Blog/AI In HR

Recruiting automation: what to switch on first, and what to keep

Flowmingo Editorial TeamFlowmingo Editorial Team18 min lezen
Recruiting automation: what to switch on first: the stage email Flowmingo sends when a candidate moves to Shortlisted

Your 1 open role has 200 to 300 applications, and every vendor wants to sell you recruiting automation without saying which part of your recruiting workflow to hand over first.

The pile comes from the market, not from you. Ashby, a vendor that analysed more than 100 million applications on its own platform, says applications per hire have tripled since 2021.

The teams reading those applications have shrunk at the same time. Greenhouse, which sells an applicant tracking system (ATS), says recruiters per organisation fell 55.6% from 2022 to 2025.

Nor does the advice you find online settle which step comes first. Of the 10 pages Google ranks highest for recruiting automation, 4 tell you to map your process and start with the step that slows it most, and none ranks the steps by risk.

That order matters, because the wrong first step costs you people you never see. Hidden Workers, a survey of 2,275 executives that Harvard Business School and Accenture published in 2021, more than 5 years ago, shows how.

In it, 88% of employers said that filters demanding an exact match to the job description screen out qualified high-skill candidates. A filter like that removes people before any person has seen them.

The order should therefore follow how much each step decides about a person. Status updates and scheduling decide nothing about anyone, so they go first, while scores and rejections decide a great deal, so a person keeps control of them.

Ranked that way, the 2 safest steps still save real time. In section 2's worked example, a team that makes 5 hires a quarter from roles of 300 applicants saves 91.6 net hours a quarter, and those 2 steps return 26.1 of them before any software judges an applicant.

Key takeaways

  • What should I automate first in hiring? Status updates and scheduling. They decide nothing about a person and return 26.1 of 91.6 net hours a quarter.
  • How many hours does recruiting automation save? About 92 net hours a quarter at 300 applicants a role and 5 hires, after 23 hours of review.
  • Can the software reject candidates for me? Our rule: only by a yes/no rule on a written requirement, such as the right to work. The UK's data regulator, the ICO, treats a score that auto-rejects as automated decision-making, so a named person signs off the rest.
  • Did the EU dates for hiring tools move? Yes, from 2 August 2026 to 2 December 2027.
  • How would I know the tool is rejecting good people? Our rule of thumb: read 20 of the lowest-scored files per role each week. If you would have interviewed more than 2, fix the criteria.

1. Hiring alone with 200 applicants a role, which recruiting automation do I switch on first?

Switch on the 2 steps that decide nothing about a person first: status updates, then scheduling. Add steps that score people later, and keep rejections under a named person's sign-off.

Recruiting automation is software that does a hiring task so that a person need not repeat it, whether the task is an update, a booking, a question or a score.

A pile of 200 to 300 applicants is ordinary now. Ashby's May 2026 release puts the average at more than 300 applications per hire, and Greenhouse counted 244 applications per job in 2025.

We ordered the 6 steps by how little each one decides about a person, and then by how easily someone can check what it produces. A yes/no rule can be read and tested, but a score cannot, which is why knock-out questions, step 3, come before the AI interview, step 4.

Hours did not set the order. The clearest sign is CV scoring, which returns the most time, 40.0 net hours a quarter in our worked example, and still comes fifth. The order is our judgment rather than a published ranking.

One law draws a similar line. Colorado's AI law, SB 26-189, takes effect on 1 January 2027. It leaves out tools that only prepare information for a person to review, but it covers a tool that ranks, scores or recommends.

This account comes from a summary by Proskauer, a law firm, rather than from the bill text. In the summary's words, the law excludes tools used solely to "summarize, organize, translate, draft, route, or present information" for human review of administrative processing.

In our reading, that exclusion covers status updates and scheduling. The table shows the main risk at each step and what a person still does once software takes it over.

Its last row names the rules a rejection must meet, which sections 7 and 10 explain.

Step Main risk A person still
1 Status updates Hollow or false updates Writes templates and dates
2 Scheduling and reminders Wrong time zone, reminders to declined people Handles exceptions
3 Knock-out questions A proxy rule rejects good people Owns the list, samples declines
4 AI interview Candidate trust; high-risk in the EU Reads every report
5 CV scoring Silent filtering, name and order bias Reads 60 CVs a role, signs off the cut
6 Rejection Breaking EU, UK or Colorado rules on automated decisions Decides and writes the reason

Recruiting automation in order: status updates, scheduling, knock-out questions, AI interview, CV scoring, then rejection decided by a person

Automation ladder: steps 1 and 2 carry low risk, step 3 medium, steps 4 and 5 high, and step 6, rejection, stays with a person.

