People & HR

How to run an interview loop as Small HR Teams

People & HRFor Small HR Teams4 apps12 steps~24 min to set up

You're the scheduler, the note-taker, the follow-up chaser, and the debriefer — all in a two-person HR team running hiring for 150 employees. A single engineering hire can involve 6–8 interviewers, four scheduling rounds, a take-home, a debrief, and a Greenhouse update that nobody remembers to do. You're copying candidate names from Greenhouse into calendar invites by hand, texting hiring managers the night before to remind them what to ask, chasing debrief notes that live in three people's inboxes, and building a hiring-stage tracker in a Google Sheet that's already out of date. Calendly helps with the booking link. Nothing helps with everything else.

Build this on Starch → See the apps used

People & HRFor Small HR Teams4 apps12 steps~24 min to set up

Outcome

What you'll set up

A scheduling system that reads your interviewers' Google Calendars and generates candidate-ready booking links for each interview stage — no back-and-forth email required

A meeting notes workflow that transcribes every interview, extracts structured feedback by competency, and drops it into a searchable archive your hiring team can actually find

An interview loop tracker — built as a custom Starch app — that pulls candidate stage data from Greenhouse, surfaces which loops are stalled, and flags roles with no scheduled next step

The Starch recipe

Apps, data, and prompts

The combination of Starch apps, the data sources they pull from, and the prompts you use to drive them.

Apps used

Scheduling Meeting Notes Knowledge Management Email Agent

Data sources & config

Starch connects directly to Google Calendar via scheduled sync to read interviewer availability in real time. Greenhouse is connected from Starch's integration catalog; the agent queries it live when the loop tracker needs current candidate stages. Notion is synced on a schedule so interview playbooks and rubrics are searchable inside Starch. Gmail is connected via scheduled sync so the email triage app can monitor candidate threads and draft replies. Interview recordings are brought in through browser automation — no API needed — if your video conferencing tool doesn't have a direct connection.

Prompts to copy

Build me a scheduling app for our interview loop. It should pull from Google Calendar for each of our five interviewers, let me set stage-specific meeting types (30-min phone screen, 60-min technical, 45-min values), add 15-minute buffers between back-to-back interviews, and generate a single booking link per stage that I can paste into a Greenhouse email template.

Build me a meeting notes app for candidate interviews. After each call, it should transcribe the recording, generate a one-paragraph summary, extract a structured debrief with fields for 'technical signal', 'communication', 'culture fit', and 'hire / no-hire recommendation', and save it to a searchable archive tagged by candidate name and role.

Build me a knowledge base of our interview playbooks. Pull from our Notion pages where we store competency rubrics and role-specific question banks, auto-detect when a rubric hasn't been updated in 90 days, and let any interviewer search 'what should I ask a senior backend engineer?' and get the right guide instantly.

Set up an email triage app for my recruiting inbox. Flag any email from a candidate who has been waiting more than 48 hours for a reply, draft a holding response I can send with one click, and remind me every morning about any interview confirmation that hasn't been acknowledged.

Run these in Starch → or paste them into your favorite agent

Walkthrough

Step-by-step

1 Connect Google Calendar via scheduled sync. Starch pulls availability for all five interviewers on a rolling 3-month window — this is the source of truth for every scheduling decision.

2 Connect Greenhouse from Starch's integration catalog. The agent queries it live to pull current candidate stages, assigned interviewers, and open roles whenever your loop tracker refreshes.

3 Connect Notion via scheduled sync. Your existing competency rubrics, question banks, and offer templates become searchable inside Starch — no copy-paste required.

4 Connect Gmail via scheduled sync. The email triage app monitors your recruiting inbox, flags candidate messages waiting more than 48 hours, and drafts replies so you're not context-switching all day.

5 Install the Scheduling starter app from the Starch App Store and customize it: add your five interviewers, configure stage-specific meeting types (phone screen, technical, values), and set buffer rules. Starch generates a unique booking link per stage.

6 Install the Meeting Notes starter app. Point it at your video conferencing recordings (via browser automation for platforms without a direct connection) and set up structured debrief fields specific to your competency rubrics — hire signal, technical bar, communication, culture fit.

7 Install the Knowledge Management starter app and point it at your Notion playbook pages. Interviewers can now search 'what do I ask a PM candidate about stakeholder management?' instead of pinging you on Slack.

8 Build a custom interview loop tracker: tell Starch 'Build me a dashboard that pulls every open role from Greenhouse, shows each candidate's current stage, the date of their last activity, and flags any candidate with no scheduled next step in 5+ days.' Starch assembles it from your live Greenhouse data.

9 Set up an automation: 'Every weekday morning at 8am, check Greenhouse for any candidate who advanced to a new stage yesterday, generate a scheduling link for the next stage, and draft an email to the candidate with the link.' Review and send with one click.

10 Set up a debrief follow-up automation: 'After each interview is logged in Meeting Notes, send a Slack DM to the interviewer with a link to fill out their structured debrief. If no debrief is submitted within 24 hours, send a reminder.' No more chasing notes manually.

11 Build a weekly hiring summary automation: 'Every Friday at 4pm, pull all candidate activity from Greenhouse this week, count interviews completed, offers extended, and open roles with no movement, and Slack me a summary.' Share it with the CFO and hiring managers.

12 As your playbooks evolve, the Knowledge Management app flags rubrics older than 90 days. Update them in Notion; Starch re-syncs on schedule and the updated version is searchable the next day.

See this running on Starch

Connect your tools, describe what you want, and the agent builds it. Closed beta is free.

