AI drafts. Humans review and decide.
WhiskerMatch uses AI to reduce administrative friction — not to make placement decisions, not to invent medical claims, and not to replace the judgment of shelter staff. AI drafts are review aids. Humans retain full decision-making authority.
Routine work, not judgment work.
AI is presented as doing the administrative work that slows shelter staff down — drafting, structuring, summarizing, and organizing. It does not make placement decisions.
Drafts structured fields
AI drafts intake fields, profile summaries, and care notes from source material. Every draft is reviewable and editable before it ships.
Summarizes for review
AI groups applications, flags missing documents, and surfaces overdue items so staff can triage faster — not so AI can decide who gets approved.
Organizes and routes
AI suggests queue assignments, due dates, and next actions based on workflow rules. Staff confirm or override every suggestion.
Prepares, never decides
Placement calls, approvals, and publishes stay with named staff who know the animal. AI prepares the paperwork. Humans own the decision. AI drafts are review aids, not replacements for judgment.
Hard lines that do not move.
These are not policy preferences. They are product constraints.
Never auto-reject applications
No application is rejected based on model output. Staff triage, respond, and decide. AI summarizes for speed, not for judgment.
Never invent medical claims
AI does not diagnose, prescribe, or invent medical status. Care content is drafted from reviewed sources and signed off by credentialed humans.
Never produce opaque match scores
No 94% compatibility number pulled from a model. Fit signals come from what the shelter actually observed — surfaced with source and limits.
Never accept pay-to-rank
No partner, brand, or shelter can pay to change discovery order. Not by AI. Not by humans. Not ever.
Never publish without human review
No record, profile, or response reaches the public or an applicant without a named staff member reviewing it first. AI drafts do not ship on their own.
AI access to sensitive context is opt-in and bounded.
The public site does not send inquiry or shelter data to an AI service. Before product-stage processing, organizations will see what may be sent, for which task, under which provider settings, and how to keep a human-only workflow.
Minimum approved context
Only the source material needed for a specific drafting, structuring, summarizing, or missing-field task may be sent to an AI service.
Sensitive data needs explicit authority
Applicant PII, medical notes, behavior notes, and uploaded documents are not sent without an explicit product agreement, documented purpose, and appropriate consent or authority.
No model training by default
Organization data will not be used to train third-party models unless the organization explicitly opts in. Provider retention and training settings must be verified first.
Human-only path
A shelter can opt out of AI processing. AI-prepared output remains marked, editable, and blocked from public or placement action until a person reviews it.
Stated in public so it can be checked.
These boundaries are published so adopters, shelters, and reviewers can hold WhiskerMatch to them. AI drafts are review aids; humans decide. If you see something that contradicts these commitments, we want to know.
See the system that keeps humans in control.
Talk through one workflow and inspect exactly where preparation stops and human review begins.