Readiness Engine Assessment: Method, Data, and Limits

Readiness Engine Assessment

Readiness Engine is an AI-assisted developmental leadership assessment. It analyzes a structured interview or other consented language and returns evidence-linked interpretations about leadership capacity and growth.

The company currently says the instrument is early in validation, developmental, and not validated for selection decisions.

Product at a glance

flowchart LR
    A["Consented video, audio, or transcript"] --> B["Transcription and evidence extraction"]
    B --> C["AI-assisted analysis against rubric"]
    C --> D["Range, dimensions, and evidence"]
    D --> E["Human review when included"]
    E --> F["Developmental report and client use"]

The diagram summarizes public documentation. It does not disclose or independently reconstruct the proprietary instrument.

Construct and methodology

The current methodology page says the product looks for perspective reach, tension handling, evidence discipline, and the ability to coordinate different scales and time horizons.

It grounds those claims in adult-development, action-logic, hierarchical-complexity, dynamic-skill, and integrative-complexity traditions. It says the proprietary components include the interview, questions, scoring anchors, calibration, and reports.

Evidence for those research traditions is not the same as evidence for the product. The instrument needs its own reliability, validity, fairness, version, and outcome record.

Six dimensions

The current site presents strategic complexity, relational and emotional intelligence, team climate, resilience under fire, purposeful agility, and coachability with identity flexibility.

These dimensions can provide a useful coaching vocabulary. A buyer should still ask how each is defined, which excerpts count, how overlap is handled, how raters agree, what range is reported, and what evidence would change the interpretation.

Three service tiers

The April 2026 privacy policy and terms describe three tiers.

The standard tier uses one AI model. The triangulated tier cross-references three models. The human-reviewed tier has a qualified reviewer exercise oversight over the output.

The terms require the human-reviewed tier when a client uses output as a basis for a consequential decision. They also state that the output is developmental guidance and that the company does not warrant accuracy, completeness, or suitability for a particular purpose.

Using more models can reveal disagreement. It does not guarantee truth. Human review can add context. It does not make an unvalidated selection method valid.

Validation boundary

Readiness Engine says it is correlating outputs with Big Five measures, integrative-complexity scoring, partner behavioral analytics, and pilot outcomes. It says results will be published as they become available.

The methodology states that the product is not validated for selection decisions. It is not designed or validated as the sole basis for hiring, promotion, termination, or other employment decisions.

The clearest current use is developmental. A report may generate coaching hypotheses, name a pattern, or structure a conversation when the participant can see and contest the evidence.

Reliability and fairness evidence needed

Public procurement evidence should include inter-rater reliability, test-retest behavior, prompt equivalence, minimum input, model-version stability, uncertainty, reference population, languages, roles, cultures, disability accommodations, subgroup performance, and error analysis.

The Trust and Ethics page says the company audits models for bias and designs scoring to exclude accent, verbal polish, and storytelling style.

These are commitments. Ask for the audit method, auditor, data, groups, sample sizes, metrics, findings, remediation, and monitoring schedule.

Data and subprocessors

The privacy policy describes identity data, consent, assessment responses, video, audio, transcripts, scores, profiles, developmental classifications, device data, usage, and audit trails.

It names Google Cloud Platform for infrastructure, storage, and AI processing; Willo for video interview collection and transcription; Vercel for dashboard hosting; and n8n for workflow automation.

The policy says identifiable inputs and outputs are not used for training without separate consent. It says standard assessment data is retained for 12 months, with possible extensions under client contracts or law. Anonymized research data may be kept indefinitely.

Participant and client control

The policy describes access, correction, deletion, portability, consent withdrawal, contest, and human-review rights. It says the client is a separate controller or business for its own handling.

The terms say withdrawal stops future consent-based processing but does not retract outputs already delivered to a client. That makes the buyer's own governance part of the product risk.

Decision and legal boundary

EEOC employment-test guidance says employers remain responsible for their tests and selection procedures. SIOP's AI assessment recommendations emphasize use-specific validation evidence.

NIST's AI Risk Management Framework adds ongoing governance, measurement, transparency, privacy, and fairness work.

No product document can transfer the buyer's legal, ethical, or decision responsibility to the vendor.

Venture Step conversation

Logan Yonavjak demonstrates the product thesis in episode 104. Read the [[Logan Yonavjak on Measuring Founder Readiness|episode story]], [[Can Coachability Be Measured|coachability analysis]], and [[How to Evaluate an AI Leadership Assessment|assessment buyer guide]].

Editorial and verification notes

Venture Step has not purchased, scored, security-tested, audited, or independently validated Readiness Engine. Dalton completed an earlier assessment but his result was not available in the recording and is not inferred.

Recheck methodology, tiers, models, human review, validation studies, sample populations, subgroup results, privacy, retention, subprocessors, security, participant rights, client controls, and regulatory status before publication or procurement.

AI assisted with research organization and drafting. Dalton Anderson remains responsible for the product boundary and publication decision.