Try it on Starch →

Worked example

April 2026 Engineering Hiring Loop — 3 Open Roles

Sample numbers from a real run

Roles being filled	3
Active candidates across all stages	14
Interviews scheduled this week	11
Debrief notes missing (before Starch)	7
Hours spent on scheduling email per week (before)	4
Hours spent on scheduling email per week (after)	1

In April you're filling a senior backend engineer, a product designer, and a customer success manager simultaneously — 14 active candidates, 11 interviews scheduled this week. Before Starch, your Monday morning was: open Greenhouse, find which candidates advanced, open five Google Calendars in separate tabs to find overlap, send three rounds of 'does this time work?' emails, manually confirm in Greenhouse, and text each interviewer a reminder the night before. That was four hours a week, conservatively. With Starch, the Scheduling app reads all five interviewer calendars, you paste the stage-specific booking link into your Greenhouse email template, and candidates self-schedule. The Monday morning automation checks Greenhouse for any candidate who advanced Friday, drafts the scheduling email with the correct stage link, and has it ready in your Gmail drafts at 8am. You review and send in 15 minutes. After each interview, Meeting Notes transcribes the call and surfaces a structured debrief: technical signal (strong — solved the distributed caching problem correctly), communication (clear, asked good clarifying questions), hire/no-hire (hire, pending values interview). The debrief is tagged under 'Senior Backend Engineer — Candidate: Priya S.' and searchable forever. The 24-hour Slack reminder fires to every interviewer who hasn't submitted. Out of 11 interviews this week, you got 10 debriefs within 24 hours. Last month you got 4 out of 11. The loop tracker dashboard shows one senior backend candidate stuck in 'values interview scheduled' for 8 days with no confirmed date — you catch it before the candidate goes cold and sends the scheduling link that afternoon.

Measurement

How you'll know it's working

Time-to-schedule per stage (target: under 24 hours from stage advance to candidate booking confirmed)

Debrief completion rate within 24 hours of interview (target: 90%+)

Candidates with no next step scheduled after 5+ days (goal: zero open roles with stalled pipelines)

Hiring manager response time to debrief requests (tracked via Slack automation)

Interview loop cycle time by role: days from first screen to offer extended

Comparison

What this replaces

The other ways teams handle this today, and how the Starch version compares.

Greenhouse + Calendly + Google Docs for debriefs

Greenhouse tracks stages, Calendly handles booking, but nothing connects them — you're still manually generating links per stage, and debrief notes live in Google Docs nobody searches.

Lever

Lever has built-in scheduling and feedback forms but its forms are rigid — you can't customize debrief structure by role without an admin-level configuration change and it doesn't surface stalled loops proactively.

Rippling + built-in ATS

If you're already on Rippling, its ATS is fine for offer letters and onboarding but thin on interview loop management — no transcript capture, no debrief tracking, no cross-system automation.

Notion + Zapier for the glue

Notion stores the playbooks and Zapier can trigger some notifications, but you're building and maintaining every Zap yourself, there's no natural-language authoring, and it breaks silently when Greenhouse changes an API field.

ModernLoop or Prelude (dedicated interview scheduling tools)

Purpose-built for interview scheduling at scale, but they're an additional point tool — they don't connect to your debrief notes, your Greenhouse stage data, your email triage, or your weekly hiring summary in one place.

On Starch RECOMMENDED

One platform — scheduling, meeting notes, knowledge management all running on connected data. Setup in plain English; numbers stay current via scheduled syncs and live agent queries.

Try it on Starch →

FAQ

Frequently asked questions

We use Greenhouse as our ATS. Does Starch actually connect to it, or do I have to export CSVs?

Greenhouse is available through Starch's integration catalog — no CSV exports. When your interview loop tracker or scheduling automation needs current candidate stage data, the agent queries Greenhouse live. You connect it once from Starch's integration browser.

Our interviewers use a mix of Zoom and Google Meet. Can Starch capture transcripts from both?

Google Meet recordings can be pulled in directly. For Zoom, Starch can automate through your browser — no Zoom API required — to retrieve recordings and pass them to the Meeting Notes app for transcription and debrief extraction.

We store our interview rubrics in Notion. Will those actually be searchable inside Starch?

Yes. Starch syncs your Notion pages on a schedule, so your competency rubrics and question banks become searchable inside the Knowledge Management app. When a rubric hasn't been touched in 90 days, the app flags it for review. You keep editing in Notion; Starch re-syncs automatically.

Is Starch SOC 2 Type II certified? We handle candidate data.

Not yet — Starch is not currently SOC 2 Type II certified. If your company or your candidates require that certification before you can connect recruiting data to a new tool, that's an honest blocker worth knowing upfront.

What if a hiring manager never fills out their debrief? Can Starch actually chase them?

That's exactly what the debrief follow-up automation does. You tell Starch: 'After each interview is logged, send a Slack message to the interviewer with a debrief link. If nothing is submitted in 24 hours, send a reminder.' The automation runs on its own — you don't have to manually track who's missing. You can also set a second reminder at 48 hours if you want.

We're a two-person HR team. Is this going to take weeks to set up?

The Scheduling, Meeting Notes, and Knowledge Management starter apps are available in the Starch App Store — you install them and customize, you don't build from scratch. The custom interview loop tracker (pulling from Greenhouse, flagging stalled candidates) is built by describing what you want in plain English. Most teams have a working version of each piece within a few hours, not weeks.

Can Starch replace Greenhouse entirely?

No, and it shouldn't. Greenhouse handles your ATS workflows — job postings, offer letters, compliance tracking, candidate records. Starch connects to Greenhouse and builds the surfaces around it: the scheduling automation, the debrief tracker, the weekly summary, the stalled-pipeline alert. Think of it as the glue layer, not a replacement.