A chatbot responds to questions. An agentic AI takes action. The Celeste Panel in DealCloud doesn’t just surface information — it creates deal records, ingests documents, updates the CRM, and extracts structured data from unstructured sources, all within DealCloud. The deeper distinction is that Celeste operates with your firm’s own data, your permissions structure, and your defined workflows, which means it behaves like a trained professional working inside your system rather than a general-purpose assistant that has no context for how your firm works. For firms looking to go further — incorporating playbooks, a semantic layer, and agentic workflows that span your entire technology stack — that is the vision for Celeste for the Firm, currently in limited availability.

Yes. Using the Celeste Panel in DealCloud, you can generate structured deal documents — including one-pagers — in Word, PowerPoint, PDF, or Excel, drawing on the deal information already held in DealCloud and from the web. For firms that want to go further and score deals systematically against defined investment criteria using dedicated playbooks, that capability sits within Celeste for the Firm, which is currently in limited availability. Speak to your Intapp account team to understand what’s available for your firm today.

Yes. Celeste can ingest meeting notes and other unstructured documents, extract key details and action items, and create or update the corresponding records in DealCloud, so, the output of a meeting flows directly into your firm’s workflow rather than sitting in a document.

Not yet, but it is coming. The current workflow focuses on data extraction and record creation — Celeste reads the document, surfaces the relevant fields for your review, and creates the deal record once approved. The ability to automatically attach the source document, such as a CIM or teaser, directly to the corresponding DealCloud record is on the near-term roadmap. If this is a priority for your firm, we’d encourage you to raise it with your Intapp contact.

Celeste is designed to detect potential duplicates. If two teams create records for the same company under slightly different names, or upload the same CIM at different times, Celeste will flag a potential match before any record is created and prompt the user to review the existing entry. Visibility across teams respects the access controls and information barriers already configured in DealCloud — each team sees only what they are permitted to see, and Celeste surfaces a duplicate flag without exposing the other team’s data.

Access to AI models is table stakes. The firms that win will be the ones whose AI knows their business — the workflows they’ve refined, the playbooks they’ve built, the institutional knowledge accumulated across DealCloud and the Intapp suite. Those playbooks live in Celeste and outlast any individual LLM. When a better model comes along, you don’t lose your firm’s IP — you just run it on a better engine.

Within DealCloud, the Celeste Panel works directly with the data your firm already holds in DealCloud — deal records, contacts, companies, and relationship history — and surfaces it intelligently through natural language queries and automated workflows. For firms with data spread across additional systems, the broader vision for Celeste is to connect across your entire technology stack, including third-party data sources, document repositories, and other Intapp products. That capability sits within Celeste for the Firm, which is currently in limited availability. Your Intapp account team can walk you through what’s possible today and what’s on the roadmap.

This is managed through what is called the schema contract — a configuration layer where you specify which data fields Celeste should read from and write to. For firms with more complex data models, such as multiple EBITDA fields or financial information split across separate objects, this configuration step is where you point Celeste to the right locations. For most firms the basic schema contract is straightforward to set up, and the DealCloud team supports this during onboarding.

The Celeste Panel can access DealCloud data, Intapp Graph, and in real time, the web. These connections are visible and individually configurable within the interface — users can see which data sources are active and toggle web access off if preferred. Firms that require strict data boundary controls can configure Celeste to operate exclusively within their DealCloud environment.

For a significant portion of routine updates, yes. Rather than navigating to a record and updating fields manually, users can instruct Celeste in natural language to update deal status, log a note, or add a contact, and Celeste executes the change directly in DealCloud. Looking ahead, within Celeste for the Firm, Intapp is also building agents that automate data updates without any user prompt at all — so when a contact changes roles, a portfolio company publishes new filings, or firmographic data shifts, records stay current automatically rather than waiting for someone to notice and update them.

Through the semantic layer. Within Celeste, your firm builds and maintains ontologies and glossaries — a translation layer that teaches Celeste how your firm thinks and speaks. When your data model has custom objects or firm-specific terminology, Celeste needs that context to give accurate answers. It’s what separates an AI that understands your firm from one that’s just sitting on top of your data.

The Celeste Panel in DealCloud is designed to be low-lift from an IT perspective — enabling it is a feature-level configuration in the product, not a complex integration project. The main consideration is the schema contract, which defines which fields and objects Celeste should interact with. For firms with a straightforward DealCloud setup this is minimal, while firms with more customised environments will work through a structured configuration exercise with Intapp’s support rather than a technical build. End-user training requirements are also lighter than you might expect, since the interface is designed to be intuitive enough for professionals who have never worked directly with AI tools before.

Yes. Celeste Panel in DealCloud is built to be self-configurable, and Intapp is providing documentation and enablement collateral to support clients through setup. The intention is that teams without prior AI implementation experience will find the process approachable, and Intapp is deliberately shortening the time between feature enablement and live usage in the hands of end users.

Yes, Copilot integration sits within the Celeste for the Firm capability set, which is currently in limited availability. A number of clients are already working with it in this phase, and the broader commercial launch is being finalised. Timeline and pricing are best discussed with your Intapp account team, as both depend on your existing licensing and the specific integration scope required.

DealCloud’s Celeste Panel runs on Claude by Anthropic as its default model. On accuracy, the important distinction is that Celeste’s extraction capability is not a general-purpose model applied to documents — it has been specifically configured for the private capital document types firms work with every day, such as teasers and CIMs. The approach is deliberately weighted toward precision, meaning Celeste will only extract data it is confident is correct rather than returning a broader result that includes uncertain or inferred information. Hallucinations in document extraction have rarely been observed, and Intapp actively monitors quality as the product scales.

Intapp has been the custodian of sensitive professional services data for decades — deal history, client relationships, compliance records. That doesn’t change with Celeste. Your data stays within your governed infrastructure, is never used to train models, and is never shared externally. This is a meaningful distinction from interacting directly with a general-purpose tool like ChatGPT, where confidential information can inadvertently flow through a consumer interface that wasn’t built with your obligations in mind.

The firms Intapp works with are increasingly framing this as a strategic investment rather than a cost line, and evaluating it the same way they evaluate any other investment — on the return it generates. For private capital specifically, that return shows up as the ability to raise and deploy capital more efficiently, exit positions at better valuations, scale the business without proportionally growing headcount, and reduce both operational and portfolio risk. The question senior leaders are asking is not whether AI is worth the investment, but whether the AI they are committing to can deliver those outcomes within their specific workflows and governance requirements.

Not at all. Celeste is LLM agnostic — it’s not tied to any single model, and that’s deliberate. The AI landscape is moving fast. A year ago, few people were talking about Claude. We don’t know what the leading model will be two years from now, and your firm shouldn’t have to rebuild every time it changes. Your workflows, your data, and your firm’s intelligence stay constant regardless of which model is running underneath.

No — Celeste enforces your existing permissions, it doesn’t bypass them. The controls you’ve defined across the Intapp suite extend into every AI interaction. A banker on the wrong side of a wall won’t get answers that breach it through Celeste, any more than they would through a direct DealCloud search. For firms where MNPI controls are critical, this isn’t an afterthought — it’s built into the architecture.

DealCloud with Celeste currently operates under a fair use policy, meaning there is no consumption-based billing within normal usage parameters. Usage that falls significantly outside fair use thresholds may be subject to rate limiting. For firms that previously purchased Intapp Assist, the upgrade to DealCloud with Celeste is included at no additional cost.