Five candidates, five phone screens, and no two-minute recap that misses half of what they said.

Thursday's calendar has a tight screening funnel: five candidates moving to the phone-screen stage, scheduled back-to-back from 9:00 to 11:30. Each screen is a half-hour call. The questions are the same for each candidate — background, technical thinking, what they're looking for in a role — but the answers are all different, and keeping them straight while still listening to each person is the problem the recruiter is hired to solve. Without a recording, the recruiter has thirty seconds of notes between calls. With one, the five recordings sit on the Mac, each with its own full transcript, and comparing candidates becomes a matter of reading, not remembering.

8:55am — the Mac is recording, the first candidate is two minutes away

The Mac starts recording before the video call connects. Zoom doesn't know it's happening and nobody has to approve anything on the candidate's side — the microphone and the system audio start as two separate tracks the moment the recruiter joins the meeting room. This is the same setup the recruiter uses for every call with every candidate.

9:00am — the first candidate, and the first signal of a mismatch

The first candidate comes in with a strong resume, but ten minutes in, it's clear they've misunderstood what the role is. They're talking about their past work in data infrastructure, which is not what the position needs. By the end of the call, it's obvious the mismatch is deeper than a misread job description — they want a senior IC role and the hiring team needs someone who can grow into team leadership within two years. They're both looking for different things, which is useful to know before the hiring manager spends an hour on a technical screen.

9:30am — moving to the second candidate, with context already written

By 9:30, the first candidate's summary is sitting on the Mac. The recruiter has already typed up the decision before the second candidate joins the call at 9:30. Rather than starting the second screen distracted by trying to remember what the first candidate said, the next call gets full attention, because all the first call's content is already captured.

9:30am — the second candidate, and the first real contender

The second candidate has two years in the role they're applying for, at a company that uses the same tech stack the hiring team uses. By the middle of the call, it's clear they understand the work and they're interested in growing into the leadership part of the role — that matches what the job needs. They ask thoughtful questions about the team structure and the growth opportunities, which is a signal that they're already thinking about how they fit into the org. They also ask about remote-work flexibility, and they're willing to move to the area if the team is in-office two days a week. The summary will make the hiring manager's decision straightforward.

10:05am — three candidates down, two to go, and the data is already searchable

By 10:05, three phone screens are done. Three recordings are sitting on the Mac. Three summaries are written. The second candidate is now clearly ahead of the first — not just in the recruiter's memory of the conversation, but in a summary that captures the key decisions and questions. If the hiring manager asks "what did the second candidate say about remote work?" the recruiter doesn't have to rely on memory; they have the summary and, if needed, the exact timestamp where the candidate said it.

10:10am — the third candidate, and the first red flag

The third candidate has the right background, but early in the call it becomes clear they don't actually understand what the product does. They've done a surface-level research, but they're not sure how their skills connect to this specific problem. This is a red flag — not because they can't learn the product, but because it signals they haven't put in the preparation time the hiring team would need from a senior person. The transcript captures the moment this becomes clear, and the summary flags the gap.

10:45am — the fourth candidate, and the one with a different strength

The fourth candidate has less direct experience in the role than the second candidate, but their background in cross-functional work is exactly what this team needs to grow. They've worked with product, design, and customer success, which means they understand how the decisions the engineering team makes ripple out to the rest of the company. This is a different kind of candidate — not a perfect resume match, but a strong cultural and structural fit for the team's next phase.

11:15am — the fifth candidate, and the closing summary

The fifth candidate is solid but not exceptional. They have the skills, they understand the role, they're interested in growth, but they don't bring anything surprising to the conversation. In a vacuum, they're a good hire. In comparison with the second and fourth candidates, they land in the middle.

11:45am — all five screens done, and the comparison is a matter of reading

By 11:45, all five recordings are on the Mac. Five summaries. Five transcripts. The recruiter is not trying to reconstruct what each candidate said from notes that captured maybe 40% of the conversation. Instead, each summary sits as a short document the hiring manager can read in five minutes, and every candidate's exact phrasing on key topics sits in the transcripts, searchable by phrase.

Between screens — comparing candidates without memory

If the hiring team decides to move four of the five candidates into technical screens, the decision isn't made from four different recruiters' interpretations of the same interview. It's made from four transcripts, four summaries, all on the same Mac, all written in the same voice — the candidate's own voice in the transcript, and the recruiter's assessment in the summary.

