A shortlist with reasons.
Hiring decisions, yours.
Hiring on top of your day job? One gives you a small crew of AI bots to read applications against your must-haves, prepare a consistent first phone screen that identifies itself as AI, and draft candidate and hiring-manager updates.
Six applications.
Every one gets a reason.
Reading the must-haves…
Coached support teammates
“Coached four support teammates through weekly case reviews.”Résumé · page 1 ↗
The unread application pile
298–348 inbound applications per hire across startup-size groups.Ashby · 2026 report ↗
Phone screens that won’t fit the calendar
The same first questions, between meetings you already have.
Candidates left waiting
No next step. No update. Another applicant drifts away.
The hiring manager asking for updates
Who’s ready? Who’s stuck? Whose scorecard is still missing?
Spend Monday reviewing people,
not catching up on the pile.
Give the crew the same job-related criteria you would use. Keep the decisions in your hands.
- TodayRead the pile between meetings.
- With Review recommendations with résumé quotes and written gaps.
- TodayRepeat the same first-screen questions.
- With Test one disclosed AI screen; inspect its answers and evidence.
- TodayRemember who still needs an answer.
- With Review a candidate-message draft and the list waiting for your decision.
- TodayRebuild the pipeline story every Monday.
- With Read a scheduled update with stages, blockers and next actions.
The reading, the screens,
and someone to set it up.
Example roles. The Architect proposes a crew around your work; you review the setup.
Alex
Applications & follow-upReads each résumé against your scorecard. Writes the reasons, the gaps and the next update.
Otto
Calls & conversationsBuilds and tests the first phone screen, with an AI disclosure and quoted answers.
Architect
Your workspace partnerProposes your crew, Candidates table and handoffs for you to confirm.
The applications came in.
What happens by Monday?
Choose a role. Change a must-have. Follow the evidence all the way to your decision.
What does this job need?
Change the criteria used to read this batch.
I’ll quote the evidence for each must-have and flag what needs a person to check.
Library · 2 sources
Job description: use the role’s listed responsibilities.
Interview guide: ask the same core questions; a person makes each hiring decision.
Confirm the table in One before importing your ATS CSV.
Open any recommendation. Read why.
Same questions. Answers you can check.
Fictional Test Chat · Simulated · English · No audio or real call
Hi, I’m an AI assistant helping Hearthway with this role. You can ask for a person at any point. Shall we begin?
Waiting for the test answers
Ready for your Monday review.
No decision made. Review the résumé and screen first.
Every application, with reasons
Screen answers, with quotes
Here’s where this role stands
Candidate follow-up · Draft for your review
Thank you for taking the time to apply to Hearthway. We’re reviewing applications against the role’s requirements. We’ll write again once the hiring manager has decided on next steps.
You send from your ATS. Review the recipient list and use its bulk-send action. If a bot sends through Gmail or Outlook with Ask before sending external email on, each email needs a separate approval card.
Keep the rest of the hiring week moving.
Draft the scorecard
Turn the job description into must-haves and screen questions. Ask the bot to question criteria unrelated to the job.
Read the actual documents
Open text-based PDF résumés on the bot’s computer or attach them in chat. Scanned PDFs need a readable version; there is no OCR.
Keep the role brief close
Put the job description and interview guide in the Library. Open the source passages behind a reply.
Bring the candidate list
Confirm a Candidates table, then import an ATS CSV export. Imports deduplicate on email or phone.
Hand off the next draft
A stage change can wake another bot with the record to draft an invitation. Contact fields arrive masked; the address must come from you or an accessible ATS or email thread.
Prepare the interview loop
Draft a panel brief and follow-up questions from the screen record. Arrange times through a connected calendar in chat.
Nudge for interview feedback
Ask for a reminder draft that names the missing scorecard and the candidate waiting on it.
Write the Monday update
Schedule the pipeline summary in the bot’s chat. Allow it to post through connected Slack if that is where your hiring manager reads it.
Organise reference notes
Summarise the notes you provide against the role’s criteria, with open questions for a person to check.
Bring your ATS.
Keep your hiring decisions there.
Connections give the crew tools, not a new hiring system. Connected apps are shared by your bots in the Space.
ATS · one-click connect
Read the applications you permit. Keep hiring decisions with a person in the ATS.
GreenhouseOne-click connect
Reads candidates and interview stages
ATS & screening · with your key
AshbyConnect with your key
Reads applications and interview schedules
Breezy HRConnect with your key
Works with hiring and people records
HireflixConnect with your key
Works with hiring and people records
Email & calendar · one-click connect
GmailOne-click connect
Finds messages and drafts replies
One-click connect
Reads mail and drafts replies
Google CalendarOne-click connect
Reads availability and creates events
CalendlyOne-click connect
Reads bookings and availability
Cal.comOne-click connect
Reads schedules and meeting details
Video & chat · one-click connect
ZoomOne-click connect
Works with meetings and recordings
Google MeetOne-click connect
Reads schedules and meeting details
One-click connect
Reads schedules and meeting details
SlackOne-click connect
Finds messages and posts updates
Also: Built into One. Per-bot chat; ships with the next release.
The separate per-bot chat connection ships with the next release.
Forms, sheets & role documents · one-click connect
TypeformOne-click connect
Works with forms and signed documents
Google SheetsOne-click connect
Reads and updates spreadsheet rows
AirtableOne-click connect
Finds and works with workspace records
NotionOne-click connect
Finds pages and updates databases
Google DriveOne-click connect
Finds and reads shared files
Forms · with your key
JotformConnect with your key
Works with forms and signed documents
ATS, HR & offers · setup required
These need a one-time configuration that the hosted service lacks today.
