Answers
What hiring looks like now that everyone has an assistant
837 articles on screening, interviewing, assessment design, policy, teams and pipeline. Every claim carries a source.
First reads
- Twelve Interview Questions That Survive an AI-Prepped Candidate
- What Interview Questions Show How a Candidate Works With AI?
- Behavioral Questions Still Work If You Ask for Events, Not Stories
- Send the Questions in Advance and Change What You Score
- A Good Question Is One a Prepared Stranger Cannot Answer
- Build the AI Question From the Role's Own Tasks in an Afternoon
- What AI Resume Screening Does to Your Pile Before You Open It
- Your ATS Rejects Only What Your Rules Tell It To
- Do AI Detectors Work on Interview Transcripts or Written Responses?
- Is It Legal to Reject a Resume on an AI-Detector Result?
- What Replaces an ATS Keyword Filter When Every Resume Matches?
- Should You Skip the Resume Screen and Go Straight to a Work Sample?
7 beats
Where the writing goes
114 articles
Assessment design
Work samples, take-homes and rubrics, and what an exercise still measures once an assistant can produce the deliverable.
Open the beat
110 articles
Screening
Resumes, detectors and keyword screens, and what a document is still evidence of when every one of them is polished.
Open the beat
130 articles
Interviewing
The loop, the panel and the questions, and what an interview round can and cannot show.
Open the beat
89 articles
Policy
Disclosure rules, adverse impact and records, and what a hiring decision has to be able to show. General guidance, not legal advice.
Open the beat
64 articles
Teams
Calibration, panels and month one, and whether the people in the room agree on what they saw.
Open the beat
93 articles
Pipeline
Applications per opening, and what a tripled pool does to a process that was built for a smaller one.
Open the beat
237 articles
Roles
The seats the AI economy is creating: who fills them, where they come from, and what it takes to hire one.
Open the beat
114 articles
Assessment design
- An AI-Use Score That Tracks Polish Is Measuring Polish If the AI-use line agrees with the presentation rating on nearly every candidate, one impression likely produced both. The re-rating test, and the anchors that fix it.
- Public Questions Are Real Practice and a Bad Prediction Company-tagged question lists come from candidate recall, age without saying so, and sit in a corpus models solve at high rates. What that changes about preparing.
- Why a Game-Based Assessment Has No High Score to Chase What a game-based assessment records, why your tier depends on employees you cannot see, and what a practice pack buys when it cannot buy the target.
- Which Roles Absorb a Career Changer Who Doesn't Know Your Field Sort the role's tasks by how long a wrong answer survives before something outside the person catches it. That column, not tool fluency, decides the fit.
- A Review Round Tests What to Check, Not Whether They Can Build Reviewing model-written code measures what to check and when to stop. The round only grades anything if you author the diff and write the key first.
- What a Paste Looks Like to a Coding Assessment, and What a Flag Does Coding platforms say they can record paste activity, tab switches and code-writing patterns. What that produces is a flag for a person to read, not a verdict.
- What Missing Empty and Error States Tell You About a Designer Missing empty, error and loading states in AI-made screens are what the request produced. What separates designers is the list they make before you point at anything.
- Hand a Candidate Five Documents and Ask Which One Cannot Go In A policy question everyone passes tests nothing. Hand a candidate five work items and two tools, ask what each would need before it could be used, and grade the reason.
110 articles
Screening
- An AI Title Tells You What the Last Employer Called the Job An AI job title records what the last employer called the role. Read the bullets for the decision, the check, and what would have gone wrong unchecked.
- A Five-Year Career Break Costs Tool Exposure, Not Judgment A five-year break costs tool exposure, which refreshes in months, and not the judgment the job runs on. How to test the second, and where a silent discount lands.
- Under 50, the Age Gap in AI Use Is Five Points, Not a Generation Among US adults, chatbot use runs 66% at ages 18 to 29 and 61% at 30 to 49. A five-point gap is not a skill difference, and the fix is one piece of evidence from both.
- What a Hackathon Win and a Kaggle Rank Leave Untested A hackathon win and a Kaggle rank are dated, third-party scored and easy to check. Neither tests whether the target was right. One follow-up closes most of the gap.