A recruitment chatbot is a channel rather than a step, so its risk follows what it does with applicants' answers. It sits at step 1 when it sends updates, step 2 when it books interviews, step 3 when it asks knock-out questions and step 4 or 5 when it scores people.

2. How many hours does recruiting automation save at 300 applicants a role, 5 hires a quarter?

About 91.6 net hours a quarter in our worked example, which is the 114.8 hours that software removes minus 23.2 hours of review that stays with a person.

Say you make 5 hires a quarter from 5 roles of 300 applicants each, which is what Ashby's average of more than 300 applications per hire implies. That gives you 1,500 applicants and about 114 interviews a quarter, because Greenhouse counts 22.7 interviews per job.

For the first round, we assume 20 candidates a role get a 15-minute AI interview, 100 across the 5 roles.

That length is the middle of the 10 to 20 minutes for a full human interview reported by Jabarian and Henkel, researchers at Carnegie Mellon and Erasmus University Rotterdam whose field test of AI interviews appears in section 5.

The steps do not overlap, which matters for the total. Those 100 AI interviews need no booking, so step 2, scheduling, counts only the 14 interviews still held live.

The review column is the work that stays with a person: checking 12 sent messages and 24 declined files a role, and rebooking 1 interview in 10 by hand. Every minute in the table is our assumption, so time your own work and use your numbers instead.

Step By hand today (assumed) Units a quarter Gross hours Review that stays Net hours
Status updates 1 minute per applicant 1,500 25.0 1.0 (60 messages, 1 minute) 24.0
Scheduling and reminders 10 minutes per interview 14 (114 minus 100) 2.3 0.2 (1 in 10 rebooked, 10 minutes) 2.1
Knock-out questions 0.5 minute per applicant 1,500 12.5 2.0 (120 declined files, 1 minute) 10.5
AI interview 15 minutes per call 100 25.0 10.0 (100 reports, 6 minutes) 15.0
CV scoring 2 minutes per CV 1,500 50.0 10.0 (60 CVs a role, 5 roles, 2 minutes) 40.0
Total 114.8 23.2 91.6

CV scoring removes the most work, 50.0 hours of first reads, and hands 10.0 of them back to a person who reads 60 CVs a role. At 2 minutes a CV, that is 2 hours a role, and it is the price of a safer cut.

The 2 safest steps, status updates and scheduling, return 26.1 net hours between them, 28.5% of the total.

Recruiting automation saves 91.6 net hours a quarter at 300 applicants a role; status updates and scheduling return 26.1 of them

Net hours by step: 24.0, 2.1, 10.5, 15.0 and 40.0 hours a quarter add to 91.6. Status updates plus scheduling return 26.1 (illustrative example, not customer data).

One cost is missing from the table. The weekly audit in section 8 adds about 13.3 hours a quarter, 20 files a week at 2 minutes each for 5 roles open about 4 weeks, which brings a cautious net down to about 78.

3. Can I safely automate status updates and scheduling, or will candidates feel ghosted by a bot?

Yes, these 2 are the safest steps of recruiting automation, because they decide nothing about a person. Keep every message true: a bot that promises a reply it cannot keep feels like ghosting.

Silence is what hurts your employer brand. In the 2025 Candidate Experience (CandE) benchmark, 31% of North American candidates had not heard back 1 to 2 months after applying.

Talent Board founded the benchmark and Survale has run it since 2025, and that year's edition drew more than 66,000 completed candidate surveys.

In that edition, candidates also complained about automation: AI or automated screening was among their top 3 negative themes. Yet every one of the benchmark's top-10 winners used some AI, which suggests the complaint is screening that no person seems to check, not AI as such.

The first 2 steps of recruiting automation need no special tool. A mail merge or your ATS's automatic email can send the 2 messages below, and a booking link covers scheduling. Send the acknowledgement within 1 hour of the application.

Subject: We got your application for [Role] at [Company]

Hi [First name], thanks for applying for [Role]. We will tell you by [date, within 7 days] whether you move on. To ask for an adjustment, reply to this email.

[Name], [Company]

Subject: Still deciding on [Role]: update by [new date]

Hi [First name], we promised an answer about [Role] by [date] and are not there yet. We will write by [new date, within 7 days].

