Babylist · Labor planning and insights

Foundation and
build scope

What we agree before kickoff, and what we decide after the foundation findings. October 2026.

Foundation objective

Prepare priority labor data for one fulfillment center.

Set an initial baseline, and work out what the first application release can feasibly do.

Boundary To approveProposed
Site and periodOne fulfillment center; up to 12 consecutive months, ending at an agreed cutoff
Dataset familiesUp to three: workforce roster and employment events; warehouse output and labor hours; staffing or volume forecast
Source reviewInventory of Paylocity, Lattice, Korber, Greenhouse, Excel or Sigma, and Aerotek or UKG requirements. Not a commitment to ingest every system.
Working locationOne Babylist-approved project data room; original inputs preserved
PreparationNormalize dates, IDs and labels; resolve deterministic duplicates; document missing values; join only on reliable keys
AnalysisBaseline up to five measures; investigate up to three priority questions
Access testingTest approved access paths; record missing fields, approvals, refresh limits and fallbacks
Time and fee1–2 weeks from usable inputs and access; $5,000 proposed
Babylist · Foundation and build scopeTerminal AI
Foundation02
Before kickoff

Decisions to confirm with Babylist To confirm

01Data scope
Site, history, source reports, fields and business owners
02Questions and measures
Three priority questions, five baseline measures, agency/direct linkage and the source of truth for each metric
03Data handling
Approved storage, authorized users, and limits on employee data and AI processing
Outputs and acceptance

What you receive, and how it is accepted

OutputAccepted when
01Organized project data roomAgreed inputs and outputs present, labeled and accessible to named owners. Source versions and date coverage recorded.
02Prepared dataset and recordCleaned tables, data dictionary and repeatable preparation steps. Lineage, row counts and open issues recorded.
03Metric definitions and baselineFormulas, periods and filters for up to five measures, reconciled with reference reports or differences explained.
04Initial findingsUp to three questions with figures, limits and next actions. Unsupported questions documented, not concluded.
05Build recommendationCapabilities, sources, refresh, dependencies, access, acceptance tests, schedule and a fixed price or revised estimate.

The prepared dataset is a working analytical dataset, not an automated production pipeline.

Babylist · Foundation and build scopeTerminal AI
Foundation03
Responsibilities

Who does what in the foundation

Terminal AI
  • Prepares the data
  • Documents data issues
  • Documents the findings
  • Writes the build recommendation
Together
  • Metric definitions, reviewed by both teams
  • The baseline, reconciled with Babylist reports
Babylist provides
  • Approved data and access
  • A sponsor and an operations owner
  • HR and Finance definitions
  • IT, reference reports, agency feed

Agency data is included only where Babylist authorizes it.

Not included

What the foundation does not cover

Data and systems

  • A production application
  • Ongoing automated refresh
  • A full enterprise warehouse
  • Source-system repair
  • Company-wide data cleanup
  • Rebuilding missing history
  • Unsupported custom extraction

People and findings

  • Probabilistic employee matching
  • Individual attrition scoring
  • Automated employment decisions
  • A guaranteed finding
Babylist · Foundation and build scopeTerminal AI
After the foundation04
Deployment scope

Decided with the findings To complete

Completed and approved after the foundation, before the build starts.

DecisionWhat to define
01Users and workflowsPilot site, user groups, role permissions, decisions and tasks supported
02Data and refreshSystems, reports, fields, history, feed type, frequency, failure handling
03DashboardsNamed views, measures, filters and comparisons, each measure defined
04Retention and planningCohorts, horizon, volume and productivity assumptions, attrition method
05Agency and directAllocation rules, lead times, costs, target mix, a named human owner
06AlertsTriggers, thresholds, recipients, channel, frequency and expected action
07Ask or other agentsQuestion types, data access, evidence shown, limits, permitted actions
08Deployment and securityHosting owner, sign-in, permissions, model provider, retention, monitoring
09Acceptance and handoffReference data, tolerances, example answers, tests, training and docs
10Commercial agreementInclusions, dependencies, milestones, fee, payment, stabilization, costs
Commercial

The $20,000 deployment allowance is indicative until this scope is approved.

Next step

Approve the foundation · $5,000 proposed. Confirm site, history and source reports; name the sponsor and source owners; agree questions, measures and data limits.

Babylist · Foundation and build scopeTerminal AI