AI candidate screening that starts the moment someone says they are interested
AI candidate screening checks interested candidates before anyone spends interview time on them. RecruitingBot sends knockout and pre-screening questions, runs a short auto scored skills test and combines the answers with the CV score into a ranked shortlist your team approves.
- GDPR built in
- Human approves every shortlist
- No hiring fees
Interest is where screening really begins
A CV tells you what someone has done. It cannot tell you whether they would accept the salary band, work the hours, start within your timeline or solve the problems this team actually has. Recruiters usually learn those things on a 30 minute phone call, and too many of those calls end with "this is not going to work" within the first five minutes. That is the time AI candidate screening gives back.
In RecruitingBot, screening is triggered by a reply. When the reply handler classifies a message as interested, the candidate receives a short set of questions written for that role, followed by a skills test if the role has one. People who replied "not now" are kept for later, and people with questions get answers drawn from the role's facts first. Nobody is screened before they have said yes.
Knockout questions that save both sides time
Knockout questions cover hard requirements that a CV rarely confirms: legal right to work in a country, willingness to be onsite three days a week, a required license, an earliest start date, or a salary expectation inside the band. Each knockout has an expected answer set by you. A candidate who answers outside it is marked as not meeting that requirement, and the record shows exactly which question and which answer caused it.
Keep the list short and genuinely binary. Three or four knockouts are plenty, and each should be something you would truly not move on. If you would bend on a requirement for an exceptional person, it belongs in pre-screening instead, where it informs the ranking without ending the conversation. A person on your team still decides what happens to anyone who misses a knockout, so an honest edge case can be pulled back in.
| Question | Type | Expected answer |
|---|---|---|
| Are you authorized to work in the United States without sponsorship? | Knockout | Yes |
| Can you work from our Denver office on Tuesdays and Thursdays? | Knockout | Yes |
| What base salary range are you targeting? | Pre-screening | Compared with the role band |
| How much notice do you need to give your current employer? | Pre-screening | Recorded for planning |
| Describe a quota you missed and what you changed afterwards. | Pre-screening | Read for substance and ownership |
Pre-screening answers read for substance
Open questions are where good candidates separate themselves, and they are also what recruiters most often skip because reading forty written answers takes an afternoon. RecruitingBot reads each answer against the role rubric and notes which criteria it supports. An answer describing a concrete situation, an action the candidate personally took and a measurable result reads very differently from three sentences of general enthusiasm.
Candidates sometimes ask a question back instead of answering. The reply handler answers within the facts you entered for the role, such as the salary band, location policy, team size and interview steps. When a question goes beyond those facts, it is left for a person on your team to answer, and you can take over any conversation at any moment from the inbox. The candidate then simply continues the thread with a person, with the full history visible to whoever picks it up.
- Specific situation. The answer names a project, account or problem rather than a general habit.
- Personal ownership. The candidate explains what they did, not only what the team achieved.
- Result and reflection. There is an outcome and some sign the person learned from it.
- Fit with the facts. Notice period, salary and location answers are checked against what the role offers.
Skills tests that score themselves
For many roles, a short practical test says more than an hour of conversation. RecruitingBot generates a role specific test of multiple choice and short answer questions from the rubric, which you can review and edit before it goes out. A data analyst might get a query to interpret and a chart to critique. An SDR might get a cold email to improve and a discovery question to sequence.
Multiple choice answers are scored instantly. Short answers are scored against a model answer and the criteria behind it, with the reasoning shown next to the score so a reviewer can disagree quickly. When you edit the test, aim for 15 to 25 minutes in total, because long unpaid assignments drive strong candidates toward employers who respect their time. Monthly skills tests are 50 on Starter, 200 on Growth and 600 on Agency.
