JustineAIEL · Eve-Legal {F5/reasoner}

JustineAI™ — Employment Class Edition.

Plaintiff-side employment class action and FLSA collectives. Class certification motion practice, damages modeling at class scale, FLSA fluctuating-workweek calculations, opt-in conversion tracking, EEOC charge orchestration.

EL inherits the full Eve-Legal Operating System — case management, deadlines, conflict check, trust accounting, time tracking, e-filing, e-signature, client portal, Microsoft 365 and Filevine integration. The OS layer is shared across every edition; only the legal-reasoning calibration changes.

This edition is not yet generally available. The page describes the workload it will serve and lets employment-class attorneys register interest. We don’t publish dates — we publish the architecture and reach out when the edition is ready.

Justine — the JustineAI™ Digital Employee
Employment-class attorneyOn the roadmap
The EL workflow, as designed

Eight stages. One Justine. The full employment-class lifecycle.

This is the lifecycle JustineAI™ EL is designed to run end to end — the same supervisor pattern that runs personal injury today, recalibrated for employment-class practice. Every stage will be auditable, every output attorney-attested work product.

  1. Stage 01Intake & class definition

    When the edition opens, Justine will frame the putative class from the first attorney conversation — the at-issue policy or pay practice, the proposed class period, the geographic and job-title boundaries, and the limitations clock under the FLSA’s two- or three-year willfulness test. She will draft the class definition for counsel review.

  2. Stage 02Data ingestion

    Justine is designed to ingest the employer’s production at class scale — timekeeping exports, payroll registers, and personnel files — and read the full record in one 10M-token context. She will reconcile clock-in data against pay stubs per workweek, surfacing the unpaid-time pattern that survives across the entire class, not a sampled subset.

  3. Stage 03Liability theory & damages model

    She will build the liability theory — off-the-clock work, misclassification, or meal-and-rest violations — and compute per-member exposure from the production data. Each pay period is modeled, with liquidated damages and prejudgment interest layered correctly, producing a class-wide damages figure defensible at certification.

  4. Stage 04Pre-certification discovery

    Justine is designed to manage Rule 23 discovery — drafting class-list and time-record requests, auditing the employer’s production for completeness, and assembling the declarations of named plaintiffs and comparators. She will flag where the defense will argue individualized inquiry defeats predominance, before the motion is filed.

  5. Stage 05Class-certification motion

    When it opens, Justine will draft the certification brief — Rule 23(a) numerosity, commonality, typicality, and adequacy, plus the (b)(3) predominance and superiority showing, or the FLSA collective’s lenient similarly-situated standard at conditional certification. Every authority will be grounded against CourtListener for attorney attestation.

  6. Stage 06Notice & opt-in management

    Justine is designed to orchestrate notice administration at scale — generating the court-approved notice, tracking distribution and reach, and recording opt-in consent forms against the collective. She will surface response rate and conversion in real time, with the running opt-in count and exhaustion posture reported to class counsel.

  7. Stage 07Merits & expert work

    She will assemble the merits record — representative testimony, the statistical sampling methodology, and the damages expert’s underlying data — and prepare the expert for a Daubert challenge to that methodology. Justine will model fluctuating-workweek versus time-and-a-half exposures so counsel can defend the half-time computation if contested.

  8. Stage 08Settlement & allocation

    When it opens, Justine is designed to build the settlement allocation — per-member shares keyed to workweeks and pay rate, service awards for named plaintiffs, the fee and cost carve-out, and the cy pres residual. She will draft the preliminary-approval papers and the claims-administration plan for the court’s fairness review.

The supervisor patternCalibrated for EL

One Justine. Sub-agents built for employment-class attorneys.

The EL edition will run the same architecture as PI: one Justine coordinating stage-specialized sub-agents across the matter, reasoning in a single context. The sub-agents aren’t separately branded — they’re facets of the Digital Employee. These are the ones this edition is designed around.

  • Class-definition sub-agent

    Class period boundaries, job-title and worksite scope, numerosity proof, the willfulness-driven limitations window, and the commonality theory that ties every member to a single unlawful policy or practice.

    For example — A proposed class of 1,200 non-exempt warehouse associates at three California fulfillment centers. The class-definition sub-agent will draw the boundary — hourly associates clocked under a single rounding policy, March 2022 through filing — fix a three-year FLSA window on a willfulness theory, and frame the common question: whether the employer’s timekeeping rounding systematically undercounted compensable pre-shift security-screening time across every associate.

  • Damages-modeling sub-agent

    Per-member exposure from timekeeping and payroll production, fluctuating-workweek versus time-and-a-half computation, liquidated damages, prejudgment interest, and the statistical sampling methodology defensible at certification.

    For example — Against payroll registers and clock data for the full class, the damages-modeling sub-agent will reconcile each workweek, isolate the average 14 minutes of unpaid screening time per shift, and apply the 1.5x overtime multiplier where the week exceeded forty hours. It will run both half-time and time-and-a-half scenarios, layer liquidated damages and interest, and produce a per-member and class-wide exposure figure with the sampling protocol attached.