Thursday afternoon — the narrowing that sticks

By the time the screening round closes, the hiring manager has read five summaries. They've asked the recruiter clarifying questions ("what did they say about remote work?" "did they mention what they didn't like about their last role?") and the recruiter has the exact answer, not a paraphrase. The candidates moving to the technical screen have been compared on what they actually said, not on how well they stuck in the recruiter's memory between calls.

The tone you need from a candidate, vs. the tone you remember

Recruiting is pattern matching. A good screen looks for signals: Does the candidate know what they're talking about? Do they ask good questions back? Do they seem interested in the role or just looking for any job? These signals are easy to spot in the moment but surprisingly easy to misremember. After five calls in one morning, the recruiter's memory of candidate two gets a little thinner. Did candidate two ask about growth opportunities, or did that sound like candidate three? Did candidate four seem overconfident, or was the recruiter just tired by call four?

The transcript doesn't have memory problems. Every question each candidate asked sits there with the timestamp. Every moment where they paused to think about an answer, every moment where they seemed hesitant or confident, it's all in the audio. The recruiter can listen to a specific moment again — not the whole 30 minutes, but the five-second clip where the candidate answered the question about team size. That's a different kind of recall than remembering what tone they had.

The candidate who impresses at 11:15 and looks different at 11:45

Sometimes a candidate seems great in the moment and by the time the recruiter is thinking about them an hour later, something nags. Or the opposite — someone who seemed okay in the call becomes more impressive on reflection. Looking at the summary from an hour later, or two hours later, can change a recruiter's assessment. Having the transcript means the second assessment isn't based on what stuck in memory; it's based on what was actually said.

A recruiter who reads the summary from candidate two at 2:00pm that evening might notice something they missed in the moment — a casual mention of a technology they thought candidate three talked about, or a hesitation where candidate two said they'd done something but actually wasn't sure they had. The exact words are there to read again.

What today would have looked like with a hiring platform that only stores recordings

Many hiring platforms record the call and store the video, but they don't transcribe or summarize. The recruiter has to re-listen to each call to pull quotes for the hiring manager, and comparing five candidates becomes five re-listening sessions. A recruiter with five calls a day across five days is listening to 25 hours of calls a week just to remember what was said.

Searching across a week of interviews

If the recruiting push continues and the same recruiter runs screens every morning, by Friday they've done 25 candidate calls. A hiring manager asks "did anyone mention compensation expectations above $X?" Searching through five transcripts from Thursday turns up exactly who said it and what they said. Doing that across five days of calls — 25 interviews — without a search function means re-listening to all 25 hours or taking a guess about who might have said it.

Comparing two candidates who interviewed a month apart

If one candidate gets rejected in the first round and then reapplies three weeks later, the recruiter has the original transcript sitting on the Mac. A quick read refreshes the original feedback, so the second interview starts from a position of actual context rather than a vague memory that this person interviewed before. If they've made significant progress on a weakness they showed the first time, it's obvious. If they haven't changed, it's also obvious.

What stays on the Mac, and what gets shared

The recordings and transcripts and summaries stay on the Mac — they never leave for a cloud platform to be analyzed or stored. Sharing them with the hiring manager is simple: a text file that summarizes each candidate, emailed before the technical screen. The hiring manager doesn't need to log into a hiring platform; they have the summary in their email, in the format they already read for every other candidate.

Why phone screens haven't changed much, and why the Mac changes them

Phone screening is one of the oldest parts of recruiting. Hiring managers and recruiters have been doing it for decades. The fundamentals — background, fit, growth trajectory — don't change. What does change is whether the screening round produces data or impressions. An impression is what the recruiter remembered. Data is the transcript, the summary, and the searchable record of what each candidate actually said. The difference between "I think Alex was better than Casey" and "Alex asked about team structure three minutes in; Casey asked about it near the end of the call" is the difference between a guess and a fact.

What it costs, what it runs on

Coii AudioNotes is $19, paid once, covering up to three of the recruiter's own Macs, with a 30-day trial needing no card and no account. It runs on macOS 13 Ventura or later, Apple Silicon or Intel. The full argument for a recruiter's notetaker covers the case in depth, and the how-to on searching old meetings walks through the mechanics this page only showed in motion. The comparison with Otter and comparison with Granola cover what each alternative offers beyond today's five screens. For a round-up of private meeting recorders that keep data on the machine rather than in a cloud vault, and the case against meeting bots that join calls as a named participant, both connect to the privacy-first approach this workflow depends on. If comparing candidates by searching their exact words sounds useful, how to tell speakers apart in a transcript covers voice print matching, and how to export a transcript covers sending the candidate's actual words to the hiring team.