- The Chatbot Texting You Is the Application, Not the Interview A text-message hiring assistant collects eligibility and availability, applies rules the employer configured, and books a slot. What your answers actually do.
- On a 25-Minute Screen, Collect What You Cannot Judge A screener outside the domain can check whether an AI answer has the shape of a real episode, and take the content bar from the manager who owns it.
- A New-Grad AI Requirement Selects for Access, Not Skill Nearly every graduating student uses AI, so the requirement filters almost nobody. What to ask instead, and where the AI line belongs in a new-grad posting.
- A Professor Saw the Drafts, Not the Deadline A professor saw drafts, revisions and a defense, which is more process than most managers ever see. Ask what they watched happen, not how much AI the student used.
130 articles
Interviewing
- What an Accent Filter Changes, and What It Cannot Fix No study measures an accent filter against a hiring transcript. What is on record is the recognition gap, one vendor's microphone rule, and the US standard on accent.
- An AI Interview Won't Verify Your Resume. It Will Ask. An AI screen elicits rather than verifies, so an inflated resume line can survive it. What the vendors publish, who reads the transcript, and how to correct a claim.
- Which AI Interviews Work on a Phone, and Which Don't Chat and phone-call formats can run on a phone; a proctored or coding round may not. What a dropped connection costs you mid-answer, and how to set the phone up.
- Read a Follow-Up for the Part Your First Answer Skipped An AI interviewer's follow-up is best read as a gap in the first answer: the owner, constraint, number or disagreement. What to give it, and what reaches the employer.
- Talk to an AI Interviewer in Finished Thoughts, Not Monologues A live AI interviewer can end your answer after a short silence. What to do when it cuts in, asks a follow-up, or glitches, and who reads it after.
- When the AI Interviewer Loops, Say So and Keep Answering A repeat or a mishearing is usually a fault, not a test. What to say on the recording, what to send the recruiter, and the human-review routes vendors publish.
- AI Mock Interview Tools Grade Delivery, Not Your Answer What Yoodli, Big Interview and Google's Warmup actually count, what a transcript scorer ignores, and where the check on your content has to come from.
- A ChatGPT-Only History Transfers. Your Stack Is the Ramp. Most of a ChatGPT-only history transfers, because the judgment is the same in every assistant. What changes is your approval step, your index and your permissions.
89 articles
Policy
- Article 4 Reaches the Interviewers Who Operate the Tool Article 4 has applied since 2 February 2025 and lands on whoever operates AI on the employer's behalf. Which interviewers that covers, and what discharging it looks like.
- When the Tool You Banned Is Someone's Assistive Technology A candidate who uses AI to manage a disability is raising a question about what your round measures. What to ask, what never to ask, and what to offer instead.
- Put the AI Expectation in the Role, Not the Offer Letter An offer letter carries title, pay, start date and at-will language. Put the AI expectation in the written role scope and the first performance plan instead.
- When an AI Interview Asks Something It Should Not, Keep the Record A scripted question can't be redirected. Where the ADA draws the pre-offer line, why a graduation year is the same question, and what to save before the tab closes.
- Check What They Decided, Not Whose Project It Was Almost all workplace AI work is collaborative and the interview asked for the first person. Before the offer moves, ask which decisions inside the project were theirs.
- An AI Requirement Answers to the ADEA, Not to the Four-Fifths Rule An AI requirement that leaves out everyone over fifty is an ADEA question, not a Title VII one. The selection-rate math to run, and the defense an employer carries.
- Your AI Requirement Screens Before the First Application Arrives Selection-rate math needs applicants, and a posting requirement removes people before they become one. Test the requirement against the task, then vary the wording.
- Don't Ask a Candidate to Open Their Own AI History A personal AI account holds a former employer's material and the candidate's own health and money questions. Supply the session you want to watch instead.
64 articles
Teams
- The Admin Console Counts Prompts, Not Judgment Admin consoles report active days, prompts submitted and a last activity date. What that settles about an internal applicant, and what to ask for instead.
- The Backfill Req Is a Task Inventory, Not a Portrait Spend the notice period on a task inventory: what the person leaving runs, what breaks when it is wrong, who notices, and what they check before trusting an output.
