AI Features Schedule

This schedule lists every feature of the Sitterz Service that uses an artificial-intelligence model, what each one does, what information it receives, and the limits that apply to all of them. It forms part of the Terms of Service and the Data Processing Addendum.

It exists so that a business using Sitterz can answer its own questions — and its own regulators — from a document, rather than from an assurance in a sales call.

There is no AI feature in Sitterz today. The website-design assistant that earlier versions of this schedule listed was withdrawn on September 1, 2026, before the pilot; no feature of the Service sends any information to an artificial-intelligence model, and no model provider is engaged.

If a feature that uses a model is added, it gets a row here — what it does, who can run it, exactly what it receives, what it never receives, whether it touches personal information, and whether its output could bear on a decision about a person — before it ships, and this schedule's version moves with it.

The limits in the next section are not conditional on a feature existing. They bind any AI feature Sitterz ever adds, and they are why the product does not read, score, or assess what a business writes.

No AI feature ranks, scores, rates, matches, or assesses the suitability, quality, safety, or fitness of any individual, and no AI output is intended to be, or may be used as, a substantial factor in any decision regarding the employment, engagement, assignment, scheduling, compensation, discipline, or termination of any person.

That is a product constraint, not an aspiration. Concretely, Sitterz will not build:

• caregiver matching, recommendation, ranking, or “best fit” of any kind;
• a trust score, safety score, risk score, or any composite signal about a person;
• screening, verification, or applicant evaluation;
• detection or flagging of concerning content in daily logs, incident notes, or messages. This one is worth naming: a vendor that advertises detection takes on a duty to detect, and every signal it misses becomes a failure. Sitterz stores what a business writes and does not read it.
• assessment of a child's development, health, behaviour, or wellbeing.

Sitterz does not use Customer Data to train, fine-tune, or improve any model — ours or anyone else's — and does not grant any provider the right to do so. Engaging a model provider under terms that prohibit training on what we send is a condition of using it at all.

Two honest qualifications, because an absolute we cannot support is worse than a narrower statement that is true. First, a provider generally retains a limited amount of what it receives for its own abuse-monitoring purposes; that is not training, and it is not something any customer of any model provider avoids. Second, the inputs listed above contain no record about any client, child, or staff member in the first place — which is the real protection here, and it holds regardless of what any provider's terms say.

Sitterz will keep this schedule current, will notify customers of a material change, will retain records relating to AI features for at least three years, and will provide the information a business reasonably needs to give its own notices.

The business decides whether to use an AI feature, reviews and edits every output before relying on it, and is responsible for its own obligations under any law governing AI notices, assessments, or automated decisions. A business must not use the Service in a way that would make Sitterz a developer or deployer of a high-risk or consequential-decision system, and must not use any output as a factor in a decision about a person.

Sitterz is not the business's agent and does not act on its behalf in any employment-related function.

AI output is machine-generated, may be inaccurate or incomplete, and must be reviewed before it is relied on. It is not advice of any kind, and it is not a determination about any person.