acted on it, and Y happened” begins to look like an asset.

You can start with organized files, transcripts, images, notes, and a simple database. AI can help enormously with transcription, tagging, comparison, retrieval, and pattern finding.

But do not accidentally give away the thing you are trying to create. Some of this material may contain trade secrets, customer information, proprietary methods, copyrighted material, or rare knowledge whose value depends partly on staying private. Agree early on who owns what, and be careful about loading the crown jewels into AI systems without understanding how your data is stored, used, and protected.

Protect the knowledge layer before building the intelligence layer.

Find the Expensive Decision

Eventually, ask:

Who would save time, money, or mistakes if they could use this knowledge?

The buyer may already be standing beside your expert. An employer may need to train junior people faster before experienced workers retire. A manufacturer may want fewer failures in the field. A service-company owner may want to scale without diluting quality. An insurer may care about avoiding expensive mistakes.

Or perhaps you become the buyer yourself by using the knowledge to operate a business better than competitors.

The first product does not need to be an AI company. It might be training, diagnosis, inspection, remote expertise, or simply a better service business.

Try to sell something early. People are generous with compliments. Asking them to pay $500 produces better information.

Then You Discover the Second Asset

Suppose the experiment works.

The obvious asset is the expertise you captured.

But something else happened. You learned how to find a promising niche, identify the right practitioner, earn trust, recognize what matters, capture the work, protect the material, turn messy experience into something usable, and discover what someone will pay for.

You built the machine that builds the asset.

That is your knowledge-mining machinery.

Now try it somewhere else.

What you learned beside a mason might help you investigate agricultural equipment. A method developed around refrigeration might transfer to scientific instruments. Your second experiment should be faster than the first, and the third should improve again.

At that point, you may have found more than an interesting niche. You may have found a profession.

Now Jump Ahead Ten Years

Imagine yourself high inside an old clocktower with an experienced technician and two affordable robots.

One robot is inspecting. The other is positioning tools and components. The machines themselves are commercially available. Your competitors can buy the same ones.

What they cannot simply order is everything your team has accumulated over the years: the strange failures, expert corrections, sounds, measurements, physical routines, bad repairs, edge cases, decisions, and outcomes.

The robots are not the advantage.

Knowing what they should know is the advantage.

And because you learned how to mine professional knowledge rather than merely collect clock knowledge, you know how to find the next body of expertise they will need.

The Time Is Now

Experienced practitioners are retiring. Wearables can capture what people see and hear. Sensors are cheap. AI can work across images, audio, video, documents, and measurements. Robotics is improving.

Meanwhile, enormous amounts of valuable working knowledge remain inside individual people and small groups, largely untouched by all of it.

If you are in university today, you do not need to know what company you want to build ten years from now.

Find something worth knowing. Find someone who knows it unusually well. Earn the relationship. Work beside them. Capture what everyone else misses. Protect it. Make it useful. Try to sell something. Then do it again.

After 18 months, you may discover the niche was a dead end.

Or you may discover that you have found a profession nobody told you existed, created an asset you never knew you could own, and learned how to build the next one.

The companion field guide goes much deeper into choosing a field, working with experts, capture methods, workflow and operating-pattern capture, intellectual-property protection, using AI carefully, testing whether knowledge transfers, and finding businesses that can grow from it.


Written with help from AI, shaped through iteration, and pushed toward clarity by a human who did not want the machine to settle for the obvious.

Previous
Previous

Every Business Needs an AI Pirate

Next
Next

Doomerism for Dummies