You make scanned drawings and legacy records searchable with AI by bringing them into storage your organisation controls, extracting their text with optical character recognition (OCR), indexing that text with its metadata, and using retrieval-augmented generation (RAG) to answer questions with citations. CDE Migrate™, CDE Archive™ and CDE Backup™ bring the records into your own Azure tenancy, and AI Insights in the Utopiaverse™ applies RAG and OCR search across them.
Why legacy records matter
Asset owners and long-standing contractors hold decades of drawings, specifications, letters and reports. Much of it was scanned from paper, exported from retired platforms or saved as image-based PDFs. It remains essential. Upgrades, maintenance, heritage works and claims all depend on knowing what was built and why. Yet most of it is effectively invisible, because keyword search cannot read an image.
What OCR does, and its limits
OCR reads the text in an image and turns it into searchable text. Modern OCR handles typed documents well and copes with many drawing title blocks, notes and tables. Handwriting, faint scans and dense drawings are harder, and results vary. Good practice is to keep the original image beside the extracted text, record the OCR confidence where available, and always let people open the original to check. Our article on preparing CDE data for AI covers this and related steps.
How RAG builds on OCR
Once the text is extracted, it is split into passages and indexed by meaning. When someone asks a question, RAG retrieves the most relevant passages and passes them to the AI model, which writes an answer and cites the documents it used. For scanned drawings, title block details such as drawing number, revision, date and project are especially valuable, because they let the AI filter and order results. Our article How Does RAG Search Work on Construction Documents? explains RAG in more detail.
How to set it up, step by step
Identify where legacy records sit: retired CDEs, file servers, scanning projects and archives.
Choose the Azure tenancy and region using the global data region map.
Bring the records into Azure with CDE Migrate™ for retired platforms, CDE Archive™ for closed projects, or CDE Backup™ for file servers still in use.
Apply OCR and extract title block metadata where possible.
Map metadata to a common schema with your current records.
Index the content and pilot with real questions from engineering and operations teams.
The network server to Azure Blob Storage guide covers bringing file server content across.
Example questions
Which drawings show the original foundations of the 1970s grandstand?
What load rating was specified for the bridge when it was strengthened?
Where are the original specifications for the tunnel ventilation system?
Which letters discuss the heritage requirements for the station façade?
Each answer cites the scanned document, so the person asking can open the original image and check it.
Build your own AI or use AI Insights
With the records in your tenancy, your data team can build its own OCR and RAG pipeline with Azure AI services, and can expose the index to approved AI assistants through an MCP server. AI Insights provides RAG and OCR search ready to use, running natively on Azure in your tenancy, across legacy records and current CDEs together.
Security and residency
Legacy records can include sensitive infrastructure details. Keep them in your own tenancy and region, and choose a provider that does not store files in transit. Utopia Digital's products stream files without keeping copies, encrypt data in transit and log every action, and Utopia Digital holds SOC 2 Type II and Cyber Essentials. Evidence is in the Security Trust Portal.
Legacy records and the Utopiaverse™
In the Utopiaverse™, scanned and legacy records sit beside current content from Forma, Aconex, ProjectWise, Procore, SharePoint and other CDEs in one connected record in your own Azure. AI Insights answers questions across old and new records with RAG and OCR search and cited sources. The file viewer opens scanned drawings and current documents in one place, and dashboards report on the record across the portfolio.
The organisation gains visibility of its full history, across every project and platform, and knowledge locked in paper and retired systems becomes part of everyday work.
See RAG and OCR search in the monthly live demo, or talk to the team about your legacy records.