The Lever automation playbook for talent teams

Most mid-market talent teams we work with bought Lever for the right reason. They wanted an ATS that did not pretend candidates only come through inbound applications. They wanted a single record that connected the LinkedIn message from two years ago to the application that landed yesterday. They wanted nurture campaigns and structured interviews without buying three separate tools to bolt onto a generic ATS.

Then they spent eighteen months using Lever like an applicant database. New requisition opens, job posts go out, applications flow in, recruiters scan resumes manually, hiring managers complain about candidate quality, and the CRM half of the platform sits unused. The Apollo sourcing list lives in a separate Google Sheet. The screening rubric lives in Notion. The handoff to BambooHR is a recruiter retyping data on offer-accept day. Lever becomes a system of record instead of an operational layer.

The fix is not more training on Lever's UI. It is the automation layer around it. Below are the four plays we build most often for Heads of Talent and VP People at companies hiring 30-100 roles a year. Each is concrete, names the adjacent tools, and is something we have shipped, not a thought experiment.

The Gap Most Lever Customers Have

Before the plays, the symptoms. If you recognise three or more of these, the gap is real and the automation layer is missing:

• Your sourcers are exporting CSVs from LinkedIn Recruiter or Apollo and uploading them into Lever by hand, losing source attribution along the way.
• Recruiters open 80+ applications per role and skim manually. The hiring rubric exists in a doc but no one is scoring against it consistently.
• Hiring managers complain that interview feedback is missing or arrives a week late, and decisions stall because no one has aggregated scorecards.
• At offer-accept, someone retypes name, role, comp, manager, and start date into BambooHR or Rippling. The HRIS and the ATS do not talk.
• Your weekly pipeline review runs off a Google Sheet that someone updates Monday morning. Lever has the data, but no one has built the dashboard.

Automation Plays We Build with Lever

1. Outbound Sourcing Pipeline: Apollo plus Clay into Lever

Trigger: a sourcer builds a list in LinkedIn Recruiter or Apollo for an open req. Workflow: the list flows into Clay, where each prospect gets enriched with verified email, current company tenure, recent job changes, and a fit score against the role's must-haves. Anything that passes a threshold is pushed into Lever as a candidate, tagged with the source (Apollo, LinkedIn, referral), the campaign, and the recruiter who owns the req. The candidate auto-enrols in Lever's nurture sequence for that role, typically a three-touch cadence over ten days with a personalised opener built from the enrichment data.

Outcome: sourcers stop spending half their week on CSV uploads and copy-paste. Source attribution is clean from day one, which means your source-of-hire dashboard later actually means something. Reply rates on outbound typically land somewhere between 15 and 30 percent depending on role and seniority, indicative, not promised.

2. AI Rubric Screening on Inbound Applications

Trigger: a new application lands in Lever. Workflow: the application's resume and cover letter get scored against the role's hiring rubric by an LLM using the exact criteria the hiring manager signed off on, years of relevant experience, must-have skills, nice-to-haves, location, comp expectations. The score and a one-paragraph rationale write back to Lever as a structured note on the candidate profile. Applications scoring above the threshold (typically the top 10-15 percent) auto-advance to the recruiter review stage; the rest stay in the applied stage with the score visible so the recruiter can sanity-check edge cases.

Outcome: a recruiter reviewing 200 applicants now reviews 25, and every score has a documented rationale that aligns with the rubric. Hiring managers stop receiving 30-person shortlists. Time-to-first-recruiter-screen usually drops from days to hours. Combine this with Lever's native AI Screening and Candidate Insights and the screening layer becomes genuinely scalable.

3. Structured Interview Feedback Enforcement

Trigger: an interview wraps in Lever. Workflow: each panellist gets a Slack DM with a direct link to their scorecard and a 24-hour SLA. If the scorecard is not submitted in 24 hours, a reminder fires; if not in 48 hours, the recruiter and the panellist's manager are notified. Hire/no-hire decisions in Lever are blocked until every panellist has submitted a complete scorecard. Once they all land, an aggregated decision summary (average score per dimension, dissenting votes flagged, comments collated) gets written back to the candidate profile and posted to a dedicated Slack channel for the hiring committee.

