Responsible AI & Trust

Behavioral data about people.
We treat it accordingly.

GloCoach processes behavioral evidence that shapes leadership careers. The gravity of that is not lost on us. This page answers the questions your CHRO, legal team, and procurement function will ask — before they ask them.

Our principles

Six commitments we make to every participant and every client.

Explicit consent, always

Every participant explicitly consents before any behavioral data is collected. No assumed consent. No employer-side override. The participant understands what is being captured and why.

Human in the loop — no adverse decisions without review

GloCoach never surfaces an adverse recommendation (demotion, restructuring, dismissal) without a GloCoach consultant reviewing the evidence first. The system augments human judgment. It does not replace it.

Fairness auditing on every output

Every recommendation batch is audited for adverse-impact patterns before surfacing to the client. If a demographic pattern is inconsistent with behavioral evidence, it is flagged and reviewed — not passed through.

Data boundaries — your data stays yours

Behavioral data collected from your organization is used to produce intelligence for your organization. It is not used to train models on behalf of other clients. Data is segmented, access-controlled, and subject to deletion upon contract end.

Evidence transparency — every output has a trace

Every recommendation includes the behavioral evidence that justifies it. The CHRO can see what specific behaviors drove a readiness rating or a succession ranking. No opaque scores. No black-box outputs.

Validity evidence — the science behind the assessments

Assessments are designed against the GloCoach Leadership Model — a validated competency framework refined over seven years of enterprise deployment, following evidence-based psychometric principles.

Consent model
Explicit. Informed. Revocable.

Every participant goes through an explicit consent process before any data is collected. Consent is:

  • Explicit — not assumed, not embedded in employment terms
  • Informed — participant understands what behavioral data is collected, how it is used, and who sees the output
  • Specific — consent is for the specific engagement, not a blanket data processing agreement
  • Revocable — participants can withdraw consent; their data is removed from processing upon withdrawal
Note on employer-initiated assessments: When an employer contracts GloCoach for a leadership assessment, participants are notified directly by GloCoach — not routed through HR — to ensure the consent relationship is with the participant, not the employer.
Fairness & adverse-impact
Audited before every recommendation batch.

The most dangerous AI bias in leadership decisions is not obvious — it is the kind that replicates historical patterns of exclusion at scale. GloCoach takes this seriously:

  • Every recommendation batch is reviewed for demographic pattern anomalies before delivery
  • If a protected characteristic correlates with an adverse recommendation in a way that is not explained by behavioral evidence, the batch is held and reviewed by a GloCoach consultant
  • The GloCoach Leadership Model is reviewed periodically for construct bias — ensuring the competencies assessed reflect effective leadership, not historical demographic norms
  • Adverse-impact documentation is available for procurement and compliance review on request
AI architecture
RAG + prompt orchestration. Not fine-tuned.

Lance uses Retrieval-Augmented Generation (RAG) to retrieve relevant behavioral records from your organization’s evidence base, combined with prompt orchestration to structure the query and produce a structured intelligence output.

  • Not fine-tuned on your data — the base model is not modified using your behavioral data; RAG retrieval is used instead
  • Not trained on internet text — Lance’s behavioral inferences come from your organization’s evidence base, not from generic language model training
  • Every response is traceable — the behavioral records used to produce a recommendation are identifiable and can be reviewed
  • Human review on adverse decisions — no adverse recommendation is surfaced without consultant review of the underlying evidence
GloCoach does not claim Lance is “objective” — no AI system processing behavioral evidence is free from the patterns in that evidence. What GloCoach does claim is that every output is traceable, auditable, and reviewed before it informs a career-shaping decision.
Data governance
Your data. Your boundaries.

Behavioral data is among the most sensitive personal data an organization processes. GloCoach treats it accordingly:

  • Client behavioral data is segmented by organization — no cross-client data use
  • Data is encrypted in transit and at rest
  • Access is role-based and logged — CHRO and CEO only see the outputs authorized for their role
  • Individual behavioral records are not shared with line managers without explicit CHRO authorization
  • Data retention and deletion schedules are specified in the engagement contract
  • On contract end, behavioral data is deleted per the agreed schedule — not retained for model training

Procurement readiness

Everything your procurement team will ask.
Answered before the RFP.

Enterprise HR technology procurement above $100K ACV typically requires a full data governance review. This checklist reflects the standard gates. Items marked pending are in progress for the current compliance cycle.

Data & Privacy

✓ EU GDPR compliant data processing
✓ Explicit participant consent architecture
✓ Data minimization — only behavioral data required for intelligence is collected
✓ Data segmentation — no cross-client data use
✓ Right to deletion — participant and client
● DPIA (Data Protection Impact Assessment) template — available on request

Security

✓ ISO 27001 framework
✓ Encryption in transit and at rest
✓ Role-based access control
✓ Access logging and audit trail
● SOC 2 Type II — in roadmap
● Penetration testing report — available on request

AI Governance

✓ Human-in-the-loop for adverse decisions
✓ Adverse-impact auditing on recommendation batches
✓ Evidence traceability on every output
✓ No use of client data for third-party model training
● EU AI Act readiness assessment — in progress

Quality & Validity

✓ GloCoach Leadership Model — 7+ years of enterprise validation
✓ Competency construct validity documentation
● Published validity study — in development
✓ Adverse-impact audit documentation available on request

Questions about behavioral data governance?

Book a conversation directly with the team. We welcome the hard questions from CHROs and procurement — they are exactly the ones we built this platform to answer.

Book an executive briefing → See the platform