Data lineage matters, especially in the garbage-in, garbage-out world of artificial intelligence.
We can define data lineage as a “visual map” of an organization’s data assets that traces their origin, integration points and transformations (i.e., actions including filtering or aggregation) and all other data event actions through each asset’s lifecycle. Essential for tasks such as compliance and governance, data lineage allows us to perform impact analysis on data in order to assess what effect any changes to it will have on upstream applications, systems and services as it progresses.
Solidatus works in this space and the company is now driving (arguably) timely new AI functions into its platform.
The company’s Solidatus AI Lineage Assistant is a new agentic AI tool designed to give data science and software application development teams audit-ready lineage to meet the most stringent regulatory demands at speed, but there’s also full human oversight.
Build, Enrich & Maintain
This new assistant service is intended to act as an agentic AI tool that builds, enriches and maintains data lineage across complex enterprise data estates.
Looking at where users (allegedly) fall short as a result of manual processes, Solidatus says that organizations across regulated industries face a growing mismatch between what regulators demand and what teams can achieve on the ground.
“Every organisation we work with is racing to fuel their AI ambitions with complete data lineage as context, but building it manually takes months,” said Philip Dutton, founder and CEO of Solidatus. “Our new AI Lineage Assistant has access to the full capabilities of the most advanced metadata modelling tool in the market, enabling teams to move at the speed regulators demand without sacrificing the human oversight and accountability they require to deliver their AI ambitions at scale.”
In financial services, BCBS 239 (a global standard for banks to improve risk data and reporting) and DORA (EU financial entities disruption regulation) require complete lineage across every reporting chain and system. Solidatus suggests the pressure to demonstrate data traceability is intensifying across other sectors as well.
“Manual lineage mapping takes months per project and documentation becomes outdated almost as soon as it is written. The result is compliance gaps that can trigger remediation orders, stall transformation programmes, and increasingly, block the AI ambitions that depend on trusted, traceable data,” said Dutton and team.
A Stitch in Lineage Saves Time
The AI Lineage Assistant works across technical and business domains simultaneously to “stitch lineage” across systems and map physical data to business terms. It is capable of enriching metadata (so that we have more information about information) and it flags regulatory gaps. These actions can be performed within a controlled safety sandbox to provide full contextual awareness of an organisation’s data estate.
Explainable reasoning and a full audit trail let organisations demonstrate exactly what was proposed, what was approved, and when. Users interact through natural language prompts to trace upstream dependencies, identify sensitive data flow and as needed. Every proposed change is visible before it moves to production, with full undo and redo controls throughout.
Foundational Transparency & Accountability
Stewart Bond, vice president of IDC’s data intelligence and integration software service, has much to say on this subject.
“As AI becomes embedded in critical business processes, audit-ready data lineage is no longer a nice-to-have; it is a foundational requirement for transparency and accountability across complex data environments,” said Bond. “Regulators and boards increasingly expect organizations to demonstrate not just what decisions were made, but exactly what data underpinned them, where that data came from, how it was transformed and whether it can be trusted.”
Bond underlines his statement by saying that delivering that level of evidence at the speed modern enterprises require means rethinking how lineage is built and maintained. He thinks that approaches that combine AI-driven automation with human validation are emerging as the practical path forward for producing lineage that is fast to generate and defensible under scrutiny.
Solidatus has built its AI Lineage Assistant to go beyond structured metadata, ingesting legacy PDFs, spreadsheets, images and documentation to create interactive, queryable lineage.
For organisations building AI, this enables provenance tracing back to authoritative sources, a growing requirement under the EU AI Act. The assistant can also be deployed using customers’ own enterprise LLMs, keeping metadata within their governance boundary and ensuring full data sovereignty.
