Intapp Celeste
AI your firm can trust
Celeste integrates with your existing stack and works inside your compliance controls.
Intapp Celeste
Celeste integrates with your existing stack and works inside your compliance controls.
Celeste uses configurable playbooks to get work done while respecting your security and compliance requirements.
Playbooks encode your firm’s methods into customizable, repeatable workflows that Celeste executes
Skills — reusable, modular actions such as searches and calculations — power every playbook
Connectors integrate Celeste with your existing systems so every skill and playbook acts on real data
Administrators can choose to power Celeste with Anthropic Claude or OpenAI GPT — but aren’t locked into either.
Celeste integrates with Intapp DealCloud, Time, Compliance, Collaboration, Intapp Data, Microsoft 365, Snowflake, Web, and more.
ISO 27001, ISO 27017, ISO 27018, ISO 27701, CSA STAR, and Global PRP
Data and privacy
No. Client data is never used to train models, and it’s never shared across firms. Model providers operate under contractual use limits and deletion commitments. Logical tenant isolation controls are implemented throughout Intapp products.
Celeste is hosted on Microsoft Azure, with LLMs hosted through AWS Bedrock for Anthropic models and Azure for Azure OpenAI models. In accordance with our terms, U.S. client data is stored at rest on U.S. infrastructure and EU client data is stored at rest on EU infrastructure.
Client data does not cross between regions except as required to provide service in accordance with your agreement and data-processing addendum (DPA). A complete list of Celeste sub-processors can be found here.
Client data is stored for the duration of your client agreement and remains accessible during the applicable retention period following the expiration or termination of the agreement before being permanently deleted by Intapp. Prompts and outputs are never used for model training.
Security and governance
Celeste certifications include ISO 27001, ISO 27017, ISO 27018, ISO 27701, CSA STAR, and Global PRP. Learn more.
Yes. An independent security firm performs third-party penetration testing at least annually.
All data is encrypted using AES-256 at rest and TLS 1.2 or higher in transit, with keys stored in FIPS 140-2-compliant key stores.
Access is governed through Intapp Identity Management (IDM). Celeste supports default and custom roles built from granular capabilities, with SSO support through Okta and Microsoft Entra ID. Role-based controls determine which features each user can access.
Through Intapp Walls for AI and permission inheritance. MNPI, insider controls, and information barriers are enforced by design and checked each time a playbook runs, including playbooks published through MCP, before data reaches third-party interfaces.
No. Security reviews for other Intapp products do not automatically cover Celeste. Celeste operates as its own deployment with dedicated compute, data stores, and access controls, independent of products such as Intapp DealCloud.
Azure AI Content Safety screens inputs and outputs for offensive content. Playbooks constrain Celeste’s behavior to approved tools, and outputs are filtered before they’re returned.
AI, control, and customization
For reads (e.g., search, retrieval, data lookup), Celeste acts without prior approval. For writes that create, modify, or delete data in connected systems, explicit user approval is required before execution. For playbooks that run on a schedule, actions are taken on behalf of a named user with a delegated subset of their permissions.
Playbooks define goals, guardrails, and approved tools. Deterministic tools perform fixed operations grounded in your firm’s data, and Celeste stops and asks for clarification when a match is ambiguous. As with all AI-generated outputs, users should evaluate results for their specific use cases. Please see Intapp’s AI Disclaimer.
Unlike Copilot and point tools, Celeste layers your firm’s context, governance, and accountability on top of underlying AI models. Ethical walls, permission inheritance, and human-in-the-loop approvals are built into the agent runtime by design rather than added afterwards (ethical-wall enforcement requires Intapp Walls for AI).
Celeste is model agnostic. Administrators select supported models from AWS Bedrock or Azure OpenAI through the Celeste admin portal. Both providers operate under contractual use limits.
Yes. Through MCP, Celeste can surface as a governed skill within supported MCP-compatible hosts. Your firm’s ethical walls and permission controls travel with it, so Celeste can work alongside your existing tools rather than replacing them.
No. On-premises deployment is not currently available. Celeste runs on Microsoft Azure and AWS cloud infrastructure. U.S. client data is stored at rest on U.S. infrastructure, and EU client data is stored at rest on EU infrastructure.
Yes. Firms can customize existing playbooks or build new ones without coding. Platform administrators control what is published and how it runs, and new playbooks can be tested before being released to users.
Complete the form to have an Intapp expert show you how Celeste can help your firm move faster and win more business.