Responsible AI

Responsible AI at SalesOps

AI features in SalesOps assist human sales professionals. They draft, summarize, and recommend. They do not send emails, place calls, sign contracts, or take irreversible actions on their own.

Policy Commitment

Principles

  • Human-in-the-loop. Outputs are drafts and recommendations that a User reviews before acting.
  • Purpose-scoped. Only the information required for the specific feature is transmitted to the AI provider.
  • No customer-content model training. SalesOps does not use one customer's workspace data to train models for another customer or to train a shared SalesOps model, unless the customer has expressly agreed in writing.
  • No sale of AI inputs or outputs. SalesOps does not sell AI-processed customer data.
  • Transparency about limitations. AI outputs may be incorrect or incomplete. We label them as drafts, not decisions.

Where AI is used

  • Drafting personalized cold emails and call briefs

    Current Control
  • Summarizing meeting notes and follow-ups

    Current Control
  • Sales-plan reasoning and prospect prioritization

    Current Control
  • Business-card OCR extraction

    Current Control
  • Weekly-report narrative summaries

    Current Control
  • Autonomous outbound sending

    AI features never send email or place calls on behalf of a User.

    No

Data sent to AI providers

When a User invokes an AI-assisted feature, SalesOps sends the AI provider only the information required to complete the request. We use OpenAI as our primary AI provider today. See /legal/subprocessors for our current disclosed list.

Do not paste highly sensitive information (health data, government IDs, payment credentials, or information you are not authorized to share) into AI-assisted fields.

Model training and improvement

SalesOps does not train foundation models. We use commercially-available AI services under their standard enterprise data-handling terms. Provider-side retention and training-opt-out settings are configured in line with each provider's business-tier defaults.

Customer controls

  • Workspace administrators can request that AI features be disabled for their workspace.
  • Users can decline to use AI drafts and edit outputs before use.
  • AI usage generates telemetry retained only as long as needed to operate and improve the feature.

Known limitations

  • AI outputs are drafts and may contain factual errors, awkward phrasing, or outdated information.
  • AI features are not audited for compliance with any specific regulatory AI framework (e.g. EU AI Act obligations for high-risk systems). SalesOps AI features are intended as productivity tools, not automated decision-making systems.
  • Availability of AI features depends on the availability of upstream AI providers.
Information provided in the SalesOps Trust Center is for general informational purposes and does not constitute legal advice. Contractual commitments applicable to a particular customer are governed by that customer's executed agreement with SalesOps. SalesOps is a sole proprietorship based in Sudbury, Ontario, Canada.