Recruiting strategyRead at source, 24 September 2026
How to Use AI in Recruiting Hourly Workers: What Works, What Fails, and the Law
The way to use AI in recruiting hourly workers is to let it do the fast, repetitive work around a hire and keep people on the decisions. It is good at drafting job descriptions and ads, answering and texting applicants within minutes, booking interviews and chasing no-shows. It is weak, and legally exposed, when it ranks or rejects candidates: federal law applies the same adverse-impact test to an algorithm as to any other selection procedure, New York City requires a bias audit and 10 business days’ notice for automated employment decision tools, Illinois has made discriminatory AI use and failure to give notice civil rights violations since 1 January 2026, and Colorado replaced its 2024 AI Act with a new law that applies to decisions made on or after 1 January 2027. None of it creates applicants. That still takes a job ad people see and an application short enough to finish.
Where AI fits in hourly hiring
“AI in recruiting” covers very different tools, and the risk varies with them. A tool that drafts a job ad carries almost none. A tool that decides who gets an interview carries the most. For frontline roles, where most applicants apply from a phone and many have no résumé, the useful question is where in the process the tool sits.
| Stage | What AI does | Where it helps | Where it fails | Covered in |
|---|---|---|---|---|
| Writing | Drafts job descriptions, ads, texts and interview questions | Speed; plain language | Invents requirements and pay; can’t check compliance | ChatGPT for recruiting |
| Screening | Ranks, scores or filters applicants | Sorting large volumes against real requirements | Bias, proxies, rejecting people you would have hired | AI screening |
| Conversation | Chatbots answer questions, ask knockouts, re-engage | Replying in minutes, at any hour | Long scripts that lose applicants | Recruiting chatbots |
| Phone screen | A voice agent calls and asks screening questions | Volume beyond what recruiters can call | Unanswered calls; people who hang up on a machine | AI phone screening |
| Scheduling | Books, confirms and reminds | Fewer no-shows, less back-and-forth | Rarely fails; mostly a cost question | Interview scheduling |
Using AI to write job descriptions and job ads
This is the safest and most useful place to start. A language model turns your notes into a description, cuts it into a social job ad and drafts candidate texts in seconds. The failure is predictable: when you leave a gap, it fills it with a plausible requirement or pay figure you never gave it. Give it the pay, shift, location and genuine requirements, tell it not to add anything, and have a person check the draft. Our ChatGPT for recruiting page has prompts that work, and the free job description generator builds frontline descriptions without guessing. For the ad itself, see job ad copywriting.
Using AI to screen applicants
Screening is where AI is most heavily sold and most heavily regulated. For hourly roles, the screen that works is usually two or three knockout questions about things the job legally or physically requires: a license, a shift, a lifting requirement. That doesn’t need a model. Where AI ranks or scores people, it becomes a selection procedure, and the rules in the legal section below apply. Before buying, ask the vendor what the tool decides, on what data, and whether it has been tested for adverse impact on your applicants. The detail, law by law, is on our AI screening page; our guides to pre-screening questions and applicant screening software cover the non-AI options.
Chatbots and automated texting
The best-supported use of automation in hourly hiring is speed. An applicant who applies on a break is also applying to two other jobs, and the first employer to reply usually gets the interview. A chatbot or automated text can confirm the application, ask a couple of questions and offer interview times within minutes. Where chatbots lose people is length: every extra question in the script is another point to drop out. Our recruiting chatbot page separates the four jobs sold under that name, and text recruiting software covers the simpler option of a text thread with a person behind it.
AI phone screening
A voice agent calls the applicant, asks the screening questions and hands your team a transcript and summary. It helps when applicants outnumber the hours anyone has to call them. It loses applicants who don’t answer unknown numbers and those who hang up on a machine, so for hourly roles it works better as a second touch after a text than as the first contact. Details and vendor questions are on our AI phone screening page.
Where AI fails in hourly hiring
- It can’t create applicants. Every tool on this page acts on people who already applied. If too few people see the job, AI speeds up a small funnel.
- Résumé ranking fits hourly roles badly. Many good frontline applicants don’t have a résumé, or have one that says little about whether they will show up for a 6 a.m. shift.
- Proxies creep in. Distance, gaps in work history or “fit” scores can stand in for protected characteristics. Illinois law names zip codes used as a proxy for protected classes as a civil rights violation.
- Automation adds steps. A chatbot, an assessment and an AI phone screen stacked after an application can each lose a share of applicants. In Boostpoint’s 2026 benchmark of 891 managed campaigns, what happens after the click explained 70% of the difference in cost per applicant between campaigns.
- Nobody owns the decision. When a tool rejects someone, the employer is still responsible for that decision, whoever built the tool. Keep a named person accountable for every automated step.
The legal layer: AI hiring laws as of September 2026
No single federal AI hiring statute exists. What binds an employer is existing anti-discrimination law plus a small number of state and city rules. These are the ones most employers ask about.