- Read the AI Tool List Before You Answer the Budget Number A finalist's AI tool ask describes their working method. Ask what each tool replaces before answering the money, then answer from the approved-vendor list.
- Hire for Reading When Nobody Remembers Why the Code Works The condition names the requirement: reading code with no author to ask. What to put in the req, and the four-question exercise that tests it.
- Three Years of Watching Versus One Hour of Evidence Run the internal candidate through the identical exercise, marked by someone who does not manage them, then spend the three years on the episodes no session can stage.
- Mine Eight Rejections for the Bar the Manager Can't Name The manager can sort real candidates even when they cannot state a standard. Pull eight AI rejections, sort them, and read the competency your screen missed.
- The Six Weeks Before Access Are Not Evidence About the Hire A six-week access delay makes a new hire's first weeks unrepresentative. File the request at offer acceptance, plan the wait, and date the first AI-assisted check.
- Retire the Second Reader One Decision Class at a Time Review comes off one class of decisions at a time, on recorded catches rather than on a date, and some classes keep a second reader permanently.
93 articles
Pipeline
- Bring Them Back Only If the Round Can Change a Decision A round added after four candidates finished the loop is worth re-running only where a bad result would stop somebody. The test, the debrief rule, what to record.
- Putting AI in the Title Changes Who Applies, Not How Many An AI token in the title changes who applies more than how many. Put it in the title when the AI system is what the hire owns; otherwise keep the domain title.
- Read Forty Screening Notes to Find the Criterion Nobody Wrote Down Take forty screen decisions, half of them advances, and code each written reason against the criteria list. The unwritten criterion hides in the advances.
- Write the Or-Equivalent Line So a Screener Can Use It A candidate with no AI job history may have worked somewhere the tool never arrived. Name two or three substitutes in the posting, then screen against that list.
- When Everyone Passes, Check the Answer Against the Next Stage A pass rate everyone clears has two causes with opposite fixes. Sort your last fifty screens by what each answer contained, then read what happened at the next stage.
- The Week Your Engineers Spend Grading Is a Design Choice Grader hours come from an uncapped deliverable and an unwritten key, not from the format. Cap the submission, write the key first, split the passes, count the hours.
- A Vague AI Answer Is an Open Question, Not a Rejection Reject at the recruiter screen only on facts another reader can check. A thin AI answer advances with one written open question, and you count who closes it.
- Pay Above the Band for Scope, or Record It as an Exception A band exception for AI capability holds when the scope changes. What to write down, why the posting premiums cannot price one person, and how to keep one rule.
237 articles
Roles
- Who Leads Actuarial AI Transformation When a Regulator Is Watching? Hire an Actuarial AI Transformation Lead who is a credentialed actuary first: someone who can sign the opinion and explain the model to a state examiner.
- What Does an Agent Product Manager Own When the Product Acts on Its Own? An agent product manager owns behavior, not screens: tool access, the autonomy boundary, the escalation policy, and what counts as a correct run. Here is how to hire one.
- An Agent Quality Analyst Is Hired to Distrust Your Eval Scores Your eval suite grades the agent against a rubric your own team wrote. An Agent Quality Analyst reads transcripts, names failure modes, keeps the golden set honest.
- An Agent Training Environment Lead Is Hired for the Grader, Not the Task Anyone can write a task an agent should do. The scarce skill is a grader that survives an agent trying to satisfy it the cheap way, and a suite that stays hard.
- An AI Abuse Investigator Reconstructs the Trail Your Filters Only Sampled Hire an AI abuse investigator to reconstruct misuse operations across accounts, prompts and infrastructure. Anthropic posts the title under Safeguards.
- Who Fills the AI Act Enforcement Officer Seats Opening Right Now? AI Act enforcement officers come from market surveillance, model validation and security research, not policy commentary. The AI Office is hiring the first cohort now.
- Hiring an AI Agent Manager When Nobody Has Five Years of It Assess an AI agent manager on the work: a real exception queue, a task specified badly on purpose, and the story of an agent output that shipped wrong.
- Hiring HVAC Technicians When the Rooftop Unit Diagnoses Itself Hire HVAC technicians who treat an AI diagnosis as a hypothesis and still put a meter on the board. Screen by watching a call, not by asking about software.