How Companies Use AI in Hiring: What Candidates Should Know

If you have applied for a job recently, software probably touched your application before a person did. AI in hiring now shows up in how recruiters find candidates, how resumes get ranked, how interviews get scheduled, and sometimes in how your recorded answers get scored.

That does not mean a robot decides your future. In most processes, AI sorts and routes while people make the final decisions. But the sorting matters, because a low ranking can mean no one ever reads your resume.

This guide walks through each stage where employers use AI, what the tool is likely doing, what it cannot see, and what you can do differently. It also covers your rights, including when you can ask for an accommodation or an alternative process.

How AI in hiring works across the process#

Hiring has always been a funnel: many applicants at the top, one hire at the bottom. AI tools are mostly used to move people through that funnel faster. Each stage uses a different kind of tool, so it helps to separate them.

The rest of this guide takes each row in turn. If you want a deeper technical explanation of the screening row, our guide to AI resume screening covers ranking models in more detail, and our explainer on what an applicant tracking system is covers the software most employers start with.

AI assists far more often than it decides#

Most tools produce a score, a ranking, a summary, or a recommendation. A recruiter or hiring manager then decides whom to advance. The exceptions are knockout questions and some high-volume hourly hiring, where automated rules can reject applicants who do not meet a stated requirement.

That distinction matters for your strategy. Your job is to be easy for software to read and convincing to the human who reads the shortlist.

How widely employers use AI in hiring#

Adoption is growing but uneven. In LinkedIn's Future of Recruiting 2025 report, 37% of organizations said they were actively integrating or experimenting with generative AI tools in recruiting, up from 27% a year earlier. Almost a third said they were not using or exploring generative AI at all.

Keep in mind that this measures generative AI specifically. Older forms of automation, such as resume parsing, keyword search, and knockout questions inside applicant tracking systems, have been standard for years and are far more widespread.

What recruiters say they do with the time saved#

The same LinkedIn report found that recruiters using generative AI saved about 20% of their work week on average. Among those teams, 35% said they put that time toward candidate screening and 26% toward skills assessments. In other words, the time AI frees up often goes back into evaluating you more closely, not less.

How AI use differs by type of role#

The kind of AI you meet depends heavily on the job. High-volume roles lean on automation to handle thousands of applicants, while senior roles rely more on recruiters and networks. Use this as a rough guide to what to expect.

If you are applying for high-volume hourly roles, precise answers to availability and location questions matter more than resume polish. For senior roles, the resume still needs to parse cleanly, but relationships usually decide who gets a conversation.

Most candidates still don't know how AI is used#

When Pew Research Center surveyed U.S. adults, 61% said they had heard nothing about employers using AI in hiring. Awareness has likely grown since, but many candidates still guess at what is happening. The sections below replace those guesses with specifics.

Sourcing: how AI finds candidates before they apply#

Many hires start with a recruiter search, not an application. Recruiting platforms use AI to suggest candidates whose profiles match a role, rank search results, and draft outreach messages. Job boards use similar systems in reverse, recommending openings to you based on your profile and activity.

What sourcing tools look at#

Sourcing tools draw on what is visible in your profile: titles, skills, employers, location, education, and sometimes activity signals. They match those against the skills a recruiter is searching for. Skills-based search is a growing focus; LinkedIn's report found that companies running the most skills-based searches were 12% more likely to make a quality hire.

How to show up in more searches#

• Use a headline that includes your target title, not just your current one.
• List the specific skills and tools you use, in the same terms job postings use.
• Write your experience with concrete results, since recruiters read the profile after the tool surfaces it.
• Keep location and work-preference settings accurate so you are not filtered out.
• Update your profile before you start applying, not after.

Our guide to LinkedIn profile optimization goes section by section, and a free LinkedIn profile review can flag gaps in search visibility.