Three signals combined into one ranking
The shortlist is where screening pays off. Each candidate arrives with three pieces of evidence: the CV or profile score against the rubric, the knockout and pre-screening answers, and the skills test result. RecruitingBot ranks people using all three, and the shortlist card shows each part separately so nobody has to trust a single blended number. A strong test can lift a modest CV, and a polished CV with a weak test drops.
How the CV score itself is built, with weighted must haves, nice to haves and deal breakers, is explained on the automated resume screening page. On the shortlist, the hiring manager sees the ranked list through a shareable link, opens any card for the full evidence, and approves or rejects. Approved candidates move to scheduling, where they pick a slot from the manager's connected calendar.
| Signal | What it answers | Where it comes from |
|---|---|---|
| CV and profile score | Has this person done comparable work | Rubric analysis of the CV or public profile |
| Screening answers | Can this person take this job on these terms | Knockout and pre-screening questions after interest |
| Skills test result | Can this person do the core tasks today | Short auto scored test built from the rubric |
Candidate experience during screening
Screening only works if good people finish it. Every message comes from your own connected mailbox, signed honestly, inside a sending window of 8:00 to 20:00 in the candidate's local time. Send questions as one short batch rather than a drip of separate emails, and say up front how long the test takes and what happens next. Reminders go out if someone stalls, and stop when they finish or decline.
People who do not move forward receive a rejection email once a person on your team marks them as rejected, rather than silence. Every message carries a one click opt out, and data retention follows the workspace rules described on the GDPR compliant candidate sourcing page. Respectful screening protects your employer brand, which matters most in small markets where candidates talk to each other.
Where AI screening ends and your interview starts
AI candidate screening is good at the checkable parts: requirements, logistics, written reasoning and core skills. It is not a replacement for the interview, where you test judgment, collaboration, motivation and the specific chemistry of a team. RecruitingBot hands you fewer, better prepared conversations rather than making the hire for you, and it never makes an automated final decision about anyone.
That boundary also keeps the process defensible. Each shortlist is approved by a person, every answer and score is visible, and candidates can ask for their data to be exported or deleted. For a view of the full process from intake to offer, including sourcing and outreach before screening, see AI hiring software, which covers how each stage hands its evidence to the next.
Questions buyers ask
Anything else goes to support@recruitingbot.com or the help panel inside your dashboard.
What is the difference between AI candidate screening and resume screening?
Resume screening scores a CV or profile before anyone is contacted. AI candidate screening happens after someone shows interest and checks what a CV cannot show, such as availability, salary fit, written reasoning and practical skills. RecruitingBot does both and keeps the results side by side on the shortlist card.
Do candidates know they are answering automated questions?
The questions come from your mailbox under your company's name, and the process is described plainly in the messages. Candidates can reply with questions at any time and receive answers based on the role facts, and anyone on your team can take over a conversation personally whenever they want.
Can a candidate be rejected automatically for failing a knockout?
No. A missed knockout is recorded on the candidate with the question and answer, and the person moves out of the ranked list. A rejection email is sent only after someone on your team marks them as rejected, which leaves room to reconsider an edge case.
How long should a skills test be?
Short. A test of 15 to 25 minutes covering the two or three tasks the role does every week tells you most of what you need. Longer assignments lose strong candidates who have other options. You can edit the generated questions, remove any and adjust the model answers before the test goes out.
Which roles benefit most from screening questions and tests?
Knowledge work roles where core skills can be demonstrated in writing or through short exercises, such as engineering, data, product, design, sales, marketing, customer success, operations, finance, HR and support. Roles that depend on physical work or licensed practice are better served by other methods.
Do skills tests count against my plan limits?
Yes. Each test sent counts toward the monthly skills test limit of your plan, 50 on Starter, 200 on Growth and 600 on Agency. Limits count consumption in the billing period, so deleting a candidate later does not return a test that was already used.
Keep reading
Give it a role and see the shortlist it builds
Type a job title on the homepage. The demo runs instantly without an account and shows an example pipeline, job post and shortlist for that role.