Outcome: feedback latency drops sharply, usually from a week to under 48 hours. Decisions get made on evidence rather than recall. The hiring committee debates the scorecards, not the gut feel of whoever spoke loudest.

4. Offer-Accept to HRIS Sync

Trigger: a candidate moves to Hired in Lever. Workflow: structured candidate data (legal name, role, level, comp, manager, team, start date, location, employment type) flows into BambooHR, Rippling, or Workday automatically. A new-hire record is created, the laptop request fires to IT, the welcome sequence goes out, and the manager gets a Slack DM with the onboarding checklist. If anything fails (a missing field, a comp band mismatch), the recruiter gets a specific error rather than the whole thing silently breaking.

Outcome: the handoff from talent to people ops takes minutes instead of hours, and no one retypes anything. The onboarding clock starts on offer-accept day, not the first Monday someone gets around to provisioning accounts.

How Lever Should Integrate With Your Stack

Lever's open API and webhook surface is what makes the automation layer possible. The integrations that matter most for mid-market teams:

• Sourcing layer: LinkedIn Recruiter, Apollo, Clay, SeekOut for prospect discovery and enrichment before candidates land in Lever.
• HRIS layer: BambooHR, Rippling, Workday, HiBob for the post-hire handoff. Two-way sync where possible, one-way where the HRIS is the source of truth.
• Assessments: HackerRank, Codility, CodeSignal for engineering roles; Criteria, Plum for behavioural and cognitive. Results post back to Lever as structured fields, not PDF attachments.
• Video and scheduling: Zoom, Google Calendar, Microsoft 365 for interviews; Lever's Easy Book handles the candidate-facing piece without the email back-and-forth.
• Background checks: Checkr, HireRight, Certn triggered automatically when a candidate moves to the offer stage.
• Communication and notifications: Slack for recruiter and panel alerts, Gmail or Outlook for candidate-facing email through Lever's native sync.

What ROI Actually Looks Like

Numbers vary by company size, hiring volume, and motion. The ranges below are indicative, not promised, based on what we typically see across mid-market talent teams after the four plays above ship:

• Sourcer productivity: 2-3x more qualified candidates added to Lever per recruiter per week, because they stop building lists by hand.
• Time from application to recruiter screen: typically drops from 3-5 days to under 24 hours.
• Interview feedback latency: usually falls from 5-7 days to under 48 hours once scorecards are enforced.
• Time-to-hire: compresses by 15-30 percent across the funnel, varies heavily by role seniority and market.
• Recruiter capacity: one recruiter typically owns 30-50 percent more open reqs without quality dropping, because the manual work is gone.

Where Teams Go Wrong

Even teams that invest in Lever automation tend to hit the same problems. Worth flagging the patterns we see most often:

• Automating a broken rubric. If the hiring rubric is vague, AI screening just scales the vagueness. Fix the rubric first, then automate scoring against it.
• Skipping source attribution. If you push candidates in from Apollo without tagging the source, your source-of-hire dashboard six months from now is useless.
• Building HRIS sync without error handling. Comp band mismatches, missing manager fields, mistyped emails, these break silently if you do not catch them. Build the failure path before you build the happy path.
• Over-nurturing passive candidates. Three touches is usually right. Eight touches across multiple campaigns reads as spam and tanks reply rates for future roles.
• Treating Lever as the only data store. The richest reporting comes from extracting Lever events into a warehouse and joining them with HRIS data, comp data, and performance data. Keep Lever as the operational layer, not the analytics layer.

Where Moonira Comes In

We build the automation layer around Lever so it stops being a system of record and starts being the operational core of how your talent team hires. That means the sourcing pipeline, the screening logic, the interview enforcement, the HRIS sync, and the recruiter dashboards, all wired into one stack, owned by us, monitored continuously. If your team is hiring 30-100 roles a year on Lever and the bottleneck is the manual work between Lever and the rest of your stack, that is the build we do. Talk to us when you are ready to stop using a $40k/year ATS like a shared inbox.