Job recommendations work the same way in reverse#

The "recommended for you" jobs on large boards come from similar matching systems pointed at you instead of the recruiter. They learn from your profile, your searches, and the postings you click or save. If your recommendations look wrong, the fix is usually upstream: update your target titles, remove outdated skills, and stop clicking on roles you would never accept. Within a week or two, the suggestions usually get noticeably closer to what you want.

Screening: resume parsing, matching, and ranking#

Screening is where most candidates feel AI the most. When you apply through a company portal, the applicant tracking system usually does three things: it parses your resume into fields, it checks your answers to screening questions, and it may rank or score you against the job.

Parsing turns your resume into data#

Parsing extracts your name, contact details, job titles, dates, employers, education, and skills. If your layout confuses the parser, information can land in the wrong field or disappear. Tables, text boxes, headers and footers, columns, and graphics are the usual culprits. Our guide to resume parsing errors shows what broken parsing looks like.

You can check this yourself before applying. A resume parser shows your resume the way an ATS reads it, field by field, and an ATS resume checker flags sections, dates, and contact details that do not come through cleanly.

Knockout questions are the real auto-rejectors#

Questions like "Are you legally authorized to work in this country?" or "Do you hold an active RN license?" often have a required answer. Answer differently and the system can reject you automatically, regardless of your resume. Read them carefully, answer honestly, and never guess on a required credential.

Matching and ranking#

Many systems compare your resume with the job description and produce a match score or rank. Older systems lean on keywords; newer ones use language models that can recognize related terms. Either way, the safest strategy is the same: describe your real experience using the posting's own language where it is accurate.

The first version gives a matching tool nothing to work with and gives a recruiter no reason to keep reading. The second names the role, the system, the scale, and a result. For more on choosing the right terms, see our guide to resume keywords.

What the recruiter sees after the ranking#

Once a recruiter opens the shortlist, they typically see your parsed profile, your answers to screening questions, sometimes a match score or AI-written summary, and your original resume file. Many recruiters skim the summary and most recent role first, then decide whether to read further.

That means the top third of your resume carries most of the weight. Put your target title, years of relevant experience, and two or three strongest results where a reader sees them in the first few seconds. If an AI summary of your resume is generated, it can only draw on what you wrote, so vague input produces a vague summary.

Assessments: skills tests, games, and work samples#

After screening, some employers send online assessments. These range from straightforward skills tests to personality inventories and game-based exercises. AI may score the results, compare them with profiles of successful employees, or flag unusual response patterns.

Common assessment types#

Do not get hidden help#

Assessments are where AI misuse is most likely to end a candidacy. If instructions say no outside help, that includes chatbots. Employers increasingly use proctoring, response-time analysis, and follow-up interviews that ask you to explain your answers. Prepare honestly instead.

Ask about accommodations early#

If a disability affects how you perform on a timed, visual, audio, or game-based test, you have the right to request a reasonable accommodation. The Justice Department's guidance on AI and disability discrimination in hiring says employers should tell applicants what technology is being used and how they will be evaluated, so that applicants can decide whether to ask for an accommodation. It also says employers must check that tools do not screen out qualified people with disabilities.

Interviews: chatbots, scheduling, and one-way video#

AI shows up in interviews in three main ways: chatbots that run early screening conversations, scheduling assistants that book your interviews, and one-way video platforms where you record answers to preset questions.

Chatbot screeners#

Chatbots usually confirm basics: availability, location, pay expectations, work authorization, and sometimes a few experience questions. Treat each answer like a knockout question. Be precise, avoid ranges you cannot accept, and keep a record of what you said.

One-way video interviews#

In a one-way interview, you see a question, get a short preparation window, and record an answer. Recruiters may review the recordings themselves, and some platforms transcribe and score answers against job-related criteria. You usually will not know which.

That uncertainty points to one strategy: give clear, structured, specific answers that work for any reviewer. Our guide to video interview tips covers setup, and the STAR method gives you a reliable structure.

