Howwe AI Policy

Version 2026-10-22 · Last updated 22 October 2026

Download PDF

This policy describes how Howwe Technologies AB (publ) uses AI in the Howwe service, and how customer data is protected when it does. It covers two things: Howwe's own AI features, and AI assistants that customers connect to Howwe themselves.

1. Howwe's own AI features

Howwe's intelligence is built into the software. Examples:

• Strategy Planner drafts goals, initiatives and milestones from a company's strategy material.
• Strategic Pulse reads execution data and briefs leaders every month on risks and opportunities.
• Intelligence Hub researches public information about a company and its market to support planning.

The features run on Google's Gemini models through Google Cloud, which is listed as a sub-processor.

2. Our commitments

1. No training on customer data. Customer data is not used to train or fine-tune AI models, whether Google's, other external or public models, or Howwe's own.
2. Only what the feature needs. A feature sends the AI model only the data it needs for the task the user started.
3. Same access rules. AI features follow the user's permissions in Howwe. They cannot read data the user could not read.
4. People decide. The features propose, summarise and flag. A person reviews, changes and decides what to use. Strategy Planner works as a draft in a guided workshop, which users edit before they use it.
5. No decisions about individuals. The features are not used to evaluate, rank or make decisions about employees.
6. Customer choice. A customer can ask us to turn off Howwe's AI features for its organisation, and we do so within two business days.

3. AI assistants you connect yourself

Users can connect AI assistants, such as Microsoft Copilot, Claude, ChatGPT and Gemini, to Howwe through our MCP server.

• The assistant acts with the user's own permissions.
• The assistant's provider processes the Howwe data the user asks about under its own agreement with the user or the organisation. It is not a Howwe sub-processor.
• Howwe records which tool was used, when, by whom and whether it succeeded. Howwe does not record the conversation, the content, the data returned or search terms.
• The organisation decides whether its users may connect AI assistants. An administrator can ask us to turn off access through AI assistants for the organisation.

4. EU AI Act

We assess Howwe's AI features as minimal risk under the EU AI Act (Regulation (EU) 2024/1689). They are not used in any of the high-risk areas in Annex III, such as recruitment or evaluation of employees. We tell users when content is produced by AI, and we review this assessment when we add or change an AI feature.

5. Governance

• New AI features and new AI providers are reviewed for security, data protection and AI Act impact before release.
• New AI providers that process customer data are added to the sub-processor list with at least 30 days' notice.
• Employees who build or operate AI features are trained in responsible AI use. Customer data is only entered into AI tools that Howwe has approved, under business terms that do not allow the provider to train on it.
• Howwe's CTO is responsible for this policy, which is reviewed at least once a year.

6. Contact

Questions about AI in Howwe: contact@howwe.io.