Skip to content

“I liked her.”

during the interview process

(I also had to fire her)

after a few months on the job

This is what happens when the candidate is prepared and you are not. A JobKit tells you what to ask and what to listen for, so you walk out knowing what you actually heard.

Complete 3–4-level PDF bundle · $29.99 · Immediate download · Use it with any ATS or no ATS

BCUDiscovering JusticeJarvusTransitOpsAPC

The interview is tomorrow. Nobody on the team has hired this role.

A nonprofit is hiring its first Development Director. The session is already on the calendar, and nobody on the team has interviewed for the role before. The night before, the interviewer finds the kit: the questions, the order, and what a credible answer sounds like.

It is the night-before moment every interviewer knows, with a different ending.

Watch the film: one nonprofit’s first Development Director hire, start to finish.

The resume pile got bigger. The signal got worse.

Candidates can produce more tailored applications. Employers answer with more automated filtering and ranking. Candidates optimize for those systems. Employers add more automation.

One machine writes the application. Another machine grades it. Two people wait to learn what happened.

The last honest signal is live, observable human performance.

A conversation. A presentation. A realistic work simulation. A monitored exercise. Something where the team can observe how the candidate thinks, responds, and works in real time.

+45%
LinkedIn applications in one year
11,000
Applications per minute

Reported by The New York Times in June 2025. A later report documented hidden prompts and code used to manipulate AI resume screeners.

The last honest signal is still easy to waste. We call that pretend work.

01

Wrong questions

Testing what someone could learn in week two. Spending time on nice-to-haves. Asking a candidate to describe something that would be better demonstrated live.

02

Wrong delivery

The helpful rephrase and the leading follow-up. The interviewer performs the reasoning, hears it repeated back, and credits the candidate.

03

Laundered gut

Little is written down, then the debrief turns impressions into an official-looking score and the confident voice wins.

Most firing is a hiring mistake with a delay on it.

$100,000
Total impact of one bad hire at 1× salary
2.5×
The research ceiling for senior positions

Source: Center for American Progress and SHRM research on bad-hire costs, modelled on a $100K position at a 1× multiplier.

None of it appears in the first month. It arrives as missed deadlines, a second search, and a team that stops trusting the process.

Lost productivity
$35,000 · 35%
Recruiting replacement
$28,000 · 28%
Team disruption
$18,000 · 18%
Manager oversight
$12,000 · 12%
Brand and HR fallout
$7,000 · 7%

A debrief cannot recover evidence the interviews never produced.

A polished answer can still tell you almost nothing.

Read the question and pick the candidate you would advance. Then see what the interviewer was told to listen for.

The point is not that one answer sounds better. The point is that the standard exists before either candidate speaks.

JobKit Nº 220 · COO · Mid

3Round 3 of 4

Conflict navigation

You're in a planning session where two department heads are talking past each other, both insisting the other doesn't get it. You've been asked to help. What's your first move?

Ask once, as written, then allow silence. A helpful rephrase may hand the candidate the answer.

Two candidates answered. Which one do you advance?

What the kit told the interviewer to listen for

Written before either candidate spoke

Credit

  • Proposes separate conversations before joint problem-solving
  • Checks for unstated goals or pressures
  • Reflects each position back in neutral language
  • Considers timing and setting for the joint discussion

Discount

  • Immediately proposes compromise without diagnosis
  • Sides with one party based on title or relationship
  • Jumps to a framework before understanding the disagreement
  • Assumes bad faith or incompetence

Pick a candidate to mark which indicators that answer actually hit. Then try the other one — same question, same indicators.

The standard existed before either candidate answered. That is what makes A and B comparable, and what makes the rating defensible later.

Question and indicators adapted from JobKit Nº 220. Candidate answers are illustrative.

product interface

Not two products. One standard, carried from the kit into the room.

The PDF is a complete product and works with any ATS, or none. Open the same kit in FirstWho and the questions become interactive forms that capture the rating while the session is fresh.

  1. 01

    JobKit

    The interview is designed before anyone joins the call

    Level-specific questions and live exercises, written for the job in front of you. Every interviewer works from the same design.

  2. 02

    In the room

    The answer is rated against a standard that already existed

    Indicators written before the candidate spoke. The interviewer rates what they heard while it is fresh, not from memory at the debrief.

  3. 03

    FirstWho

    The ratings become a comparable record

    Weighted by what the role actually requires and laid side by side, including what nobody got around to asking.

  4. 04

    The manager

    A decision with evidence under it

    The ranking is a map of the evidence, not a recommendation. The trade-offs stay on the page, because they belong to the manager.

Whatever the team already runs

  • Lever
  • Greenhouse
  • NeoGov
  • Rippling
  • Workable
  • BambooHR
  • Pinpoint
  • JazzHR

No integration to install and nothing to migrate. The kit is the interview, and it travels.

Every question, every round, written for the exact job and level.

Three or four complete guides, one for each level of the job. Use the one that matches the opening in front of you. Save the rest for the next search.