[Name], [Company]

Each message follows 4 rules:

  • Promise a yes or no within 1 week, which is our rule. Talent Board's 2024 benchmark puts it at 1 to 2 weeks at most (ERE).
  • Sign with a name and a reply address a person reads, because 33% of rejected candidates in the 2025 CandE benchmark got a do-not-reply email.
  • Send "no news yet" on the date you promised, because a status that never changes reads as silence.
  • Never remind someone you have declined, because it shows the system lost track.

Reminders are the other half of step 2, and the right moment to send them depends on how long applicants take. Flowmingo's own records show how long that is for an AI interview, counted from the invite:

Flowmingo data · 52,078 AI interview invites sent by recruiters · 866 companies · 15 Jun to 13 Sep 2026

  • 30.9% of invited applicants finished the AI interview within 14 days, and 14.9% finished within 24 hours.
  • 24.8 hours was the median time from the invite to a finished AI interview, among the 16,071 who finished.
  • 22.2% of those 16,071 finishers came in after the first automatic reminder, sent at a median of 48 hours.

Invites to candidates from Flowmingo Network, Flowmingo's own candidate pool, are excluded. Only finishers have a time. Reminders go to invitees who had not opened the link, so this shows timing, not cause.

Recruiting workflow timing: 48.2% of candidates who finish an AI interview do so within 24 hours of the recruiter's invite

Invite finish curve: of 16,071 finishers within 14 days of a recruiter invite, 4.9% finished within 1 hour and 48.2% within 24 hours. By 48 hours 74.3% had finished, and 98.3% within 7 days.

Most invited applicants do not finish on the first day. Only 14.9% finish within 24 hours, and 22.2% of those who finished came in after the first reminder, though that shows when they finished rather than what made them.

So for a self-paced interview, send reminders on day 2 and day 5, unless your tool already does. For a live slot, use the day-before reminder in our interview invitation email templates.

4. Can knock-out questions cut my pile without rejecting good candidates?

Yes, if each knock-out is a yes/no check on a written requirement the job truly needs. Rules built on proxies, such as gaps or degrees, reject good people unseen.

In the UK, this is the 1 step of recruiting automation where an automatic no does not count as automated decision-making, so the extra safeguards for those decisions do not apply.

The UK's Information Commissioner's Office (ICO) has published use cases that show when hiring software counts as automated decision-making.

In them, a yes/no knock-out on a written requirement, such as the right to work, is not automated decision-making. A fit score that auto-rejects below a pass mark is.

The trap is the proxy, a rule that tests a stand-in, such as an employment gap or a degree, instead of the requirement itself. In the same Hidden Workers survey, 48% of employers that filter middle-skill candidates used an employment gap over 6 months.

Screening on age itself has a price. In a 2023 US case, iTutorGroup paid $365,000 to settle a suit by the Equal Employment Opportunity Commission (EEOC), which alleged that the company's software rejected applicants by age.

So test every rule on 20 past applicants before it goes live, and once it runs, re-read 24 declined files per role, 8% of a 300-applicant pile.

Rule Keep? Why
Right to work or a required licence Keep A written legal requirement
Shifts or pay range printed in the posting Keep Printed, so yes or no; never ask pay history
Employment gap over 6 months Remove A proxy, used by 48%
Degree the law does not require Remove A proxy, so score the skill instead
Age or birth date Remove Direct age screening
Graduation year Remove An age proxy, so ask for skills instead

Send the decline within 3 to 5 days of the application, the pace that 60% of the top 10 North American winners in the 2025 CandE benchmark kept. In it, state the reason, say no score applied and offer a route to a person. Our rejection email templates have the tiers.

Subject: [Role] at [Company]: our decision

Hi [First name], thank you for applying for [Role]. This role needs [requirement], and your answer says you do not have it. This decision came from that answer alone, with no CV score. If we misread it, reply and a person will look again.

[Name], [Company]

5. Will good candidates walk away, or stop trusting us, if an AI runs the first interview?

Some will, so say plainly that it is an AI and what it does with the answers. In 1 large field test, candidates interviewed by an AI got more offers, and 78% chose it when given the choice.

The trust gap is real. In a Gartner survey of 2,918 job candidates in the first quarter of 2025, only 26% trusted AI to evaluate them fairly.

In Greenhouse's 2026 survey of US job seekers, 38% had walked away over an AI interview and 44% wanted to be told upfront.

The field test is a preprint by Jabarian and Henkel, the Carnegie Mellon and Erasmus University Rotterdam researchers cited in section 2. They ran it at PSG Global Solutions, a recruitment firm in the Philippines that hires entry-level customer-service staff for client companies.

