AI Transparency

Metrics That Matter includes AI-powered features to help customers ask questions about their business, explore performance trends, generate summaries, and speed up analysis that would otherwise require manual work across multiple reports.

These features currently include conversational interfaces such as Scout, automated analysis workflows, company-scoped data tools, and other AI-assisted experiences built into our reporting and content workflows.

This page explains, at a practical level, what data those features may use, why that data is processed, and what controls are applied when AI features run.

Depending on the feature you use, our AI systems may process:

• Your prompts, questions, instructions, and follow-up replies
• Conversation history needed to answer within the same thread
• Page context from the part of the application you are viewing, such as visible filters, chart context, and summarized table data
• Account, company, and workspace context needed to scope the request
• Usage metadata such as timing, token counts, and tool execution status

We design these flows to provide the requested AI feature, preserve continuity in a conversation, and keep the model grounded in the company context relevant to the user's request.

When you connect third-party platforms such as Shopify, Meta, Google, TikTok, or other data sources supported by Metrics That Matter, AI features may use data from those connections to answer questions and generate insights on your behalf.

This can include, for example:

• Sales, orders, products, sessions, and margin-related metrics
• Advertising spend, clicks, impressions, and campaign results
• Performance data shown on reports, dashboards, and detail pages
• Structured query results returned by internal analysis tools that read company-scoped analytics tables

We use this data to provide the requested analysis or answer. We do not make connected-platform data broadly available outside the company and permission scope associated with the request.

To provide continuity, history, and reporting around AI usage, we may store AI conversations, user messages, assistant replies, and metadata associated with a request.

Stored records may include:

• Conversation identifiers and message content
• Model name and request identifiers
• Response times, token counts, and usage-cost metadata
• Tool execution events needed for troubleshooting and auditing
• Generated outputs created through AI-assisted workflows, where the feature requires those outputs to be saved

This storage helps us render past conversations, support debugging, monitor reliability, and improve the product experience within your workspace.

Some AI experiences use limited memory and personalization to make the assistant more useful over time. This can include remembering business context, preferences, goals, recurring instructions, or prior insights that are relevant to future requests.

These memories are stored as structured records and may be retrieved when a later request is processed so the assistant can respond with better continuity and less repeated setup from the user.

Memory is intended to improve product usefulness, not to expand access beyond the company and permission boundaries already applied to the request.

AI requests in Metrics That Matter are processed within a company context. Internal tools used by the AI are scoped to that tenant and are designed to respect user-specific data permissions where those permissions apply.

• Requests are associated with a company or workspace context before tools execute
• Internal data tools are run with company-scoped parameters rather than broad global access
• Permission checks and platform filtering are applied before tool execution where supported by the feature
• Access to stored conversations and account data remains subject to the access controls of the underlying product

No security model is perfect, but company isolation and permission enforcement are part of how we design AI-assisted access to customer data.

To deliver AI features, we may send relevant prompts, context, and structured instructions to third-party AI providers that process data on our behalf. Those providers generate the responses returned through our service.

Our AI features currently utilise models and services provided by OpenAI.

We configure our AI features so that customer data submitted through Metrics That Matter is not used to train third-party foundation models.

Even when third-party providers are involved, the data is sent for the purpose of operating the requested feature, not for those providers to build general-purpose profiles about your business.

AI-generated outputs can be useful, but they are not guaranteed to be complete, current, or correct in every case. Models can summarize incorrectly, miss context, or present an answer too confidently.

Customers should review AI-generated analysis before relying on it for business, financial, legal, or operational decisions. This is especially important where a response involves interpretation rather than a direct calculation from underlying data.

We recommend using AI outputs as an assistive layer alongside normal human review, not as a substitute for professional judgment.

AI-related records may be retained for as long as necessary to provide the relevant feature, maintain conversation history, support account operations, investigate issues, comply with legal obligations, and enforce our agreements.

If you request deletion, we can also delete AI-related conversation records and associated AI data held in connection with your use of these features, subject to any legal, security, or operational retention requirements that still apply.

General information about how we handle personal data is described in our Privacy Policy. The contractual rules governing your use of the platform are set out in our Terms of Service.

If you need assistance with deletion or AI-related data questions, please contact us using the details below.

If you have questions about this AI Transparency page or our AI data practices, please contact us at: