Your Best Work Is Buried on the P: Drive.
We Make the Archive Findable Again.
Past projects, details, precedents and fee proposals are spread across SharePoint, an old file server and the OneDrive folders of people who have since left. We connect them into one search for the whole practice, and put AI on top that answers from your own work and names the drawing, report or email each answer came from.
For Architecture, Interior Design and Planning Practices
A practice sells its experience. It should be able to find it.
The Problem
Most practices moved to SharePoint and took half the archive with them. The rest sits on a file server nobody wants to switch off, and finding the detail used on a similar scheme five years ago means asking whoever ran the job, if they are still here.
In the practices we talk to, the registers are just as scattered. Planning conditions, heritage consents, building regulations and fire safety items live in a spreadsheet per project. Issue sheets are retyped for each issue. Few people could say which folders are shared with which clients and consultants.
Bids and fee proposals start from whichever old one someone happens to find.
What We Do
We build one search across SharePoint, legacy servers and OneDrive, by keyword and by meaning. Ask for "curtain wall details on residential schemes over ten storeys" and get the documents, the projects they belong to and who worked on them.
Around the search we build the registers a practice runs on: one tracker for planning conditions, heritage, building regulations and fire safety across every live project, issue sheets produced from the document register, and an audit of everything shared outside the practice.
The AI answers only from your archive and always cites its source, so an architect can check the answer before relying on it.
Key Capabilities
- Search across SharePoint, legacy file servers and OneDrive
- Past projects, details and precedents found in seconds
- Planning conditions, heritage, building regulations and fire safety tracking
- Drawing issue sheets generated from the register
- External sharing audit
- A reusable library of bids, fee proposals and project sheets
- New documents classified and filed as they arrive
How We Work
1. Discovery
A fixed-price AI Discovery Day, or a longer paid Discovery for larger practices. We talk to the people who use the archive every day and rank the problems by the time they cost.
2. A Working Prototype
Built on a sample of your own files, so the partners see search results from their own projects before anyone commits to a build.
3. Into Production
The smallest useful version goes live and gets used, then grows. It runs on Xefr, our platform, or natively in your own environment if that suits you better. See how our AI consultancy works.
We built the data mapping and visualisation layer that brought millions of documents together on HS2. The same discipline applies to a practice archive.
Questions We Get Asked
Do we have to move everything into SharePoint first?
No. We search the files where they sit today, whether that is SharePoint, a legacy file server or OneDrive. If you are part way through a migration, new documents can be classified and filed into the right SharePoint location as they arrive.
Is it safe to point AI at our project archive?
The AI answers only from your own documents and cites the file each answer came from, so anyone can check it. Access follows the permissions you already have on your files, and the external sharing audit shows what has left the practice. Xerini holds ISO 27001 and Cyber Essentials certification and is registered with the ICO.
Does it work with drawings?
PDFs, Office documents and email are searched in full, by keyword and by meaning. Drawings are found by title block, file name, metadata and issue sheet. Deeper work on native model files is scoped during Discovery, against the formats your practice uses.
We already have Microsoft Copilot. Why would we need this?
Copilot works inside Microsoft 365. Most practices also hold an archive outside it, and a planning conditions tracker, issue sheets or a sharing audit need structured records as well as search. We build alongside Copilot where you have it.
Can it help with bids and fee proposals?
Yes. Past bid answers, project sheets, CVs and fee proposals become one searchable library, so a new bid starts from the closest previous one and the right project references.
How do we start, and how quickly do we see something working?
With a fixed-price AI Discovery Day or a paid Discovery. We then build a working prototype on a sample of your own files, so you see it working before you commit to a build. Most clients see initial value in three to six weeks.
Name the Project You Can Never Find
Tell us what your team spends the most time looking for. We will show you whether we can find it for them, on a sample of your own files, before you spend anything on a build.