The Job Search Is Moving Beyond Keywords. Is Your Résumé Ready for Semantic Search?

For the better part of two decades, we have taught job seekers to think about their résumés in terms of keywords.

Read the job description. Identify the important skills. Make sure those words appear in your résumé. Use the right job titles. Include the right technical terms. Build a skills section. Make yourself searchable.

There was a good reason for that advice. Recruiting technology has historically relied heavily on matching what an employer searches for with what appears in a candidate's résumé or LinkedIn profile. Recruiters learned Boolean search. Résumé writers learned ATS optimization. Job seekers learned to dissect job descriptions looking for the magic combination of words that might get them through the system.

That isn't disappearing. But recruiting technology is beginning to get much better at something that traditional search has never been particularly good at: understanding what someone's experience actually means.

LinkedIn's Hiring Assistant is a good example of where this is going. Instead of requiring a recruiter to construct a search from a collection of titles, skills, keywords and filters, an AI recruiting agent can interpret a recruiting assignment expressed in normal language.

A recruiter might traditionally build a search that looks something like:

("HR Director" OR "People Director" OR "Director of Human Resources") AND ("hospitality" OR "hotel") AND ("employee engagement" OR "talent management")

There is nothing inherently wrong with that search. The problem is what happens to everyone who belongs in the results but doesn't happen to use those words.

Now imagine the recruiter can instead say:

Find me a senior HR leader for a 1,000-person hospitality company. I need someone who has actually led culture transformation across multiple locations, not someone who has merely supported it. Hospitality experience would be great, but I'm open to people from similar customer-facing businesses with distributed workforces.

Those are two very different ways of searching for talent.

The first is primarily asking, "Who matches the words I entered?"

The second is asking, "Whose experience demonstrates what I mean?"

That distinction is going to matter quite a bit for job seekers.

Keywords aren't going away. Their job is changing.

I would not advise anyone to start stripping keywords out of a résumé. Quite the opposite. A cybersecurity executive's résumé should contain the language of cybersecurity. Someone working in FP&A should probably have FP&A somewhere on the page. A talent acquisition leader should use recruiting terminology.

Professional terminology gives both humans and technology important information about what you do.

What becomes less useful is treating the résumé as a container that needs to hold every possible variation of every keyword an employer might search.

Consider someone working in identity and access management. A keyword-heavy résumé might make sure to include:

Identity & Access Management | IAM | Identity Management | Access Management | Entra ID | Conditional Access | Zero Trust

There is value in some of that terminology. But now consider an accomplishment:

Led identity modernization across 14 business units, migrating 18,000 users to Entra ID and implementing conditional access as part of an enterprise Zero Trust security strategy.

An AI system doesn't have to find the phrase "Identity & Access Management" six times to have substantial evidence that this person has enterprise IAM experience.

It can see what they did.

More importantly, it can see the scale at which they did it.

That is where I think résumé optimization begins to change.

A résumé needs to provide evidence, not just labels.

For years, one of my biggest frustrations with keyword-focused résumé advice has been that it can produce documents filled with the right terminology while saying remarkably little about the candidate.

Take this:

Responsible for talent management, succession planning, employee engagement, workforce planning and leadership development.

There are plenty of keywords in there. From a traditional search perspective, it checks a lot of boxes.

Now compare it with:

Built enterprise talent strategy for 4,500 employees across 26 locations, introducing succession planning and leadership development programs that increased internal promotions 31%.

The second version still contains relevant terminology. But it gives a recruiting agent considerably more information to work with.

The candidate has operated at enterprise scale. They have experience with a geographically distributed workforce. They didn't simply participate in succession planning; they introduced it. Their work included leadership development. There was a measurable change in internal mobility.

An AI system can begin connecting those dots.

That is semantic search.

The system isn't simply asking whether the candidate wrote talent strategy. It is evaluating whether the candidate's experience provides evidence of talent strategy.

Numbers become more than résumé decoration.

I've always pushed candidates to quantify their work because numbers help a recruiter understand the significance of an accomplishment. Semantic recruiting gives us another reason to care about them.

"Managed a recruiting team" tells us something.

Led 14 recruiters supporting 3,200 annual hires across 47 locations while reducing time-to-fill from 52 to 34 days.

That tells us considerably more.

There is team size. Hiring volume. Geographic complexity. Operational responsibility. Performance improvement.

Now imagine an employer asks an AI recruiting agent to find someone who has led a high-volume, multisite recruiting operation.

The candidate never needs to write:

HIGH-VOLUME MULTISITE RECRUITING EXPERT

Their experience already proves it.

The same applies to budget responsibility, revenue, employee populations, geographic scope, transaction volume, customer populations, cost savings, growth and nearly every other measure of scale.

"Managed a large budget" requires interpretation.

"Directed a $38M operating budget" doesn't.

This could finally make job titles a little less powerful.

Titles have always been an imperfect proxy for capability.