Of 67,056 randomised applications, 53,660 went to an AI or a human interviewer. Offers were 9.73% with the AI against 8.70% with a person, and humans made every hiring decision.

The AI did not win on every count. Applicants rated its interviews as less natural, and those who chose the AI scored lower on language and analytical tests, on average.

Structure is the part worth keeping from human interviewing. In a 2022 review of human-run interviews, published more than 3 years ago, the same questions scored on a written rubric predicted job performance better than loose chats.

Whatever tool you use, put a notice like this one in the interview invite. Add the consent tick-box that Illinois requires, under its AI Video Interview Act, before AI analyses a video interview.

This interview is run by an AI. It asks every applicant the same questions, with a follow-up when an answer is thin, and scores the answers against the role's criteria.

[Our recruiters and the hiring manager] can see your recording, and we keep it for [period]. A person reads the report and decides. For an adjustment or a person, reply to this email.

[ ] I agree to take part in this AI-run interview.

Our own completion times show when applicants actually take a self-paced interview:

Flowmingo data · 41,126 completed AI interviews · 856 companies · interviews started 24 Jun to 19 Sep 2026

  • 60.6% of completed AI interviews were finished outside Monday to Friday 9:00 to 17:00 in the candidate's local time.
  • 34.4% finished on weekday evenings from 17:00 to midnight, and 17.6% at weekends.

Local time is estimated from the candidate's connection, and 9 to 5 on weekdays is assumed. About 61% of applicants are in India and Nigeria, and this shows timing, not preference.

Automate candidate screening in the candidate's own time: 60.6% of AI interviews finish outside weekday 9:00 to 17:00

Out-of-hours finishes: of 41,126 completed AI interviews, 34.4% finished on weekday evenings, 17.6% at weekends and 8.6% before 9:00 on weekdays. So 39.4% finished in a 9-to-5 weekday.

In other words, 6 in 10 completed interviews landed outside a weekday 9-to-5 in the candidate's own time zone. A first round held only in those office hours would have needed another slot for most of them, though this shows timing, not preference.

How Flowmingo helps

With the AI interviewer you send 1 link, and each applicant interviews in their own time. You get the recording, transcript and a score out of 10 with a reason per criterion, and who moves forward is always your call.

6. Should software score and rank my CVs, or is that where automation goes wrong?

Use scores to order the pile, and let a named person sign off the cut after reading samples above and below it.

Scoring has 2 documented failures. The first is about names: in a resume audit by Wilson and Caliskan, 3 AI models that match CVs to a job picked the top 10% for each job from the same CVs under White or Black names.

They favoured White-associated names in 85.1% of 27 race tests, though the result depends on the names used.

The second failure is about order. In a study by Rozado, 22 AI models given 2 equally qualified CVs picked the first-listed one 63.5% of the time.

A human screen is not a safe baseline either. Li, Raymond and Bergman studied 88,666 applications at 1 Fortune 500 firm, where 10% of the people recruiters chose to interview got the job, against an estimated 27% to 32% for the CV-screening models they built.

Better hit rates did not settle fairness, though. Their standard model, trained on past hires, favours groups with a proven track record at the firm, while their exploration model also picks some applicants it is least sure about, to learn how good they are.

The standard model would have cut the Black and Hispanic share of interviewees from the recruiters' 9.4% to 4.2%, while the exploration model would have raised it to 24.3%, so design decides who gets cut.

That is why the cut needs an owner, who signs the declines below the line only after reading the band under it and 20 random CVs. If more than 2 of those 20 are people you would interview, read the next 20 first.

Reading a sample has a limit. The ICO treats ranking and then rejecting without considering every application as automated decision-making.

So reading 60 of 300 does not make the other 240 declines human decisions, and what follows from that depends on where you hire.

In the EU, have a person look at each file before a decline, for the reasons in section 7.1. UK law allows automated declines with safeguards, so tell candidates you used a score and offer them a person.

Automate candidate screening safely: after scoring 300 CVs, a person reads the top 20, the next 20 and 20 at random, then a named person signs off the cut

Sort, then a person signs off the cut: read the top 20, the next 20 and 20 at random from the other 260. That is 60 of 300 CVs, and the 240 unread declines count as automated.

Order bias has a cheap fix. Shuffle the order before any AI comparison and run it twice, swapped, because the first-listed CV wins more often. Our candidate shortlisting guide has the full method.

How Flowmingo helps

CV Evaluation turns a pasted job description into criteria you edit and rate by importance, and scores every CV out of 10 with written evidence.

