AI governance

AI when it helps. Humans where it matters.

Intelligent guidance does real work inside Chaperone, but it works within firm limits, under human oversight, and never in place of a trained Peer Coach. Here is exactly what our AI does, what it will never do, and how we keep it accountable.

Technology assists the work. A person stays responsible for the care.

Intelligent guidance helps with routing, summaries, check-ins, workflow assistance, and safety signals so Peer Coaches can spend more time being present with people. It does not make care decisions on its own, and a qualified human is always in the loop where it matters.

What intelligent guidance does

AI handles the mechanics so people can focus on people

Each capability is scoped to assist a Peer Coach or care team, not to replace their judgment.

Routing

Helps connect a person to a suitable Peer Coach or support path based on what they share, so the right human responds sooner. The connection is reviewed by people.

Summaries

Drafts concise summaries of conversations and history so Peer Coaches have context fast. Summaries support a coach’s understanding; they never stand in for a coach reading the full picture.

Check-ins

Sends gentle, scheduled check-in prompts to help people stay connected between conversations. Responses are visible to the care team, and a person decides what happens next.

Workflow assistance

Reduces administrative load by surfacing tasks, suggesting next steps in a workflow, and organizing information, so coaches and supervisors spend less time on overhead.

Safety signals

Helps flag language that may indicate someone needs more than peer support, so a trained person can review and act quickly. The signal prompts human attention; it never decides on its own.

Coach support tools

Offers optional drafting help and resource suggestions a Peer Coach can accept, edit, or discard. The coach always owns the message and the relationship.

  • Replace a trained Peer Coach or own the relationship
  • Diagnose a condition or label a person
  • Provide treatment, therapy, or clinical care
  • Act as crisis counseling or a crisis service
  • Make a final care decision without human review
  • Become the primary relationship a person relies on
Where we draw the line

What our AI will never do

Some things must stay human. Intelligent guidance is intentionally limited so it can never drift into roles that belong to trained people, licensed clinicians, or crisis services. These limits are part of how the product is designed, not a setting that can be quietly turned off.

If a moment looks like it needs more than peer support, the workflow routes it to approved support and escalation processes, with a person responsible for the next step.

How it’s governed

Guardrails, oversight, and accountability by design

Governance is not an afterthought. It is built into how intelligent guidance is configured, supervised, and reviewed.

Human in the loop

A qualified person reviews and is accountable for safety-sensitive output. AI prompts attention and prepares context; people make the calls that affect care.

Admin & governance controls

Administrators configure how AI features are used in their deployment, set permissions, and control where guidance is enabled, with sensible, conservative defaults.

Auditability & logging

AI-assisted actions are recorded so teams can see what guidance was offered, who reviewed it, and what a person decided. This supports transparency and accountability.

Defined guardrails

Guidance operates within scoped instructions and limits aligned to its task. Safety-sensitive workflows route to approved support and escalation processes rather than handling moments alone.

Ongoing review & quality

AI behavior is reviewed against quality and safety expectations, and adjusted over time. People remain responsible for evaluating whether the assistance is helping the work.

Transparency to people

We aim to be clear when guidance is AI-assisted versus a message from a Peer Coach, so people understand who and what they are interacting with.

Principles at a glance

The commitments we hold AI to

Human
A person stays accountable for safety-sensitive decisions
Assist
AI supports the work; it never owns the relationship
Limited
No diagnosis, treatment, or crisis counseling, ever
Auditable
AI-assisted actions are logged and reviewable
Keep exploring

Related pages

Let’s talk

Reviewing how we use AI? Bring your questions.

Whether you are evaluating our governance for a procurement review or simply want to understand the limits, we’ll walk through exactly how intelligent guidance is scoped, supervised, and kept accountable.