A Vice President at one company may be doing work performed by a Director somewhere else. A "Head of People" could be the senior HR executive for an entire organization or the only HR employee at a 40-person startup. A Senior Manager at a massive global company may have considerably more scope than a VP at a smaller organization.

Traditional recruiter search doesn't handle that particularly well because titles are convenient search criteria.

Suppose an employer needs a VP of Talent Management.

The candidate's title is VP, People & Culture.

A traditional search heavily weighted toward titles could miss that person entirely.

But their experience shows that they led succession planning, executive development, performance management, talent reviews, workforce planning and leadership development for 8,000 employees.

An AI recruiting agent has the ability to ask a much more useful question:

The title is different, but has this person actually done the work?

If the answer is yes, the title mismatch becomes less important.

That could be a very good development for candidates whose careers don't fit neatly into standardized boxes.

Career changers may have even more to gain.

Transferable experience has always been difficult to search for.

Imagine a hotel company looking for an HR executive with experience supporting a distributed, customer-facing workforce. A traditional search might heavily favor people who already have hospitality or hotel somewhere in their profiles.

That makes sense until you consider the HR executive coming from luxury retail.

They have supported 60 locations. They understand frontline employees. They have managed seasonal hiring. They have dealt with high turnover, employee relations, geographically dispersed leadership teams and a business where the customer experience is directly affected by the employee experience.

They may never have worked in a hotel.

But much of the operating environment is remarkably similar.

A recruiter using semantic search could potentially ask:

Find senior HR leaders who have managed large, distributed, customer-facing workforces and could transition successfully into luxury hospitality.

Now the system has permission to look for similarity in the experience, not just similarity in the industry label.

That is a major difference.

It also means career changers need to stop assuming that transferable skills will be obvious simply because they know they have them.

If the résumé says:

HR Director | Luxury Retail Company

there isn't much to interpret.

If it explains that the person led HR strategy across 60 retail locations, supported 6,500 frontline employees, managed seasonal workforce planning and reduced turnover across difficult labor markets, the connection becomes much easier to make.

Semantic search can only reason from the evidence you give it.

The résumé also needs to make sense as a career, not just as a collection of jobs.

One of the more interesting possibilities with agentic recruiting is the ability to evaluate someone's career history collectively.

Imagine a progression from HR Manager to Regional HR Director to VP of People & Culture to Chief People Officer.

At each step, the employee population gets larger. The team gets larger. Responsibility expands from one market to several. Then several countries. Budget authority increases. The work moves from execution to strategy.

That progression itself tells us something about the candidate.

For years, résumés have compensated for weak career storytelling with labels:

STRATEGIC EXECUTIVE LEADER

TRANSFORMATIONAL PEOPLE EXECUTIVE

VISIONARY HR LEADER

Those phrases aren't necessarily harmful, but they aren't proof of anything either.

If someone's career demonstrates increasing organizational complexity, broader authority and larger business impact, a sufficiently capable recruiting system should be able to recognize executive progression without being told 11 times that the person is "strategic."

The résumé still has to make that progression visible.

And then there is LinkedIn.

This may be the part job seekers underestimate most.

LinkedIn is not simply a digital copy of your résumé sitting on a networking website. It is becoming part of the data environment in which AI recruiting agents search for, evaluate, and recommend people.

That changes how I think about profiles that contain a title, a company, and dates, and almost nothing else.

For years, some people intentionally kept LinkedIn sparse. The thinking was that recruiters could contact them if they wanted to know more.

That strategy makes less sense when the recruiter may have an AI agent deciding who is worth contacting in the first place.

Your experience, About section, skills, certifications, education, and accomplishments collectively help create a professional representation of who you are.

The more useful question becomes:

If an AI recruiting agent evaluated my LinkedIn profile without ever speaking to me, would it understand what I'm actually qualified to do?

Not what you call yourself.

Not what you hope to do next.

What does the evidence allow it to reasonably conclude?

That is a very different way to audit a LinkedIn profile.

So what changes for the job seeker?

Probably less than people will try to convince you (and absolutely nothing if you’ve worked with me in the past).

There will inevitably be a new wave of advice about "beating AI recruiters," just as we spent years hearing about beating the ATS. Someone will develop a semantic-search score. Someone will promise the 37 secret phrases AI recruiters are looking for. Someone will tell you keywords are dead.

Ignore most of it.

Use the language of your profession. Keywords still provide important signals, and search systems aren't transforming overnight.

But give those words evidence.

If you say you lead transformation, show me the transformation.

If you say you manage enterprise programs, tell me how large they are.

If you say you improve retention, tell me what changed.

If you say you're an executive, show the scale and complexity of what you've led.

If you're trying to switch industries, make the transferable operating environment explicit rather than expecting someone to figure it out.

And stop treating LinkedIn like the résumé's neglected little sibling.

For a long time, résumé optimization largely centered around one question:

Do I have the words the system is looking for?

The next version of recruiting technology is increasingly capable of asking a better one:

Does this person's experience demonstrate what the employer is looking for?

Keywords help answer the first question.

Your career evidence answers the second.

The strongest résumé needs both.