Practice with an mock interview practice tool so recording yourself feels normal before the real thing.

AI note-takers in live interviews#

In live video interviews, some interviewers now use AI tools to transcribe the conversation and draft notes for the hiring team. That changes little for you, but it does reward clear, complete answers: a transcript captures exactly what you said, not the impression you hoped to leave. If you are unsure whether a call is recorded or transcribed, it is reasonable to ask at the start.

Practical setup checklist for recorded interviews#

• Test your microphone; clear audio matters more than video quality
• Place the camera at eye level with light in front of you
• Close notifications and other apps
• Use the practice question if the platform offers one
• Keep each answer under the time limit with a clear ending
• Write three keywords per likely question on a sticky note, not full scripts

Background checks, references, and verification#

At the end of the process, AI and automation help employers verify identity, check references, and run background checks. Identity verification has become more common as employers deal with fraudulent applicants, so you may be asked to show ID on camera or upload documents through a secure portal.

Your rights during background checks#

When an employer uses a background screening company, the Fair Credit Reporting Act gives you specific protections. According to the Federal Trade Commission's guidance for employers, the employer must:

1. Give you a standalone written notice that a background report may be used, separate from the application.
2. Get your written permission before ordering the report.
3. Before rejecting you based on the report, give you a copy of it and a summary of your rights, so you can point out errors.
4. After a final adverse decision, tell you the screening company's contact details, that the company did not make the decision, and that you can dispute inaccurate information.

If a report contains a mistake, dispute it with the screening company right away and tell the employer you are doing so.

The bias problem: what research shows about AI hiring#

AI hiring tools learn from data, and data reflects past decisions. If past hiring favored certain groups, a model can learn to favor them too. That risk is real, and it is one reason regulators are paying attention.

What one recent study found#

Researchers at the University of Washington tested three AI models on more than 550 real resumes ranked against hundreds of job listings, changing only the names. As UW News reported, the models favored white-associated names 85% of the time and Black-associated names 9% of the time. Male-associated names were favored 52% of the time versus 11% for female-associated names.

This was a research test of specific models, not an audit of any employer's live system. Still, it shows why "the computer decided" is not proof of fairness.

How the public sees AI in hiring#

Americans are skeptical. In Pew's survey, 71% opposed AI making the final hiring decision and 66% said they would not want to apply for a job with an employer that uses AI to help make hiring decisions. Views were more mixed about specific abilities:

That last row is worth noting if you are changing careers or have a nontraditional background. People expect AI to struggle most with candidates who do not fit the usual pattern, which is exactly why referrals and direct outreach matter more for you.

What you can and cannot control#

You cannot control how a model was trained. You can control how clearly your resume communicates relevant evidence, whether a human advocate sees your application, and whether you ask questions about the process. For nontraditional candidates, a referral or a direct note to the hiring manager is often worth more than any formatting tweak.

Your rights when employers use AI in hiring#

The law around AI in hiring is moving quickly, and it varies by location. The table below summarizes the main rules candidates should know about. It is general information, not legal advice, and you should check current requirements where you live.

Federal protections apply regardless of the tool#

The ADA.gov guidance is explicit that employers must avoid using hiring technologies in ways that discriminate against people with disabilities, including tools bought from another company. The EEOC enforces these rules for private employers. The same logic applies to other protected characteristics: a biased outcome is the employer's responsibility.

New York City requires audits and notice#

Since enforcement began in July 2023, New York City's rules on automated employment decision tools require employers and agencies using covered tools to have a bias audit within the past year, post a summary of the results, and give candidates notice 10 business days before using the tool.

State and international rules#

Illinois was early to regulate AI analysis of video interviews, requiring notice, an explanation of how the AI works, and consent. Colorado's SB24-205 goes broader, requiring notice when a high-risk AI system substantially contributes to a consequential decision about you, a chance to correct inaccurate personal data, and an appeal with human review where feasible; its compliance date has already been pushed back once, so check its current status. In Europe, the AI Act's Annex III lists AI systems used for recruitment or selection as high-risk, which brings stricter obligations for providers and employers.

