AI Transparency & Use Policy

Clear about our tools.
Accountable to you.

AI should support the people responsible for your community. It should never replace their judgment or gain unrestricted access to your information.

Updated November 1, 2026

People keep the final say.

AI must not independently decide matters affecting a member’s money, access, employment or safety. A responsible person needs the context and authority to question, edit or reject a recommendation. An automated workflow also needs clear authorization and boundaries.

Access has a purpose and a limit.

Enabling assistance must not grant an AI tool blanket access to your database. Our standard is the minimum information and permissions needed for the specific task, within your organization’s authorized scope.

Your team controls participation.

Optional AI should be a choice. Our product requirement is that core membership, billing, check-in and reporting workflows remain usable without an AI-generated answer.

Transparency starts with us

Yes, we use AI-assisted tools.

Clubward is a human-led software company. We use AI-assisted tools in parts of our development, research, documentation and content work. People remain responsible for deciding what we build and what represents Clubward.

Our standard is appropriate review, testing and security consideration for the work involved. AI assistance does not remove that responsibility. It is not a substitute for engineering judgment or an assurance that generated work is correct.

Internal tool use is separate from access to a customer workspace. Customer records should not be supplied to an AI tool without a defined, authorized purpose and appropriate data boundaries. This policy sets our standards; the specific marketing workflow disclosed below explains an existing use and its limits.

Information with boundaries

Your records are not an open invitation.

A general announcement should not need your membership database. A summary of a report should not require unrelated household details. We start with the task and ask which information is actually necessary.

Our policy is that customer data must not be used to train AI models. Provider terms and configuration need to support that standard when an AI use is approved. Training, processing a request and retaining logs are separate questions; an answer about one does not answer the others.

Confidential member notes, children’s information, payment details and credentials deserve particular care. An AI use must not bypass the permissions that govern the underlying records or open access to another organization’s information.

Before you choose

Clear answers for the feature in front of you.

Before you authorize an AI use, we should explain its scope in plain language and demonstrate the relevant controls. That explanation needs to cover:

  • Purpose and availability: what the feature does, what is released and what is still planned.
  • Inputs and providers: which fields leave the platform, where they go and which service processes them.
  • Retention and access: what happens to requests, outputs and logs, who can access them and how long they are kept.
  • Authority and review: what it can read or change, who approves a consequential action and how to correct a mistake.
  • Choice and continuity: how to decline or stop the assistance and continue the underlying task.

Provider names, processing locations and retention periods belong in the disclosure for the actual feature and service agreement. We will not substitute a general AI statement for those details.

Oversight that means something

A confident answer still needs checking.

Review requires more than an approve button. The person responsible needs access to the source records or policy, an explanation of the proposed action and a way to stop it. Generated text needs an appropriate review process before it represents your club. For an automated workflow, agree the authorization, content boundaries and checks before enabling it.

Ordinary automation follows the rules your team authorizes. That is different from allowing an AI system to invent a decision or expand its own permissions. See our AI and automation guidelines.

A specific AI use

Marketing drafts, with defined inputs.

When configured, marketing draft assistance uses OpenAI to prepare a campaign name, subject, message and description. The request contains the opportunity type, title, summary and recommended action, together with a limited audience sample.

That structured sample includes up to six contacts’ first names, audience source information and statuses. It does not include their email addresses or phone numbers as structured fields. Opportunity text is also included, so unnecessary personal details should be kept out of that text.

This workflow sends a prepared request to the provider. It does not give the model a database login. Provider processing, retained inputs and outputs, access and deletion still need to be considered separately from model training.

Standard campaigns have an approval step. Approved automation can carry its authorization into later runs, including generated content. Your team needs to review the automation’s scope and content expectations before enabling it; an individual manual review is not guaranteed on every automated run.

Before a tool earns access

Purpose first. Provider second.

Before approving an AI use involving customer information, our standard is to establish the task, the minimum information it requires and why AI is appropriate. We need to evaluate the provider’s training and retention terms, processing locations, access controls and deletion options.

We also need to identify who authorizes the work, how consequential outputs are checked and how the underlying task continues without AI. Material changes to a tool, provider or use require those questions to be revisited.

These are evaluation requirements. A general policy does not establish a particular provider’s retention period, processing location or contractual terms. Ask us for the disclosure relevant to the feature or service you are considering.

A person you can reach

Help should still feel human.

Automation may help organize requests or reduce repetitive work. Our commitment is to preserve meaningful access to human assistance and a person who can take responsibility for helping you. Support availability and response commitments remain those agreed for your service.

Questions about an AI use, an unexpected result or a change to these practices deserve a clear response. Contact our team before sharing sensitive records through a website inquiry. Read our website privacy notice and platform data privacy information for their respective scopes.

Talk to a person

You deserve a clear answer.

Bring us a real task, a data concern or a question about who has the final say. We’ll walk through the relevant workflow and its boundaries with you.

Ask our team