An AI that has read your firm, not just the law
Index your own precedents and closed files so answers come back in your house style, from work you have already done. Your documents are never used to train a model.
Can AI be trained on my own firm's documents?
Writford can index your firm's own precedents, closed files and legal documents so the AI answers from them alongside published UK law. Your documents are retrieved to answer your firm's questions. They are not used to train AI models and are never pooled with another firm's material.
- Are my documents used to train an AI model?
- No. Indexing your files so your own fee earners get better answers is a different thing from handing them over to improve a model. Your material is never used as training data and is never pooled with another firm's.
- What difference does firm knowledge make to an answer?
- Published law tells you what the position is. Your own files tell you how your firm handles it: the clause you always use, the wording that worked with a particular lender, the note explaining why you changed approach. The AI can draw on both.
- Can I see how much of the firm has been indexed?
- Yes. The coverage view stays visible after indexing finishes, showing document counts, what is left, roughly how long and which documents failed, rather than a progress bar that disappears.
- Can I switch it off?
- Yes. It is a firm-level switch, the current state is stated plainly on the screen, and turning it off stops new documents being indexed immediately.
- Does it respect matter access permissions?
- Yes. Matter access rules still apply to what the AI will surface for a given person, so indexing never becomes a way around the firm's own permissions.
Answers that draw on work your firm has already done
Published law tells you what the position is. Your own files tell you how this firm handles it: the clause you always use, the wording that got past that lender last time, the note explaining why you stopped doing it the other way. Index your documents and the AI can answer from both.
- Your own precedents and closed files, searchable by the AI
- Answers in your house style rather than a generic one
- Sits alongside live UK legislation and case law, not instead of it
- New documents picked up as they are added
You can tell how much of the firm has actually been read
Indexing a firm's back catalogue takes time, and a progress bar that only appears while something is running tells you nothing afterwards. The coverage view stays visible: how many documents are in, how many are left, roughly how long, and which ones failed and why.
- Document counts and progress, kept visible after indexing finishes
- An estimate of how long the rest will take
- Failed documents listed rather than silently dropped
- Clear whether indexing is running, paused or switched off
Off by default, and yours to switch off again
A firm decides whether its own documents are used this way. The toggle is at firm level, it is honest about which state it is in, and turning it off stops new documents being indexed rather than leaving you guessing.
- A firm-level switch, not a per-user surprise
- The current state is stated plainly on the screen
- Turning it off stops new indexing immediately
Your documents answer your questions and nobody else's
Indexing your files so your fee earners get better answers is a different thing from handing them over to improve a model. Your material is retrieved to answer your firm's questions. It is not used to train models, and it is not pooled with anyone else's firm.
- Never used to train AI models
- Never pooled with another firm's material
- Processed in UK regions, like the rest of your data
- Matter access rules still apply to what the AI will show a given person