> For the complete documentation index, see [llms.txt](https://docs.senseloaf.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.senseloaf.ai/getting-started/responsible-ai-at-senseloaf.md). # Responsible AI at Senseloaf Safe, explainable, and controlled AI across every hiring workflow Senseloaf is built on the principle that AI in hiring should be transparent, fair, and predictable. Governance mechanisms are built directly into the platform so recruiters can use AI effectively without having to manage technical or regulatory complexity themselves. The goal is to ensure: * Predictable agent behaviour across all workflows * Fair and consistent candidate evaluation at every stage * Clear boundaries on what AI can and cannot do ### AI safety and anti-discrimination safeguards Senseloaf enforces system-level guardrails to ensure AI behaviour remains job-relevant, fair, and compliant across all hiring agents. Prompts that attempt to introduce unsafe, biased, or non-job-related criteria are automatically blocked or neutralised. These safeguards are designed to: * Prevent discriminatory or exclusionary evaluation logic * Keep AI outputs focused strictly on role-relevant qualifications * Ensure alignment with hiring best practices and regulatory expectations Senseloaf does not allow AI agents to act on prompts, criteria, or instructions related to legally protected characteristics. Any attempt to introduce such parameters is rejected by default. **Protected characteristics include, but are not limited to:** * Age (including 40+) * Sex or gender identity * Sexual orientation * Race or colour * Religion * National origin * Disability * Genetic information * Marital or parental status * Veteran status ### Agent-level compliance Each hiring agent in Senseloaf follows governance rules tailored to its role in the workflow. **AI Matching Agent (Resume Matching Agent)** The AI Matching Agent evaluates candidates strictly against job-relevant criteria defined in the resume strategy. * Scoring is based only on skills, experience, seniority, and role alignment * Strategy modifications made via natural language prompts are validated against governance rules before being applied * Prompts or criteria related to legally protected characteristics are blocked by default * Scoring logic remains transparent, explainable, and reviewable at all times * No candidate communication occurs at this stage **AI Interview Agent** The AI Interview Agent applies governance controls to ensure interviews remain fair, professional, and focused. * **Professional conduct enforcement** — the AI filters profanity, offensive language, and discriminatory remarks. If inappropriate language is detected, the conversation is gently redirected back to a professional, skills-focused tone * **Derailment and relevance checks** — the system detects off-topic, evasive, or irrelevant responses and nudges candidates back toward answering the interview questions, ensuring productive interviews and consistent signal quality ### Proctoring signals In addition to conversational guardrails, Senseloaf monitors behavioural signals during interviews to help recruiters assess authenticity and interview integrity.

Proctoring Signals

Proctoring signals may include: * **Tab switching** — identifies repeated context switching during the interview * **Fullscreen exit** — flags when candidates navigate away from the interview screen * **Multi-screen usage** — detects when additional displays are connected during the session #### A note on manual edits and recruiter responsibility When recruiters manually edit, add, or override prescreening questions or evaluation criteria, they are responsible for ensuring those changes comply with applicable hiring policies and regulations. Senseloaf always: * Continues to block prompts related to legally protected characteristics * Does not validate subjective intent or business justification behind manually written questions * Treats manual edits as recruiter-owned configuration decisions Candidate engagement only begins after a workflow is activated. These controls balance AI-driven safety with recruiter ownership and accountability. ### Data usage and privacy Senseloaf processes candidate data strictly for hiring-related evaluation within active workflows. * Data is used only within the scope of the configured workflow * Attachments and inputs are treated as contextual evaluation data * Access is limited to authorised users and systems ### Auditability and traceability Senseloaf maintains traceability across AI-driven decisions to support review and accountability: * **Strategy versioning** — every change to a resume matching strategy is versioned and preserved * **Preserved configurations** — prescreening and interview configurations are retained for reference * **Clear linkage** — each job role, agent, and evaluation outcome is traceable back to the logic that produced it These mechanisms help teams understand why decisions were made and how configurations evolved over time. ### Compliance alignment Senseloaf is designed to support alignment with common hiring and data protection regulations, including: * **Equal Employment Opportunity (EEO)** principles * **Data protection and privacy regulations** such as GDPR --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://docs.senseloaf.ai/getting-started/responsible-ai-at-senseloaf.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.