How to Get Found by Recruiters — and by AI — in 2026

Two things have to be true for someone to find you in 2026. First, when a recruiter or a potential client types your name, or "product designer in Lisbon," into Google or LinkedIn, a clear and current page about you should come back. Second — and this is newer — when someone asks ChatGPT, Perplexity, or Google's AI features to "suggest a few good React developers who write about accessibility," the model should be able to read a page that plainly says that is you.

Both of those depend on the same handful of things, and none of them are tricks. You become findable by being the clearest, most current, most consistent answer to "who is a good [your role]." A public page that states plainly who you are, what you do, where you are, and your best work. The same name and role across every place you show up. Real, machine-readable structure underneath. That is the whole game, and it is mostly within your control.

The uncomfortable part: you cannot "game" an AI assistant the way people once gamed search rankings, and anyone selling you guaranteed AI citations is selling something. There is no switch. But there is a durable position, and it is the one honest move available — actually be the best-described, most-corroborated person for the thing you want to be found for.

The two ways people find you now

It helps to be precise about the two paths, because they reward slightly different things.

A human searching. A recruiter sourcing candidates, or a hiring manager doing a background check before a call, or a client vetting a freelancer. They search your name, or a role plus a city, or a role plus a skill. They are scanning results to decide whether to click. What wins here is a result that is unmistakably you, recent, and specific — a page that loads fast and answers "is this the right person?" in the first screen.

A person asking an AI assistant. This is the part that changed. People now ask AI tools open questions: "who are some designers who've worked on fintech onboarding?" or "find me a technical writer who's written about API docs." The assistant reads pages from the open web, synthesizes, and names a few people. To be one of those names, a model has to be able to (a) reach a page about you, (b) parse what you do from it, and (c) be reasonably confident that the page, your LinkedIn, and your GitHub all describe the same person.

Notice the overlap. A clear, current, public, well-structured page serves both. You are not optimizing two different things. You are building one good source of truth and making sure both humans and machines can read it.

What actually makes you findable

Here is the honest list, in roughly the order that matters.

A clear public page that states the basics plainly

This sounds obvious and it is the thing most people skip. Your page should say, in plain text a machine can read (not buried in an image or a PDF):

• Your name.
• Your role — "Senior product designer," not "crafting delightful experiences."
• Your location, or at least your region and whether you're remote.
• Two or three pieces of your best work, with enough context to understand what you actually did.

A model summarizing you can only repeat what your page says clearly. If your homepage is a hero image and a vibe, there is nothing to extract. If it says "Product designer in Berlin, currently leading the checkout redesign at [Company]," that is a sentence an assistant can lift almost verbatim. Write for the reader first — but the sentences that help a human decide are the same sentences a model can quote.

This is also why a real page beats a résumé PDF for discovery. A PDF in an email is invisible to all of this. See how to turn your resume into a website for turning the document into a page.

Consistency across the places you appear

This is the one most people underrate, and it genuinely matters. Search engines and AI tools both try to resolve the scattered mentions of you into a single entity — one real person. They do that by corroboration: your site says you're a product designer in Berlin, your LinkedIn says product designer in Berlin, your Dribbble bio matches, your GitHub profile points to the same site. When those agree, a tool can be confident it's all you, and it will treat your page as authoritative about you.

When they disagree — different name spellings, a job title from two roles ago on LinkedIn, a personal site that 404s — the signal fragments. Nobody penalizes you exactly, but the confidence drops, and a model is less likely to name someone it can't pin down.

So the cheapest, highest-leverage thing you can do this week:

• Use the same name everywhere (pick "Sam Rivera" or "Samuel Rivera" and commit).
• Use a consistent current role across your site, LinkedIn, and any portfolio.
• Link them to each other. Your site links to your LinkedIn and GitHub; those link back to your site.

That mutual linking is what lets the connection be made. It's the difference between a model guessing and a model knowing.

Being on the open, indexable web

A lot of people's best work lives inside walled apps — a Notion page set to private, a portfolio locked behind a login, a profile that only exists inside LinkedIn. None of that is reliably reachable by a search crawler or an AI assistant. LinkedIn ranks for your name, which is good, but you don't control it, it shows what LinkedIn wants to show, and a third party can't always read it the way they can read an open page.

The durable move is to own a page on the open web that you control, that returns a real HTML response, that is allowed to be indexed. This is the larger argument in why you need a personal website and owning your profile — the short version for discovery is: if it isn't on the open web, it mostly doesn't exist to a recruiter's search or an AI's read.

