Quinn FAQ

Quinn is the findIQ AI assistant, operated through a chat: you describe what should happen, and Quinn carries out the actions in findIQ — on the same objects you would work on in the interface. Quinn is an additional way to operate findIQ, not a replacement for the interface. In full: What is Quinn?

Quinn reads documents and turns them into findIQ content, records expert knowledge in the chat, structures symptoms and causes that are already there, and answers questions about your data. Quinn works on the same objects as the interface — templates, symptoms, causes, routines, labels — and cannot delete templates or machines. Each task with its input and result: Working with Quinn

MCP stands for Model Context Protocol, an open standard that gives AI systems structured access to external software like findIQ. A human uses findIQ through the interface, an AI system uses findIQ through the findIQ MCP server. The MCP server itself is not the AI, only the interface in between. Details: What Is an MCP Server?

Quinn works on verified findIQ knowledge: expert input and existing documents are made available in a structured way through the findIQ AI model. Every answer therefore has an origin and an approval – unlike a general-purpose LLM that searches a file store and formulates an answer without a source.

Once enabled for your company, Quinn is directly linked from within the findIQ app.

The user signs in inside the chat with their usual findIQ credentials. From then on the AI system can never do more than that person — and for safety it can do less: deleting machines and templates is blocked, even when the person signed in would be allowed to.

Both. In principle, anything the API can do, the MCP server can do too: reading data (e.g. "Which machines does customer X have?") as well as writing or changing data (e.g. creating a new machine or updating a service case). Some actions are deliberately excluded – templates and machines cannot be deleted.

Yes. Quinn is activated per company. Your findIQ contact can enable or disable it for your organisation at any time — if it's disabled, access is removed for all users of your company.

Common systems such as Microsoft Copilot, Claude, and ChatGPT. If you connect your own system via the MCP server, you choose provider, model, and cost model yourself. With Quinn, findIQ selects the appropriate LLM provider for you.

The production endpoint is https://findiq-mcp-prod.azurewebsites.net/mcp. Guides: Setup in Microsoft Copilot Studio and Setup for Claude. On first use you will be prompted in the chat to sign in with your findIQ credentials.

Yes. An LLM handles the language part — understanding your request and phrasing the answer. What it works on is the difference: instead of searching a file store, Quinn reads and writes the structured findIQ data model, so every answer has an origin and an approval. Which model is used is answered in the next question.

Quinn is powered by Luna, a model from OpenAI, hosted in Microsoft Azure (Europe).

Azure's Terms of Service state:

• Data is not used to train, retrain, or improve the base models.

• Data is not available to OpenAI (the models run in Microsoft's Azure environment and never touch OpenAI's own services).

The default model is subject to change; EU hosting and not training on your data are not.

You don't need to configure anything — findIQ selects and maintains the model to ensure the best quality and compliance.

The default model is managed centrally by findIQ so we can guarantee consistent quality, EU hosting, and data-protection standards for every request. If you have specific requirements, please contact findIQ. If you want to choose provider and model yourself, connect your own AI system via the MCP server.

Processing runs in Microsoft data centres; your data is stored exclusively on findIQ servers in Europe. There is no permanent storage outside the EU.

More detail on hosting, backups, and logging: Tech-FAQ

Microsoft. The AI functionality is based on OpenAI models provided through Azure AI Services.

• No. Your data does not go into the public training of large language models.

• No documents, logs, or interviews are fed into external models or shared with third parties.

• In the Microsoft data centres the documents are only processed. Storage happens exclusively on findIQ servers in Europe.

Your documents are analysed for one purpose only: to extract the relevant knowledge and make it usable in findIQ. There is no further use or processing.

Quinn is free of charge during the beta. Pricing information will follow towards the end of the beta. If you connect your own AI system, the costs of your chosen provider apply in addition (licence or pay-as-you-go).

Reach out to your findIQ Customer Success contact or use the feedback option in the app — feedback directly shapes what we build next.