HeirWealth launches MCP server for AI wealth queries

HeirWealth has launched an MCP server for its wealth management suite, connecting its client data layer to AI assistants such as Claude and ChatGPT.

The product is aimed at financial advisers, accountants and family offices that use HeirWealth Atlas to view consolidated wealth data across multiple entities and structures. It lets users query that data in plain language through an AI assistant rather than relying on dashboards, exports or manual reconciliation.

The launch marks what HeirWealth believes is a first for an Australian wealth reporting platform. It reflects growing demand from advice firms to make internal data available in widely used AI tools while retaining oversight of access and data handling.

How it works

The product uses the Model Context Protocol, an open standard introduced by Anthropic and now supported by a range of AI assistants. Under this approach, an assistant connects to external data sources and tools through a standard interface rather than bespoke integrations.

In HeirWealth's case, the MCP server exposes the same consolidated data layer used by its Atlas platform. Advisers can ask questions about a client's family's holdings, reporting or exposures and receive responses based on data drawn from the firm's existing records.

The initial focus is on reporting and the adviser workflow. According to HeirWealth, an assistant can read consolidated reporting and generate portfolio insights, plain-English summaries, and risk or anomaly flags within seconds.

That could include questions about a family's total property exposure across all entities, a task that often requires staff to gather and reconcile information from several platforms. By moving that query into an AI assistant, HeirWealth aims to reduce the administrative work tied to meeting preparation, portfolio summaries and draft communications.

Control and access

Figures returned through the system come from HeirWealth's own calculation engine rather than from the AI model. This means numbers are consistent, repeatable and auditable, according to the company.

Access is governed through each firm's existing permissions and consent settings within HeirWealth. An AI assistant can only read information it is authorised to access, while client data remains governed within the HeirWealth environment.

HeirWealth also said no client data is used to train third-party AI models. The MCP server is available to Atlas firms on an opt-in basis and is designed to work with any AI assistant that supports the protocol.

For many advice firms, the broader issue is the gap between structured client data and the tools staff increasingly use in daily work. Advisers serving high-net-worth and multi-generational families often spend large parts of the week preparing reports, reconciling positions and assembling information across entities, trusts and investment platforms.

As AI assistants become more common in office workflows, firms have been looking for ways to connect them to their own records without losing control over permissions, auditability or source data. The Model Context Protocol has emerged as one way to do that by creating standard methods for assistants to call external tools and datasets.

HeirWealth positions itself above existing platforms rather than as a replacement. Its software aggregates data from across a client's structures into a single consolidated layer used for reporting and client visibility.

That focus on consolidation is central to the company's pitch for the MCP server. If advisers can query a single governed dataset through an assistant, the value lies less in the language model itself and more in the quality and completeness of the financial data being supplied.

Chief Executive Officer Ray Tubman outlined the company's case for the launch.

"Advisers have told us for years that the hardest part of their week isn't the advice; it's assembling the picture they need to give it. Our data layer already solved the consolidation problem. The MCP server takes the next step: it lets an adviser simply ask a question and get an answer drawn straight from their own client data, inside the AI tools they're already using. It removes hours of manual work without asking anyone to change how they work, and it does so within the permissions the firm already controls," said Ray Tubman, Chief Executive Officer, HeirWealth.

The launch comes as wealth managers, accounting firms and family offices weigh where AI can be used in regulated client work. The challenge has often been less about text generation and more about linking assistants to accurate, permissioned data sources that can support reporting and internal analysis.

For vendors in the sector, that creates an opening to add AI access layers to existing reporting systems. For firms using those systems, the test is likely to be whether the tools reduce preparation time without introducing new compliance or data governance concerns.

The new server is now available to Atlas users on an opt-in basis and connects to MCP-compatible assistants.