AI Transparency Statement

Last updated: July 3, 2026

RecruiterPM LLC (d/b/a HRTool.ai) publishes this statement as a developer of high-risk artificial intelligence systems under the Colorado Artificial Intelligence Act (SB 24-205, effective June 30, 2026).

AI systems covered by the Colorado AI Act

HRTool.ai's AI does not make hiring decisions. It analyzes interviews and assessment answers and provides transcripts, summaries, and job-relevant analysis to hiring managers, who make every decision. The Colorado AI Act nonetheless applies the statutory term "high-risk artificial intelligence system" to any AI whose output is provided to a human decision-maker and can inform a consequential decision such as hiring — the label reflects the importance of the decision being supported, not the safety or quality of the system, and applies even to analysis-and-summary tools like ours. Under that definition, the following HRTool.ai features are covered:

• AI transcription and analysis of recorded assessment answers — transcribes candidate audio/video answers and produces job-relevant summaries and scores against an anchored rubric.
• Live interview AI assistance — real-time transcription of live video interviews, suggested follow-up questions for interviewers, and analysis of candidate answers against a rubric.

AI assessment generation (creating draft questions from a role description) is authoring support used before candidate evaluation and is generally not a covered system; we note it here for completeness.

How our AI works

Recorded answers are first converted to text by a speech-to-text service. The transcript — and only the transcript — is then analyzed by a large language model against an anchored, job-relevant rubric to produce a summary and score of the answer's content and structure. Live interviews follow the same pattern: the live transcript grounds the AI's suggested follow-up questions and answer analyses. The AI never sees or evaluates video imagery, facial expressions, appearance, or vocal characteristics, and we do not use Customer content or candidate data to train AI models — our AI providers are contractually prohibited from doing so. AI transcripts can contain errors, which is one reason every output is decision support for a human reviewer rather than a decision.

How we manage risks of algorithmic discrimination

All candidate-evaluation AI is decision support only; a human makes every final hiring decision. Model instructions constrain the AI to evaluate the job-relevant content of answers and prohibit evaluation of, or inference from, voice, accent, tone, appearance, demeanor, emotion, or any protected characteristic; protected-attribute filtering is applied to live interviewer suggestions, and geographic (e.g., zip-code) proxies are not used. Transcript-quality gating prevents scoring from unreliable transcripts, and candidates may request human review of any AI-assisted evaluation. Candidates may also decline AI analysis before an assessment or live interview and receive equivalent human review against the same rubric; a recorded opt-out is enforced at every point where analysis could run, so no AI transcription, analysis, or scoring occurs for that submission. Consent, notice, opt-out, and evaluation records are logged to support deployer impact assessments and audits.

Information for deployers

On request, we provide deployers documentation describing the intended uses and known limitations of each AI feature, the types of data processed, the measures above, and information reasonably necessary to complete Colorado AI Act impact assessments. Contact support@hrtool.ai.

Incident disclosure

If we discover, or a deployer credibly reports, that an HRTool.ai AI system has caused or is reasonably likely to have caused algorithmic discrimination, we will notify affected deployers and the Colorado Attorney General as required by the Act.