Make interviews flow on their own — without losing the human touch.

1. Problem & Outcomes

1.1 The problem

Most interview processes silently leak:

• Candidates no-show or drop after scheduling.
• Interviewers don’t submit scorecards on time.
• Hiring managers forget to deliver feedback.
• Candidates get inconsistent instructions across stages.
• Recruiters manually chase everyone, all the time.

This creates longer time-to-fill, slower decisions, inconsistent experience, and constant operational drag.

1.2 Outcomes we’re aiming for

• Clean, predictable interview flows from screen → offer.
• Scorecards submitted within agreed-upon SLAs.
• Fewer no-shows and fewer “I never heard back” complaints.
• Standard, role-appropriate candidate communications.
• Recruiters focusing on real conversations instead of inbox herding.

2. Scope & Design Principles

2.1 In-scope

• Phone screen → HM/Tech → Panel → Final → Offer → Pre-start.
• Candidate, interviewer, and hiring manager communications.
• Time-based triggers and SLA logic.
• Email and calendar template library.
• ATS-native automation rules and lightweight orchestration.

2.2 Out-of-scope (v1)

• Automated note-taking or transcription.
• LLM-generated feedback summaries.
• Candidate sourcing or nurture campaigns.
• Post-hire onboarding workflows.

2.3 Design principles

• Candidate-first. Clear expectations, no surprises.
• Low lift. Built on native ATS features wherever possible.
• Observable. Every automation leaves a trail in the ATS timeline.
• Configurable. Stages and templates can vary by role and region.
• Fail-safe. Humans can always override or pause automations.

2.4 Global rules

• Use the candidate’s local time in scheduling messages; show both candidate and host time zones where possible.
• Always include self-serve reschedule/cancel options when supported.
• Ask about accommodations in every invite and include a phone backup.
• Do not include sensitive data in calendar bodies; link to secure docs instead.
• Standardize subject formats: [Company] · [Role] · [Stage].

3. Roles & Responsibilities

4. System Overview

Workflow A[i]gent delivers four major capability areas:

• Scheduling workflows: availability capture, reminders if candidates don’t respond, alerts when threads stall.
• Stage-based workflows: confirmations, reminders, and follow-ups for each interview stage.
• SLAs & escalations: scorecard due dates, automatic nudges, escalation paths when people are late.
• Status & delay updates: structured check-ins when internal timelines slip.

4.1 ATS implementation (conceptual)

Assumes a modern ATS (Greenhouse/Lever/etc.) with:

• Triggers on stage change and event booking.
• Time-based rules (X hours after interview end; Y days in stage).
• Reusable template library with tokens.
• Audit trail (emails/tasks logged on the candidate timeline).

Where native automation falls short, use a thin orchestration layer (Zapier/Make/internal service) via webhooks/APIs.

5. Workflow Library

Each workflow is a reusable pattern: trigger → recipient → template → logging.

5.1 Trigger glossary

• Stage change — triggered whenever a candidate moves into a new stage.
• Event booked — triggered when an interview calendar event is created.
• Time-based — triggered relative to interview start/end or time in stage.
• Manual override — recruiter-triggered for exceptional or sensitive situations.

5.2 Scheduling & availability

Goal: reduce drop-off between “stage change” and “interview booked.”

5.3 Stage workflow pattern

• Confirm — send immediately when the interview is booked.
• Remind — send roughly 24 hours before.
• Reinforce — optional same-day reminder, especially for panels.
• Follow-up — send 2–4 hours after to thank and set expectations.
• Scorecard nudges — to interviewers until feedback is in.

5.4 Scorecard & SLA automation

• +2h after interview — first reminder to submit scorecard.
• +24h after — second reminder plus ping to hiring manager.
• +48h after — escalation to recruiting lead and HM’s manager.
• Optional: block progression to the next stage until required scorecards are complete.

5.5 Status & delay updates

• Structured “we’re still actively interviewing” notes when panels or debriefs take longer than expected.
• Templates for roles on hold, hiring manager travel, or shifting priorities.
• Optional “what’s happening behind the scenes” messages for late-stage or executive candidates.

5.6 Offer & pre-start workflows

• Offer scheduling link or direct coordination from the recruiter.
• Offer call confirmation with agenda.
• Post-acceptance email covering paperwork, IT, and first-day expectations.
• Reminder 1–2 weeks before start with logistics and prep materials.

6. Configuration & Implementation

Typical implementation steps:

• Map current stages → Screen, HM/Tech, Panel, Final, Offer.
• Define SLAs for each stage (scorecards, response times, etc.).
• Create a template library (confirmation, reminder, follow-up, delay, regret).
• Wire triggers using ATS automation features: on stage change, on event booked, on time since interview.
• Pilot on 1–2 roles before expanding.
• Train recruiters, coordinators, and hiring managers on what’s automated vs manual.
• Monitor and tune: no-show rate, SLA adherence, average days per stage.

6.1 Role-based customization

• Sales — heavier prep, expectation setting, panel logistics.
• Engineering — linkage to assessments, code samples, portfolio reviews.
• G&A — fewer stages and faster response expectations.
• Leadership — more stakeholders, tailored messaging, longer cycles.

7. Metrics & Reporting

Suggested metrics:

• No-show rate by stage.
• Average days in each stage.
• Percentage of scorecards submitted within SLA.
• Candidate satisfaction (CSAT / NPS) where collected.
• Template usage rate versus ad-hoc emails.
• Volume of delay updates sent and roles covered.

7.1 Metrics Dictionary alignment

Workflow emits standardized timestamps and event objects that map to the shared Metrics Dictionary fields, enabling time-in-stage, time-to-offer, SLA adherence, and pass-through reporting.

8. Future Enhancements

• LLM-assisted delay updates from structured prompts and guardrails.
• Automatically summarize interviews (future Interview A[i]gent) and attach structured notes to scorecards.
• Dynamically choose templates based on candidate risk (e.g., critical hires, late-stage executive candidates).
• Proactive SLA risk alerts in Slack / Teams when interviews or scorecards are trending late.
• Interview quality analytics across interviewers and stages to inform training and calibration.

9. Fit in the A[i]gents Suite

• Upstream: Screening A[i]gent decides who should move into interviews.
• Core: Workflow A[i]gent orchestrates interview-stage comms, SLAs, and status updates.
• Downstream: Metrics A[i]gent ingests clean events and timestamps to compute time-to-fill, pass-through, and SLA performance.

Together they form an AI-ready, automation-first recruiting operating system.

10. Risks & Mitigations

• Risk: Over-automation makes the process feel cold.
  Mitigation: reserve key moments for human outreach; keep templates warm; allow pause/override.
• Risk: Incorrect triggers send the wrong communication.
  Mitigation: test with dummy candidates; phased rollout; log every automation.
• Risk: Hiring managers ignore SLAs.
  Mitigation: add escalation paths, publish SLA reports, and make compliance visible.
• Risk: Workflows become stale as the business changes.
  Mitigation: review metrics quarterly; adjust templates/SLAs; keep a lightweight change-log.

Appendix A — Template Legend

Token placeholders (candidate name, role title, location, interview date) keep templates reusable across roles and offices.