Customer OS · Skylarq

AI Transparency

Last updated: June 2026

At Skylarq, we believe in being transparent about the AI that powers Customer OS. This page explains which AI providers we use, how we use them, and how we protect your data.

1. AI Providers We Use

Customer OS uses large language models to turn the sources you connect into a useful, living memory of every customer. Each provider is used in accordance with its usage policies and on terms that prohibit training on your data.

Anthropic (Claude)

Our primary model provider, used to build and update your customer briefs and to answer questions over your connected data.

Your MCP client

Customer OS plugs into the AI client you already use — Claude, ChatGPT, Cursor, and other MCP-compatible tools — via the Model Context Protocol.

2. How We Use AI

AI powers Customer OS across a few core areas:

  • Building your customer memory — reading the sources you connect to assemble a structured, continuously updated brief on every account
  • Answering your questions — responding to plain-language questions about your customers, with citations back to your own data
  • Surfacing what matters — identifying promises made, open commitments, risks, and accounts going quiet
  • Summarization — condensing long email threads, meetings, and conversations into the parts that matter

3. Anthropic (Claude)

Anthropic's Claude models power the core capabilities of Customer OS:

  • Building and updating the living brief on each customer and account
  • Answering plain-language questions over your connected data
  • Summarizing threads, meetings, and messages
  • Surfacing risks, commitments, and recommended next steps

Compliance: Our use of Anthropic's API follows their usage policies. Data sent to the API is not used to train Anthropic's models.

4. Your AI Clients & MCP

Customer OS is built on the Model Context Protocol (MCP). Rather than locking you into one assistant, it serves your customer memory to the AI client you choose:

  • Claude, ChatGPT, Cursor, and other MCP-compatible clients can securely query your customer memory
  • Those clients are provided and controlled by you, and their use is governed by their own terms and privacy policies
  • Customer OS only shares the context needed to answer the request you make from your client

5. Data Privacy & AI

We take a privacy-first approach to AI. Here's how we protect your data when processing it through AI providers:

  • Minimal data sharing — We send only the data required to fulfill each request.
  • No model training — Your data is never used to train third-party AI models. We use API tiers that contractually prohibit training on customer data.
  • Encrypted transmission — All data sent to AI providers is encrypted in transit using TLS 1.2 or higher.
  • Workspace isolation — AI processing is scoped to your workspace; your data is never mixed with another customer's.
  • No provider retention — We do not allow providers to retain your data beyond the time needed to fulfill a request.

For more details on our overall data practices, see our Privacy Policy.

6. Responsible AI Commitment

Skylarq is committed to the responsible use of artificial intelligence. Our guiding principles:

  • Transparency — We clearly disclose which AI providers we use and how they are integrated
  • Accuracy — We cite sources for AI answers and continuously improve our prompts and validation
  • Human oversight — AI-generated briefs and recommendations are presented for your review; you stay in control
  • Bias monitoring — We monitor for and work to mitigate bias in AI-generated outputs
  • Clear communication — When content is AI-generated, we make that clear in the product

7. Updates & Contact

We will update this page as we add or change AI providers. Material changes will be communicated through the product or via email.

If you have questions about our use of AI, please contact us: