LoginSign Up
← Back to all sparks
S

Spark Hire

HR
Velocity6.3

Spark Hire moved its AI from reviewing what candidates submit to capturing the interview itself.

recruitingai interview analysiscandidate evaluationats hygienehiring workflow
◆Current state
Spark Hire ships across two products, Meet and Recruit, and the last month has been unusually dense. AI Notetaker put a model inside the live interview, generating structured summaries and suggested evaluation notes; role-aligned ratings followed within weeks, using the job description, scorecard and interview questions to rate answers and organise findings into pros, concerns and items for further review. Recruit gained automatic duplicate merging, pre-screen answers that write through to candidate fields, and LinkedIn-driven lead status updates.
◆Where it's heading
The AI work has a clear direction: capture more of the hiring conversation, then reason over it against the role definition. Each release makes the next possible — notetaking produces the transcript, the job description and scorecard supply the criteria, and pre-screen field mapping makes the structured half searchable. The Recruit side is running a parallel data-hygiene arc, since automated evaluation is only as good as the candidate records underneath it.
◆Prediction
Expect the role-aligned rating logic to reach further back into the funnel — screening and shortlisting against the same job-description criteria — and continued work on the record quality that scoring depends on.

◆Recent moves

  1. 6d ago

    Unified navigation

    Meet's navigation was rebuilt: a dedicated Settings menu, a My Jobs section on the Assessments page, Job settings renamed to Assessments & AI, and a product menu for switching between Meet and Recruit. Housekeeping that follows a month of feature additions, and the cross-product menu quietly reflects the two products being sold as one.

  2. 7d ago

    Prioritize candidates with role-aligned AI ratings and takeaways

    AI Video Review now rates answers against the job description, scorecard and interview questions, and organises findings into pros, concerns and further review. A substantial extension of the notetaking capability shipped days earlier rather than a new direction — the transcript existed, this gives it criteria.

  3. 7d ago

    Keep your lead pipeline moving

    LinkedIn InMails sent through Recruiter System Connect now move a lead's outreach status automatically, and candidates can be converted back to leads with their history intact. Pipeline plumbing for the sourcing half of Recruit.

  4. 8d ago

    Spend less time reviewing duplicate applications

    Recruit now merges confirmed duplicate candidates automatically, adds a Detach process for incorrect merges, and renames auto-resolve to cool-down period. Record hygiene of exactly the kind the AI evaluation features depend on.

  5. 11d ago

    AI Notetaker

    ⚡ SPARK

    AI Notetaker sits in the interview itself, producing a structured summary and suggested evaluation notes afterwards. It is the point where the product stops reviewing artifacts candidates submit and starts capturing the live conversation, which is what makes the later role-aligned ratings possible.

  6. 14d ago

    Turn pre-screen questionnaires into actionable candidate data

    Pre-screening answers, including dropdown and single-choice questions, now save straight to candidate fields and can drive knockout rejections, with the results searchable and reportable. It converts questionnaire text into structured data — the same instinct as the AI work, applied without a model.