How to ask about AI or request an accommodation#

You do not need to be confrontational. A short, professional email works.

You do not have to name your specific diagnosis in a first request. Describe the functional limitation and the kind of adjustment you need.

Questions you can ask a recruiter about the process#

Asking how you will be evaluated is a normal, professional question. Recruiters field it often, and the answers help you prepare. Good options include:

• Which steps in the process are automated, and which are reviewed by a person?
• How are assessment or recorded interview responses evaluated?
• Is there an alternative format if a step is not accessible to me?
• Who will review my application after the initial screen?
• How long is a typical step, and when should I expect to hear back?

What to do after an automated rejection#

A fast rejection, sometimes within minutes, usually means a knockout question or a rule rather than a judgment of your whole background. Start by rereading the posting and your screening answers to find the mismatch. Then run your resume through a parser to rule out formatting problems that hid your experience.

If you still think you fit the role, look for a person to contact: a recruiter, the hiring manager, or someone on the team. A short, specific note about your fit can get your application a second look, especially for roles that stay open. If nothing changes, many employers let you reapply after a waiting period, often several months, so note the date in your tracker.

How to adapt your application to AI-assisted hiring#

Most of the advice that helps with AI screening also helps with human recruiters. That is good news: you do not need two strategies.

Practical rules that work for software and people#

• Use a simple, single-column layout with standard section headings like Experience, Education, and Skills.
• Mirror accurate job-title language in your summary or headline.
• Name tools, certifications, and methods exactly as the posting does, when you have them.
• Quantify results so the human reading your ranked resume sees impact, not duties.
• Answer screening questions precisely, especially on authorization, licenses, and availability.
• Apply early in a posting's life, since some recruiters review the first strong matches before the rest.
• Find a human path through a referral or a short note to the hiring manager.

Our guide to building an ATS-friendly resume covers formatting rules in detail. If you apply through large portals, our walkthrough on Workday, Greenhouse, Lever, and other ATS portals covers each system's quirks. And before you rework your whole resume, check the claims in our guide to ATS myths; much popular advice is outdated.

Check your match without stuffing keywords#

Run your resume and the posting through a resume keyword scanner to see which important terms are missing. Add the ones you genuinely have, in context, inside real bullets. Avoid hidden text, white-font keyword lists, or instructions aimed at AI screeners. Recruiters see the parsed text, and those tricks can get you flagged.

Using AI yourself without crossing the line#

Employers use AI to sort applicants; applicants use AI to write and prepare. Both are normal now. The line is honesty: tools can help you express real experience, but they cannot invent it, and they cannot replace you in a live evaluation.

If an application asks whether you used AI#

Some applications and assessments now ask directly whether you used AI tools. Answer truthfully and briefly, for example: "I used an AI tool to check my resume against the posting and to proofread; all content reflects my own experience." Employers who ask are usually trying to understand how you work, not to disqualify everyone who uses modern tools. What does damage trust is a "no" that later turns out to be false, or polished written answers you cannot match in conversation.

For a full breakdown of tools by stage, see our guide to AI job search tools. If you are worried about how AI-written material reads to recruiters, our article on whether employers can tell if a resume is AI-generated explains what reviewers actually notice.

Myths and facts about AI in hiring#

Fear fills the gaps when processes are opaque. Here are common beliefs, checked against what the sources above show.

Next steps for candidates#

Start with the parts of the process you control. Run your resume through a parser to confirm it reads cleanly, rewrite your summary so it names your target role and real results, and keep your profile skills current for recruiter searches.

Then plan for the human stages. Practice recorded answers until they feel natural, prepare a short accommodation request if you might need one, and look for a referral or direct contact for every role you care about. AI may sort the pile, but people still choose whom to hire.