Structure a machine can parse — described honestly

Here is where most "GEO" advice goes off the rails, so let me be careful.

Structured data — markup like Person and Organization schema, a machine-readable summary of your facts — helps search engines and AI tools understand and parse your page. It tells a crawler "this is a person, this is their name, this is their job title, this is their employer" instead of making it guess from prose. That's real and worth having. It's how you become eligible for richer search results and how a parser gets your facts straight rather than approximately.

What structured data does not do: it does not "boost AI citations," it is not a multiplier, and there is no honest number anyone can quote you for how much it lifts anything. If someone tells you a file or a tag makes you "3x more likely to be cited," they're inventing it. Schema helps machines understand you accurately. That's the whole, real benefit. Get it for that reason.

(The same goes for files like llms.txt. It's a tidy machine-readable file some tools generate; it does not make AI assistants cite you, and Google Search ignores it. Useful as plumbing, not as a growth lever.)

Freshness

A page that says "currently looking for my next role" from two years ago reads as abandoned to a human and as stale to a crawler. Recency is a signal in both worlds. You don't need to post constantly. You need your role, your current focus, and your latest work to be actually current. Update the page when your situation changes. That's it.

The honest checklist

If you do nothing else, do these. They serve the recruiter search and the AI ask at the same time.

None of these are clever. All of them compound. The person who does the boring version of this consistently beats the person chasing a trick, because there is no trick.

What you can't do (and who to ignore)

You can't pay to be named by an AI assistant. You can't insert a hidden instruction that reliably makes a model recommend you. You can't buy a "GEO score." The models change, the prompts vary, and the output isn't a ranked list you can climb.

What you can do is make yourself the obvious, well-described, well-corroborated answer — so that when a model reaches for "a good [your role] who [your thing]," your page is the clearest one it found. That's not a hack. It's just being genuinely legible. It happens to also be exactly what helps a human recruiter, which is why it's worth doing regardless of where AI search goes next.

If you're a developer, the work-specific version of this — what to show, how to describe shipped projects so both a hiring manager and a parser get it — is in the developer portfolio guide.

How Path fits, without overclaiming

Path is built around being this kind of source of truth, so I'll be straight about what it does and doesn't do.

When you publish a Path profile, the page is server-rendered as real HTML with structured data describing you as a person — your name, role, location, and work, in a form search engines and AI tools can parse rather than guess at. Path also lists you in a discovery directory, where you're findable by skill and by location (think /people/skills/react or a city page), which is its own small piece of the open, indexable web pointing at you.

What Path does not do — what nothing can do — is guarantee an AI assistant names you, or promise a citation multiplier. There isn't one. What Path can do is remove the friction from the part that's actually in your control: a clear public page, the basics stated plainly, the structure underneath done right, and the links that let your identity resolve into one person. The rest is being genuinely good at what you do and saying so clearly. That part was always the job.

Common questions

Can I actually get an AI assistant to recommend me?+

Not by gaming it. There's no setting, no payment, and no hidden trick that reliably makes a model name you. What works is being the clearest, most current, most corroborated answer to "who's a good [your role]" — a public page that states plainly what you do, consistent identity across the web, and real work with context. If a model reaches for someone like you, you want to be the legible option it found. That's the only honest play.

Does structured data or an llms.txt file make me more likely to be cited by AI?+

No, and be skeptical of anyone who says it does. Structured data helps search engines and AI tools understand and parse your page accurately — it gets your facts straight instead of guessing them, and makes you eligible for richer search results. That's a real benefit worth having. But it is not a citation multiplier, and files like llms.txt are ignored by Google Search and skipped by most AI crawlers. Treat them as useful plumbing, not a growth lever.

Isn't LinkedIn enough to get found?+

LinkedIn is genuinely useful and it ranks well for your name, so keep it current. But you don't control what it shows, it gates a lot behind logins, and a third party (or an AI tool) can't always read it the way it reads an open page. The strongest setup is a page you own on the open web that states the basics plainly, plus a LinkedIn that agrees with it and links to it. They corroborate each other.

What's the single highest-leverage thing I can do this week?+

Make your name and current role identical everywhere you appear — your site, LinkedIn, GitHub or Dribbble — and link them to each other. This is what lets search and AI resolve all the scattered mentions of you into one confident entity. It costs ten minutes and it's the cheapest real improvement most people are missing.

How often do I need to update my page?+

Not constantly. You need your role, your current focus, and your latest work to be actually current — not a "looking for my next role" line from two years ago. Update it when your situation changes. Recency reads as active to both a human and a crawler; staleness reads as abandoned.