JobKit Nº 1395 · Forward Embedded Software Engineer · Mid · The cover, and the count · p. 1 of 111.

  • Part 2 of the guide, The four behaviours: four numbered habits — ask each question as written and allow silence, record specific evidence from the answer, rate against the level's expectations before the debrief, and repeat the same questions and scale for every candidate — each split into a Do column and an Avoid column.

    01 · How to ask, and how not to

  • At a glance page: scope and decision rights at Mid level, the four things a successful hire produces, a four-round plan with the question count and minutes for each round, five priority factors shaping the role, a split of 25 must-have and 18 nice-to-have requirements, and totals of 450 minutes of panel time and 225 minutes of live candidate time per candidate.

    02 · The scope, and what a hire must produce

  • How the process runs: the order from application through resume screen and four interview rounds to offer, a table giving each live round its purpose, scheduled minutes and question count, a bar showing how 69 question asks distribute across the rounds, and what the process costs each side — 450 minutes of panel time, 225 of live candidate time, and 171 further minutes of candidate work outside the live rounds.

    03 · Four rounds, and the 171 minutes outside them

  • Coverage matrix listing 21 capabilities against the application, resume screen, the four interview rounds and the exercises, with the expected proficiency for each and the number of questions in every cell; the Recruiter Screen column is empty for every capability, and four nice-to-have capabilities rest on a single question.

    04 · Coverage, and the round that assesses nothing

  • Score and decide page, white on navy: a Look for column of eight sample positive indicators beside a Watch out for column of eight caution indicators, drawn from the 207 positive and 207 caution indicators spread across the kit's 52 questions, with a note that written indicators give interviewers a common reference so the debrief argues about evidence rather than impressions.

    05 · 207 to look for, 207 to watch out for

Scroll the spread. Five of the 111 pages of JobKit Nº 1395.

The kit, live in the room.

Everything above fits in a PDF. Open the same kit in FirstWho and it becomes the interview form: the question on screen as written, a field for the gist of the answer, and the written indicators to rate it against, while the answer is still fresh.

Attention stays on the candidate instead of the paperwork.

A candidate answers. The kit holds the question, the standard, and the place to put the rating. Simulation, audio muted.

The interview

  • Ask as written, once. Then allow silence. The delivery guidance travels with the question.
  • The next question is already on the screen. Nobody is inventing it while the candidate is speaking.
  • Notes capture the gist of what was said, while it is fresh. Not from memory at a debrief, not from a transcript days later.

The rating

  • One to five against the written indicators, question by question.
  • Strong Yes, Yes, No, or Strong No, plus the level the candidate demonstrated.
  • Every interviewer rates independently. Nobody sees the room's verdict before entering their own.

The record, written while the answer is fresh.

Ratings land during the session, so the hiring manager is not chasing feedback for a week. FirstWho applies the weights the role requires, set in the kit before anyone met a candidate, and lays the pool side by side.

Forward Deployed Engineer · Senior

Weighted by the capabilities this role requires, set in the kit before anyone met a candidate.

Ratings submitted before debrief
Candidate comparison for Forward Deployed Engineer, Senior
CandidateCustomer discoveryTechnical deliveryLevel demonstratedWeighted
R. OkonjoSenior4.1
A. ChenStaff4.0
M. SilvaUnobservedMid3.1

Calculated from interviewer-entered ratings. Not a hiring recommendation. The manager still has to decide which trade-offs are acceptable.

The scorecard view from a live search: weighted score, capability radar, and the interview clip, per candidate. Demo data, from the live product.

Try one job end-to-end

The level of AI faking in technology interviews is off the charts. FirstWho is a game-changer for hiring managers who want to see through the smoke and mirrors to learn about the full candidate.
Paul Kim's profilePaul Kim · Software Engineer Team Lead, Capital One
FirstWho's approach focuses on the fundamentals of great hiring: knowing who and what you're looking for, and preparing interviewers for honest, authentic conversations that matter.
Jeff Liberty's profileJeff Liberty · CEO, Good Trouble Strategic Advisors
FirstWho provides the tools but also the training, because good hiring decisions depend on people.
Ashli Molina's profileAshli Molina · Staff Technical PM, Transit Tech

These reviewers evaluated the method and materials directly. Their comments validate the approach, not a claimed customer outcome.

AI helps build the kit. It does not sit in the interview for you.

FirstWho uses AI extensively to research and synthesize role-specific interview materials inside a practitioner-designed process. People direct and review the work before a JobKit ships.

During hiring, interviewers evaluate candidates. FirstWho records, calculates, ranks, and visualizes the ratings they enter. It does not decide what an answer deserved or whom the company should hire.

  • No AI interviewer

    No screening bot, chat, or conversational agent conducts the live interaction.

  • No notetaker as a substitute for attention

    A transcript can preserve words. It cannot listen, interpret, or recognize why a follow-up became necessary for the interviewer.

  • Recording is not the issue

    Recording or transcription can be legitimate. The line is using a machine to replace human attention and contemporaneous evaluation.