LeverNeeds a one-time setup
Works with hiring and people records
WorkableNeeds a one-time setup
Works with hiring and people records
RecruiteeNeeds a one-time setup
Works with hiring and people records
BambooHRNeeds a one-time setup
Works with hiring and people records
GustoNeeds a one-time setup
Works with hiring and people records
PersonioNeeds a one-time setup
Works with hiring and people records
WorkdayNeeds a one-time setup
Works with hiring and people records
DocusignNeeds a one-time setup
Works with envelopes and signatures
A reason to review.
A person to decide.
Set the workflow: the crew writes recommendations in One. You make every hiring decision in your ATS. There is no ATS-specific approval setting that enforces this boundary.
People choose who moves forward.
Check the résumé, the screen and any gaps. One’s Terms require human review for employment decisions.
Read the Terms →“I’m an AI assistant.”
Turn on Start with a disclosure. If a candidate asks for a person, record a callback request and have a person follow up. No live transfer.
Use criteria tied to the job.
Review the must-haves yourself. This workflow uses application evidence and answers, with no emotion, face or voice-trait analysis. A bot’s suggestion to rethink a criterion is not a bias audit.
Keep the words behind the result.
“Coached four support teammates through weekly case reviews.”
Ask for résumé quotes in the written recommendations. Screen records keep per-field quotes, confidence and a History.
Set the limits before calling.
Calls a bot places ask for approval by default. Calling handoffs are rules you confirm once; their calls then run within the limits.
Review before sending email
To: candidate’s existing email thread
Subject: Your application update
Turn on Ask before sending external email. One email, one card.
Illustration only. Nothing is sent.
Employers remain responsible for the legal duties that apply where they hire, including notices, assessments, discrimination law, records, and consent for automated calls or recordings. One provides disclosure settings, criteria you set and records with reasons; keep decisions with people. These controls are not a claim of legal compliance or a bias audit. Not legal advice.
One role. One batch.
Your review.
- 01
Start with one open role
Share a readable job description, the interview guide and your job-related must-haves.
- 02
Review a small batch
Confirm the Candidates table. Enable the tools needed to read files, then check the first recommendations against the original résumés.
- 03
Test before candidates hear it
Try the screen in Test Agent where the calling engine is available. Turn on Ask before sending external email and keep ATS decisions with a person.
What stays with you.
Does One decide who gets hired or rejected?
In this workflow, bots write recommendations and reasons in One’s Candidates records. You decide and make the change in your ATS. This is a workflow you set, not a technical lock: a bot connected to your ATS could change a candidate’s status if instructed. One’s Terms require the human review, notices and assessments the law requires.
What does it look at, and can candidates find out why?
Give the bot the readable applications, your job description and a job-related scorecard. Ask it to quote the résumé behind each recommendation and explain missing evidence. Voice-agent Outputs keep per-field quotes, confidence and Pass/Fail criteria from the screen. A person can review those reasons and explain them to a candidate; there is no candidate explanation portal. Missing evidence is a question to check, not proof that someone lacks a skill.
Does the phone screen say it is an AI?
Turn on Start with a disclosure and write the greeting before testing. Our example starts by saying it is an AI assistant. Calls run in English only. One does not analyse emotion, faces or voice traits in this workflow.
Can candidates ask for a person?
Yes. Build the screen to record a callback request and end when someone wants a person. A person follows up. The voice agent cannot transfer the call to a human.
Does it work with Greenhouse, Ashby or Lever?
Greenhouse is a one-click catalogue connection; Ashby connects with your key. Lever, Workable, Recruitee, BambooHR, Gusto, Personio, Workday and DocuSign need a one-time setup that the hosted service lacks today. These are connected-app tools, not native ATS integrations. Availability depends on the integration catalogue and the account’s permissions. An ATS CSV export is another way to bring in records.
Can it schedule interviews?
A bot can work with connected Google Calendar, Calendly or Cal.com in its text conversation. A phone agent cannot book a calendar slot during a call; it can record a callback request for follow-up.
Can it source on LinkedIn?
No. This recruiting workflow does not source on LinkedIn or automate LinkedIn browsing, messages or any other LinkedIn activity. Start with applications you already have.
Can we try the screen before any candidate hears it?
Yes, where the calling engine is available. Test Agent supports browser audio, manual chat and a simulated caller. Tests do not dial a carrier, and test records never trigger a real call. This page is a scripted, fictional preview; it uses no microphone and collects no candidate data.
Can One place live phone screens today?
Live phone calls are not enabled on the hosted service. Calling requires a deployment with the calling add-on, a carrier account and your numbers. Calls a bot places ask for approval by default. A calling handoff is a rule you confirm once; subsequent calls use its daily cap, cooldown, calling hours and Do not call list, without a per-call card.
How do candidate messages get sent?
Bots normally act without asking. Turn on Ask before sending external email for Gmail or Outlook: each email needs its own approval card. Always allow this tool removes that gate for the tool. For a large batch, review the message and recipient list and use your ATS’s own bulk-send action yourself. Prepare drafts only removes every tool, including file reading; use it for text-only drafting, not a bot that must read résumés on its computer.
Where does candidate data go?
It depends on your deployment, model provider and the connected tools you use. Contact email and phone fields in Records are masked in what models read; that does not redact the résumé files you supply. Share only appropriate candidate information and review the data paths, retention and provider details on our Trust page.
Read about data and trust →