Solidatus adds AI governance tools to lineage platform

Solidatus has launched version 2026.3 of its data lineage platform, adding features designed to connect governed data lineage with AI agents.

The release introduces support for Model Context Protocol (MCP), a Bring Your Own LLM option, and updates to the Solidatus AI Assistant, including persistent sessions and new workflow tools.

The changes are aimed at organisations that want to use AI systems without losing control of data provenance, oversight, and audit trails. Solidatus, whose customers include HSBC and Deutsche Bank, focuses on data lineage software for regulated industries.

The update centres on what Solidatus describes as three areas: governance of AI, governance by AI, and governance for AI. In practice, this means giving external AI tools access to governed lineage information, letting customers choose their own large language model providers, and keeping records of AI-assisted changes within the platform.

MCP support allows external AI tools and agents to connect to Solidatus and work with lineage model content. Those tools can access lineage that has already been modelled, approved, and maintained by human users.

Bring Your Own LLM allows organisations to run the Solidatus AI Assistant with their own approved model provider and keys. For companies in heavily regulated sectors, this may be appealing because it gives them tighter control over which AI models are used internally.

Every AI-assisted change remains staged for human review before acceptance. Once accepted, it is captured as an attributed, timestamped revision in the model history.

Assistant changes

A large part of the release focuses on the day-to-day use of the Solidatus AI Assistant. Conversations now continue across browser sessions, allowing users to return to an earlier investigation without starting over.

Solidatus has also introduced Chat Rollup. When a long session reaches its context limit, the tool carries active states, filters, traces, and key entities into a new conversation so work can continue.

This is particularly useful for complex lineage analysis, where an investigation can span multiple systems, data transformations, and business definitions. Instead of rebuilding the trail from memory, users can continue from the last working context.

A new Prompt Library has been added for teams running repeatable governance tasks. It allows users to create, store, and reuse prompts, which could help standardise recurring regulatory or internal control questions.

Quick Actions and queued follow-up requests have also been added. These features allow users to surface contextual next steps and line up additional tasks while the assistant is still generating a response.

Navigation and reporting

Beyond AI features, the release includes changes to navigation within Data Domains. A new Domain Spine keeps sections such as Home, AI Assistant, Reference Hub, Lineage Models, saved Data Maps, and related domain content within one click.

Lineage models now sit more clearly within Data Domains, and searchable entity hierarchy tables have been added to help users explore nested structures without leaving the domain view. These changes are intended to make it easier to move through large, complex data environments.

Analytics and reporting workflows have also been updated. Users can now export the records behind any metric to CSV from the analytics report drill-down panel, giving them direct access to the underlying evidence behind governance measures.

Reports can also be edited and deleted directly from the Data Domain home page. Data Maps now save without reloading and are intended to retain display rules more consistently.

The broader release also includes platform modernisation work covering compatibility, maintainability, search, and the handling of large or complex sessions. Solidatus has also improved navigation and Data Map rendering across larger models.

Businesses across sectors have increased their use of AI agents and low-code or no-code tools, but governance concerns remain central, particularly where firms must explain how an output was produced and what source data informed it. In that context, data lineage tools have become part of the wider control framework around AI deployment.

Philip Dutton, Founder, Solidatus, commented on the rationale behind the update.

"AI governance is quickly becoming one of the defining challenges for enterprises, particularly in regulated sectors where organisations need to prove not just what a system said, but what data it relied on and who approved the outcome," said Philip Dutton, Founder, Solidatus.

"Solidatus 2026.3 gives AI assistants and agents access to the governed lineage they need, while keeping model choice, human oversight, and auditability firmly in the hands of the organisation," said Dutton.