| Where | Rule | What it requires | In effect |
|---|---|---|---|
| Everywhere (federal) | Title VII and the Uniform Guidelines, 29 CFR 1607.4(D) | A selection rate for any race, sex or ethnic group below four-fifths of the highest group’s rate is generally regarded as evidence of adverse impact. The EEOC says a required test “must be necessary and related to the job.” | Guidelines since 1978 |
| New York City | Local Law 144 | A bias audit of an automated employment decision tool within one year of use, a published summary of results, and notice to candidates and employees 10 business days before use | Enforced since 5 July 2023 |
| Illinois | AI Video Interview Act (820 ILCS 42) | Before AI analyzes a recorded video interview: notify the applicant, explain how the AI works and what it evaluates, and get consent; limit sharing; delete on request within 30 days | Since 1 January 2020 |
| Illinois | Public Act 103-0804 (HB 3773), amending the Human Rights Act | Civil rights violation to use AI that “has the effect of subjecting employees to discrimination” on a protected basis, or to use zip codes as a proxy, in recruitment, hiring and other decisions; and to fail to give notice that AI is used | Since 1 January 2026 |
| California | Civil Rights Council automated-decision systems regulations | Applies existing Fair Employment and Housing Act rules to automated-decision systems used in employment | Since 1 October 2025 |
| Colorado | SB 26-189, which repealed and reenacted the 2024 AI Act (SB 24-205) | Notice when automated decision-making technology is used in consequential decisions, including employment; a plain-language explanation within 30 days of an adverse outcome; a right to request human review | Takes effect 1 January 2027; applies to decisions made on or after that date |
Three practical notes. First, Colorado’s original AI Act never took effect: its start date was pushed to 30 June 2026, and SB 26-189, signed on 14 May 2026, replaced it before then. Second, the Illinois Human Rights Act amendment directs the Department of Human Rights to adopt rules on when and how notice must be given, so check the department’s current guidance before you write yours. Third, the EEOC’s earlier technical-assistance pages on AI in employment selection returned “page not found” when we checked; the laws they described still apply. Our AI screening page covers each rule in depth, including what to ask a vendor.
Rule of thumb: if a tool only drafts, schedules or reminds, the legal exposure is low. If it scores, ranks, filters or rejects people, treat it as a selection procedure: know what it decides, test it for adverse impact, give any notice your location requires, and keep a person accountable for the outcome.
One AI system you already use: the ad delivery algorithm
If you advertise jobs on Facebook or Instagram, Meta’s delivery system decides who sees the ad. Employment ads have to run in Meta’s Special Ad Category, which limits targeting options and sets a minimum 15-mile location radius. That is why the ad itself carries the targeting: pay, shift and town in the first line tell the right people it is for them. Boostpoint writes and runs those campaigns; see how we run social job advertising and the 2026 benchmark data on what an applicant costs by role.
How to start using AI in recruiting this quarter
Start with drafting
Use a model to write job descriptions, ads and candidate texts from your facts. Have a person review every one before it goes out.
Fix speed before sophistication
Measure how long an applicant waits for a first reply. Automated texting or a short chatbot that replies within minutes usually does more for hourly hiring than any scoring tool.
Automate scheduling and reminders
Let applicants book their own interview and get reminders. It cuts no-shows without judging anyone.
Add screening last, and carefully
If you add a tool that ranks or rejects, check the rules where you hire, ask the vendor for adverse-impact testing, and keep a person responsible for the decision.
Frequently asked questions
How can AI be used in recruiting?
For hourly roles, the strongest uses are drafting job descriptions, ads and candidate messages; replying to applicants within minutes by text or chatbot; booking interviews and sending reminders; and running phone screens when volume outruns your team. Using AI to rank or reject candidates is the riskiest use, legally and practically, and needs testing for adverse impact and human accountability.
Is it legal to use AI in hiring?
Generally yes, but the employer remains responsible for the result. Federal anti-discrimination law applies to AI tools as to any selection procedure. New York City requires a bias audit and notice for automated employment decision tools, Illinois requires notice and bars discriminatory AI use since 1 January 2026, and Colorado's replacement AI law applies from 1 January 2027. This is not legal advice; check the rules where you hire.
What is the Illinois AI hiring law?
Illinois has two. The Artificial Intelligence Video Interview Act, in force since 2020, requires notice, an explanation and consent before AI analyzes a recorded video interview. Public Act 103-0804, which took effect on 1 January 2026, amends the Human Rights Act to bar AI that has the effect of discriminating on a protected basis, bar zip codes as a proxy, and require notice that AI is used.
When does the Colorado AI Act take effect?
The original act, SB 24-205, never took effect. Its start date was delayed to 30 June 2026, and it was repealed and reenacted by SB 26-189, signed on 14 May 2026. The new law takes effect on 1 January 2027 and applies to consequential decisions, including employment decisions, made on or after that date.
What does NYC Local Law 144 require?
Employers and employment agencies using an automated employment decision tool for jobs in New York City must have had a bias audit of the tool within one year of use, publish a summary of the results, and notify candidates and employees 10 business days before using it. The city's Department of Consumer and Worker Protection has enforced the law since 5 July 2023.
Does AI screening work for hourly jobs?
Partly. Knockout questions about genuine requirements work well and don't need AI. Ranking résumés works poorly because many good hourly applicants have thin or no résumés. Each added automated step can also lose applicants: in Boostpoint's 2026 benchmark, what happens after the click explained 70% of the difference in cost per applicant between campaigns.
What is the four-fifths rule?
It is the federal rule of thumb for adverse impact in 29 CFR 1607.4(D). A selection rate for any race, sex or ethnic group that is less than four-fifths, or 80%, of the rate for the group with the highest rate will generally be regarded as evidence of adverse impact. It applies to AI screening tools just as to tests and interviews.
Will AI get me more applicants?
Not by itself. Chatbots, screening and scheduling tools act on people who have already applied. More applicants come from putting the job in front of more of the right people, with pay, shift and location up front, and an application short enough to finish on a phone. AI can help write that ad; it can't place it.
AI can sort applicants. It can't find them.
Boostpoint writes and runs Facebook and Instagram job ads for frontline employers, with pay, shift and location up front and short in-app applications.
Book a DemoLegal sources read at source on 24 September 2026: NYC DCWP, Illinois General Assembly (820 ILCS 42; Public Act 103-0804), Colorado General Assembly (SB 26-189 signed act, section 5), California Civil Rights Council, EEOC, eCFR. Benchmark: Boostpoint 2026 Social Job Advertising Benchmark, 891 managed Meta campaigns (1,334 campaign-months).