Sun Dahan

Chapter 12 of 13

A precise dark mechanical engine with gears and shafts in gold edge-light

Findable

How Finding People Works Today

Right now, finding the right person is mostly a person with a screen. A recruiter opens LinkedIn, searches by title and keywords, and opens profile after profile. When a role is posted, the resumes land in an applicant tracking system, and someone works through the pile: reading, comparing one against another, ranking, sorting into yes, maybe and no. Sometimes the software scores or filters first. Whether people came in through a posting or were found by search, the work is the same. Read a lot of documents quickly, judge them against a brief, build a shortlist.

Will it stay like this? I don't think so. Nothing in that job is out of reach of an AI agent. Reading a profile, comparing it to a brief, ranking, shortlisting: I believe an agent can do all of it today, if it has three things. Access to the people, a database or a set of places to search, and instructions for what a good match looks like. The bottleneck is access and data.

Once agents run the search, the order of hiring can change. The agent looks first, and the posting opens when it comes back empty, as the fallback for the roles it couldn't fill. A job seeker who waits for a posting is waiting for the second step of a process that started without them. This is early, and I can't tell you the year it reaches your industry. The direction is clear enough to plan for.

An Agent Can Only Find What Exists

Today an agent finds people where people already leave a trail: LinkedIn, GitHub, Stack Overflow, personal sites, conference pages. The important word is published. An agent can't read your judgment or your record if it lives in a drawer, in a private folder, or in the heads of three former managers.

Developers are the easy case, because their work is public by default. A salesperson, an operations lead or a finance manager leaves almost nothing an agent can read, apart from one document on a hard drive and one profile. The people who get found first are the ones who already published. Everyone else is invisible, however good they are.

This is the Chapter 2 problem showing up earlier. The page already throws most of a person away, and now the search can't even reach the page.

"A job seeker who waits for a posting is waiting for the second step of a process that started without them."

"LinkedIn Already Does This"

The usual answer is that LinkedIn already does this. It is the largest professional database in the world, and every recruiter has a search box into it. If it solved findability, sourcing would be easy, and Chapter 6 described a recruiter's job that is still hard.

A LinkedIn profile is a list of self-described titles and keywords, written by people who know the keywords matter, so profiles for the same role look the same. That is the flattening from Chapter 4, in a different template. It also lives inside LinkedIn. What an outside agent can see of it is what LinkedIn lets it see, and there is no place on it to publish your full resume where the open web can search it. The closed matching databases that keep launching have the same limit: you can be inside one and still invisible to everything outside it. A profile held in someone else's system, in someone else's format, is findable only on their terms.

What Has to Exist

This is where I think it's going, and what I think has to exist. It doesn't depend on who builds it.

It has to be open. Each person's record is public and readable by anything on the web. An agent finds it through the search it already uses, not through a private connection it would have to be persuaded to build. A registry that agents have to be convinced to query starts with nobody in it and no reason to join.

It has to be plain. It doesn't need to be a beautiful site. It's a record: your name, your resume, your links, and the structured fields machines read.

It has to be guided. Software tells you what's missing and asks the questions a resume never answers, so the record ends up richer than the document you started with. Most people don't know what an agent would look for, so the tool tells them.

And it has to stay yours to release. The record is public, your contact details aren't, until someone reaches out through it and you say yes. You see your photo and a progress bar, and you hear nothing until someone wants to talk to you for a real reason.

Before Anyone Builds It

You don't have to wait for the tool. Most of this you can do with what exists.

Start by owning your name. Buy the domain with your name on it. It is the one address on the web nobody else can take from you.

Publish your resume as a page, not a file. A PDF behind a button is hard for a machine to read, and the same content as a plain page is easy.

Say the same thing everywhere. Use the same name, titles and dates on your site, your LinkedIn and anywhere else you appear, so an agent that finds two copies sees one person.

Keep the page open and plain. Text that loads without a login, a script wall or a block for crawlers is what a machine can actually reach. Structured data such as schema.org's Person type is cheap to add and fine as housekeeping, but I wouldn't count on it to get you found.

Write down what you did and what happened. The Impact Page from Chapter 13 gives an agent something to find that a bullet point never could.

Then check what comes back. Ask an AI assistant who you are and read the answer. What it says, and what it misses, is a rough picture of what a sourcing agent sees today.

That is the first reason to buy the domain before you need it. Chapter 13 covers the rest.

What does it mean to be findable to an AI agent?

That a software agent searching the open web can locate you, read what it finds, and understand it without guessing. In practice it means having a public page about you, stating the same facts everywhere you appear, and labelling them in a format machines read.

Can AI agents really screen and shortlist people today?

In my view the reading, comparing and ranking is within reach now. What limits them is access to the people and the data, which is why where you have published matters.

Isn't LinkedIn enough?

It helps, and you should keep it. But a LinkedIn profile is the format LinkedIn chose, held in a place LinkedIn controls, and profiles for the same role tend to look alike. It is one place you can be found, and it isn't the only one that will matter.

Does this mean job postings are going away?

No. In this picture the posting becomes the fallback: it opens when a search comes back without the right person. And the resume is still required (Chapter 3). The change is where the process starts, not whether postings exist.

Should I put my resume online for anyone to read?

Put online what you're comfortable having anyone read, and keep your phone number and home address out of it. The goal is to be found, not to be contactable by everyone. Add contact details in a form you control.