The problem
AI is only as useful as the information it can reach. Most businesses keep that information in many places: shared drives, old email, spreadsheets, accounting software and a few laptops. Files are duplicated, out of date or named in ways only one person understands.
Point an AI tool at that and you get confident answers built on poor material. Getting the data in order first makes everything that follows work better.
What JadeIT does
A data audit
We find out what data your business holds, where it lives and who looks after it. We also check its quality: what is missing, duplicated or out of date. You get a plain summary, not a technical report.
Clean-up and organisation
We help you tidy up. That means agreeing sensible names and folders, removing duplicates, archiving what is no longer needed and fixing the records that matter most. We do it step by step, so day-to-day work carries on.
A searchable knowledge base
We make your company documents searchable and ready for AI. Staff can ask a question in everyday language, such as “what is our returns policy for corporate clients?”, and get an answer that points back to the source document. Access follows your existing permissions, so people only see what they are allowed to see.
What you get
- A clear map of your business data.
- Cleaner, more trustworthy records.
- Faster answers for staff, without hunting through folders.
- A solid base for AI tools that give reliable results.
- Less personal information kept than you need to keep.
Is your data AI-ready?
AI can only work with data it can find, trust and read. Our AI-ready data check looks at six things.
- Findable
- Do you know what data you have and where it lives?
- We catalogue it with DataHub or OpenMetadata.
- Accurate and complete
- Is it correct, with no gaps or duplicates?
- We profile it with DuckDB and test it with GX Core.
- Consistent
- Does the same thing mean the same thing everywhere?
- We agree shared definitions and test them in dbt.
- Current
- Is it up to date, and does someone own it?
- We track freshness and ownership in OpenMetadata.
- Permitted
- Is personal information identified, and may it be used under POPIA?
- We find it with Presidio.
- Readable by AI
- Can AI actually read your documents?
- We check them with Docling and Apache Tika.
You get a short report with a score for each check and the first fixes to make. See the tools on our Foundations page.
Ask for an AI-ready data check
How it works
For IT leads, in short.
- Inventory. We catalogue sources such as file shares, SharePoint or Google Drive, mailboxes, databases and line-of-business applications, with owners, volumes, formats and sensitivity.
- Quality checks. Profiling for duplicates, gaps, stale content and inconsistent formats, with a prioritised fix list.
- Ingestion. Documents are extracted, split into passages and indexed for both keyword and semantic search. We have long experience with Elasticsearch and use it where it fits.
- Permissions. Source access controls are carried into the index so that results respect them.
- Retrieval. AI answers are grounded in retrieved passages and cite their sources, which makes them easy to check.
Curious what state your data is in? Get in touch for a first conversation about where to begin.