The top rating, Must Have, gives a criterion the most weight but is not a knock-out: nobody is filtered out automatically, so you decide who moves forward.

7. Does a person need to read every rejection, or can recruiting automation auto-reject low scorers?

Our rule: auto-reject only on a yes/no rule for a written requirement. Every other rejection needs a named person who has read the evidence, or goes out as an automated decision with a route to a person.

The rule has 5 clauses:

  1. The only automatic no is a written knock-out applied to everyone, owned by a person.
  2. A named person with authority to overturn it makes or signs off every other rejection (Article 26 and Article 14 of the EU AI Act, from 2 December 2027). Unread declines count as automated (section 7.1).
  3. Form a view before the AI's pick, because in a 2025 experiment by Wilson, Caliskan and 2 colleagues, 528 people followed a deliberately biased AI up to 90% of the time.
  4. Every rejection carries a plain reason and a route to a person.
  5. Keep logs for at least 6 months in the EU, at least 4 years in California and, per law-firm summaries, 3 years in Colorado.

7.1 Can I auto-reject in the EU or UK without a person looking, or does GDPR Article 22 ban it?

In the EU, not on a score alone: Article 22(1) of the General Data Protection Regulation (GDPR), the EU's data protection law, gives people the right not to face a significant decision based solely on automated processing.

In the UK you can, with safeguards, since the rules changed on 5 February 2026. But the ICO treats a fit score that auto-rejects as automated decision-making, so those safeguards apply to it.

The GDPR's own Recital 71 names e-recruiting without human intervention as an example of such a decision. In the 2023 SCHUFA ruling, a credit-scoring case, the EU's Court of Justice treated a score as the decision when users rely on it heavily.

That ruling concerned credit rather than hiring, so applying it to a CV score is an inference. Still, on that reasoning, a score you follow almost every time can count as the automated decision, even if a person clicks the final button.

8. How would I know recruiting automation is silently rejecting good people, and what breaks first?

Look at the pile you never see, because a silent rule hides there. Sample the lowest scores and declines every week, and compare outcomes across demographic groups, as check 4 below explains.

In our judgment, proxy rules break first, then scoring and order bias, and then candidates gaming the tool. In Greenhouse's 2025 survey of 1,200 US job seekers, 41% admit using prompt injections, hidden text meant to get past AI filters.

A vendor's own security can fail as well. In June 2025, 2 researchers, Ian Carroll and Sam Curry, got into McHire, a recruitment chatbot from the vendor Paradox.ai that McDonald's franchisees use, with the default login 123456.

They said personal data on more than 64 million applicants was reachable, a figure they estimated from record numbering. Paradox.ai's own account says the researchers viewed 5 candidates' details.

The login stopped working the evening they reported it, and Paradox.ai confirmed a fix on 1 July 2025.

For the rules and scores, run these 5 checks on our rule-of-thumb cadences:

  1. Weekly, about 40 minutes a role: read 20 of the lowest-scored or declined files.
  2. Weekly: count actions a person reopens or corrects within 48 hours.
  3. Monthly: compare the tool's top 20 with your own on 10 CVs.
  4. Quarterly: divide each group's selection rate by the top group's, where lawful to collect. Treat a ratio under 80%, a long-standing US rule of thumb, as your own monitor.
  5. Quarterly: re-run 5 CVs with names hidden and order swapped.

Recruiting workflow checks: 5 reviews a hiring team runs weekly, monthly and quarterly to see what screening tools hide

Weekly audit: 5 checks, 2 weekly, 1 monthly and 2 quarterly. They run from reading 20 low-scored files to re-running 5 CVs with names hidden and order swapped.

There is a reason check 4 uses the 80% line only as your own monitor. On 9 June 2026 the Justice Department's Office of Legal Counsel said the guidelines behind it rest on an unconstitutional reading of Title VII.

The weekly read also tells you when to act. If more than 2 of the 20 would have earned an interview, fix the criteria, which is our rule of thumb.

9. Is my screening tool high-risk under the EU AI Act, and what are the new dates?

If the AI in your recruiting automation filters CVs, scores or ranks candidates or evaluates them in an interview, yes. The duties for employers moved from 2 August 2026 to 2 December 2027.

What decides it is Annex III point 4(a) of the EU AI Act, which lists AI that analyses and filters applications and evaluates candidates as high-risk. It names no scheduling or status messages, so in our reading steps 1 and 2 of this guide fall outside it.