  • Certification sub-agent

    Rule 23(a) and (b)(3) elements, the FLSA collective’s similarly-situated standard, predominance and superiority argument, the anticipated individualized-inquiry defense, and authority grounding for the brief.

    For example — Drafting the (b)(3) showing, the certification sub-agent will marshal the rounding policy as the predominant common question and pre-empt the defense’s individualized-inquiry attack by demonstrating that liability turns on the uniform policy, not member-by-member proof. It will assemble named-plaintiff declarations against comparator records, cite the controlling circuit authority grounded on CourtListener, and flag one typicality risk — a lead plaintiff who worked a salaried stint mid-period — for counsel to resolve.

  • Notice & opt-in sub-agent

    Court-approved notice generation, distribution and reach tracking, opt-in consent recording against the collective, response-rate and conversion reporting, EEOC charge timelines, and parallel state-agency exhaustion.

    For example — After conditional certification, the notice & opt-in sub-agent will generate the approved notice, track distribution across 1,200 last-known addresses, and record each consent form against the collective. It will report a running opt-in count and conversion rate to class counsel, flag the members whose EEOC right-to-sue letters are nearing expiry, and reconcile the parallel DFEH exhaustion posture for the state-law Rule 23 claims riding alongside the FLSA collective.

What it will do

The EL workload — described before it ships.

What follows is the architectural commitment for this edition — the workload Eve-Legal F5/reasoner will carry once we open it. The reasoning core is shared with PI. The calibration, taxonomy, and document pipeline are EL-specific.

  • Class certification motion practice

    Rule 23 / FLSA collective certification briefs anchored to your class definition, numerosity proof, commonality of injury, typicality of plaintiff, and adequacy of representation.

  • Damages modeling at class scale

    Per-class-member damages computed from production data (timekeeping, payroll, personnel records) with statistical sampling defensible at certification.

  • FLSA fluctuating-workweek calculations

    Half-time vs time-and-a-half exposures modeled per pay period across the class. Liquidated damages and prejudgment interest layered correctly.

  • Opt-in conversion tracking

    Conditional certification opt-in flows tracked in real time. Reach, response rate, and opt-in plaintiff demographics surfaced for class-counsel reporting.

  • EEOC charge orchestration

    Per-class-member EEOC charge timelines, right-to-sue letters, exhaustion tracking, and the parallel state-agency analog handled at class scale.

The reasoning, imaginedWhat EL will surface

The same depth of reasoning — on employment-class matters.

We proved the reasoning on personal injury. Here is what the same compositional fabric is designed to surface once the EL edition opens — the findings a employment-class practice would want caught, each presented with its evidence for the attorney to weigh.

  • Rounding-policy detection

    Justine will surface — On a typical wage-and-hour matter, Justine will reconcile clock-in punches against paid time across every workweek in the class and surface the systematic asymmetry — rounding that consistently favors the employer on disputed punches — isolating the uniform practice that anchors commonality, with the source rows behind each reconciliation shown for attorney review.

  • Misclassification exposure

    Justine will surface — Reading the full personnel and payroll record, Justine is designed to surface the exempt classification that fails the duties test — assistant managers who spent the majority of each shift on non-exempt tasks — and model the unpaid overtime that follows, per member, across a class period the defense assumed was time-barred until the willfulness evidence reset the clock.

  • Fluctuating-workweek math

    Justine will surface — Justine will model both damages theories side by side: the half-time fluctuating-workweek computation the defense will argue for, and the time-and-a-half method plaintiffs will press — showing the gap between them with liquidated damages and prejudgment interest layered, so counsel can defend the chosen figure at certification.

  • Predominance defense

    Justine will surface — Before the certification motion is filed, Justine is designed to read the matter the way opposing counsel will — flagging the three named plaintiffs whose individual schedules invite an individualized-inquiry argument against (b)(3) predominance, and assembling the comparator declarations and uniform-policy proof that keep the common question dominant over member-specific variation.

  • Opt-in conversion

    Justine will surface — Across a conditionally certified collective, Justine will track notice reach and opt-in conversion in real time — surfacing that response is lagging among a particular shift cohort whose addresses went stale — so class counsel can authorize a reminder mailing before the opt-in window closes, with the running consent count reconciled against the class list.

  • Citation discipline

    Justine will surface — Every statute and decision Justine cites in the certification brief — the controlling similarly-situated standard, the circuit’s predominance authority, the state meal-and-rest cases riding alongside — will be grounded against a verifiable authority on CourtListener, reasoned from real law and never invented, and presented as attorney-attested work product for counsel’s signature.

The AI reasons; the attorney decides.

When it opens

Tell us about your EL practice.

We onboard new editions in waves. When EL opens, founding-firm slots and beta access go to attorneys who told us about their practice early. The form tags your inquiry as EL automatically — you don’t need to repeat the practice area.

Solo and small EL practices: we will reach out the moment self-serve opens. Mid-size and enterprise: we will schedule a discovery call in the wave-prep window before general availability.