Connecting your data sources

Once data has been prepared, a secure connection must be established between the client's data repositories and K. The preferred access model should be agreed upon between the client and Owkin during the pre-sales phase.

Three connectivity options are available:

• Transfer to Owkin-managed storage Data is transferred into an Owkin-managed environment where it is stored and served directly to K. This is the simplest option and is well suited for clients who prefer a fully managed approach.

• Expose customer-managed storage Data remains in the client's own infrastructure. The client provides the necessary credentials and configuration so that K can access the data remotely. This option preserves the client's existing data residency and governance controls.

• Connect an existing data platform K integrates directly with platforms such as Databricks or Snowflake, leveraging the client's existing governance policies and data pipelines. This is the preferred option for organizations that have already invested in a centralized data platform.

Manual data or document upload will be supported in soon for small volumes, subject to customer configuration and policy.

For proprietary assets like screening results, assay data, target assessments, or compound libraries, the same pattern applies: bring the data in your preferred tabular or platform-backed structure, then align it to the supported model.

Owkin can also provide tooling and services for transformation, and optional AI enrichment can extract additional features before analysis.

---

K Pro's connection can be tailored to the use case. For instance, when connecting to a client's data platform, we can establish a live connection to a specific view of the data, or alternatively connect to a safeguarded version that requires manual updates — the choice depends entirely on the use cases being served. The same flexibility applies to other integration methods, such as connecting to storage (whether Owkin-managed or client-managed).