Small teams should not count on a pass. We found no small-business exemption in Article 26 for deployers, the employers who use the tool. This means a team using a high-risk tool has until 2 December 2027 to be ready.

The only allowance for size we found is in the fines. For breaking employer duties they reach EUR 15,000,000 or 3% of worldwide turnover, whichever is higher, or whichever is lower for small and mid-sized enterprises.

The ban on inferring emotions in hiring, from the face, voice or gestures, can cost up to EUR 35,000,000 or 7%. That is why it matters what an AI interview scores.

Scoring the words of an answer against a rubric is different from reading the face or voice, and vendor question 3 in section 10.1 asks which one a tool does. This is general information, not legal advice.

As for the timetable, the new date for employers comes from an amending law, Regulation (EU) 2026/1744, in force since 27 July 2026 (EUR-Lex, Steptoe). Not every date moved: the emotions ban and Article 50(1), the duty for an AI that talks to people to say it is an AI, kept theirs.

Date What starts
2 Feb 2025 The Article 5 bans, including reading emotions in hiring
2 Aug 2026 Article 50(1), a provider duty: AI that talks to people tells them it is an AI
2 Dec 2027 Annex III duties for employers using hiring AI: oversight, logs of at least 6 months, notice

Recruiting automation under the EU AI Act: duties for hiring tools moved from 2 August 2026 to 2 December 2027

EU dates for hiring AI: the ban on reading emotions applies since 2 Feb 2025. Employer duties for hiring tools start 2 Dec 2027.

10. Do NYC, Illinois, Colorado or California rules reach a 15-person company, and who is liable?

Assume they can, because New York City's 2023 FAQ on its law for automated hiring tools sets no size threshold, and a law firm reads Colorado's law as hard to avoid for web-hiring employers. The employer is ultimately responsible, and the vendor can face claims too.

We found no small-employer exemption in Colorado's law either. As with the rest of this section, that is general information, not legal advice.

Each of the 4 places asks for something different, so read the rows for the places where you hire, and note that NYC's rule depends on where the job sits:

Where Rule and start date What you owe
NYC Local Law 144, enforced 5 Jul 2023 For NYC-office jobs: independent audit for bias within a year, public summary, notice to NYC residents 10 business days before
Illinois HB 3773, 1 Jan 2026; AI Video Interview Act No discriminatory AI or zip code proxy; notice. For AI video analysis: tell the applicant, explain what it checks, get consent, delete a video within 30 days of a request
California Civil Rights Council, 1 Oct 2025 Vendors count as your agent; keep automated-decision data at least 4 years
Colorado SB 26-189, 1 Jan 2027 Notice; plain description within 30 days of an adverse decision; human review on request where commercially reasonable

10.1 If my vendor's tool discriminates, am I liable, and what should I ask before I sign?

Yes, you can be, because California treats a vendor that screens for you as your agent.

In the Mobley case, a federal court in California let claims proceed against Workday, a software vendor, as an agent. That shows a vendor can face claims too, but it does not let the employer off. Ask these 7 questions before you sign:

  1. How do you classify the tool under EU Annex III?
  2. Can you share an independent audit for bias from the past year?
  3. Does it score the words or the voice?
  4. Where is candidate data stored, and how do I delete it?
  5. How do you test security, given McHire?
  6. Can I export the reason and log for each decision?
  7. Who tells candidates it is an AI?

10.2 I run a 15-person company with no HR: what is the minimum I should do this month?

Do these 4 things, which are our minimum rather than a legal standard:

  1. List every recruiting automation tool that scores, ranks or screens, and switch off automatic rejection except yes/no knock-outs.
  2. Add 1 line to postings saying AI helps sort applications, if true.
  3. Name the person who signs off rejections, and treat unread declines as automated.
  4. Keep applications, scores and reasons for the periods in section 7, and check outcomes by group each quarter.

How Flowmingo helps

Flowmingo is a free AI interviewer for recruiters, and every hiring decision stays with your team. When you reach step 4, try Flowmingo for the AI interview step of your recruiting automation.

11. Sources

Every study, law and survey in this guide links to a source below; Flowmingo figures come from Flowmingo's own platform data.

Meld je aan en start in ongeveer 60 seconden. Geen creditcard, geen gesprek

Meld u gratis aan
Flowmingo Editorial Team

Flowmingo Editorial Team

We write practical guides for recruiters and HR teams who want to hire faster and more fairly. Each guide draws on hiring research, employment rules and Flowmingo's own data from real interviews, and lists its sources.

LinkedInYouTubeFacebook