How we use AI

AI that supports your judgement — never replaces it

Loup uses AI in six careful, clearly-labelled places — to take the writing and searching off your plate, so managers have more time for people. You stay in control of every decision.

📊

Inspection-readiness review

When a manager asks for it, Loup reviews the records in your system — compliance, training, supervisions, audits, incidents and reviews — against your regulator's framework.

  • A readiness picture for each of your regulator's key areas, using their own rating words
  • The evidence it found, what is going well and what needs work
  • Your three most important actions
  • Progress compared with your previous reviews
⚠️

Incident review

Press AI review on any incident for a second pair of experienced eyes.

  • Possible causes to look into, and safety checks that may have been missed
  • Whether it may need a safeguarding referral or a notification to your regulator — with one click to start the notification
  • Patterns with earlier incidents, such as repeated falls at the same time of day
  • Suggested actions you can add as tasks, and a draft of the lessons learned
✉️

Complaint letters

Draft the acknowledgement and the full response letter from the complaint record in seconds.

  • Warm, clear, plain-English letters in your agency's name
  • Explains how to take the complaint further — the right Ombudsman or regulator for your country
  • Anything it can't know is left in [brackets], with a checklist before you send
  • Edit freely, print on your letterhead and keep it with the complaint
📈

Monthly management report

Turn your figures into the report your directors, commissioners or inspectors want to read.

  • A written summary: what went well, what needs attention, risks and next steps
  • Quotes your real figures — nothing invented
  • Last three months side by side, followed by the full evidence pack
  • Ready to print or save as a PDF
💬

Ask your system

Ask a question in your own words and get an answer from your own records.

  • “Who has training expiring before December?” “Which carers are overdue a supervision?” “How many falls have we had this quarter?”
  • Follow-up questions understood
  • Answers only from what that person is allowed to see
  • Read-only — it can never change or delete anything
💡

Feedback action plans

Once feedback or staff survey results are in, Loup can suggest a practical action plan based on the scores and comments.

  • Clear actions linked to the areas that scored lowest
  • You choose which actions to keep, edit or delete
  • Each action gets an owner and a due date
  • Shows inspectors you listen and act on what people tell you
Our principles

Responsible by design

Only when you ask

AI never runs in the background. Nothing is sent for analysis until someone presses the button or asks a question.

People decide

Every suggestion is reviewed by your team. AI never saves, sends or changes anything on its own — you choose what to keep.

As little personal data as possible

The people you support are shown to the AI by first name and initial only, and health details are never sent.

For managers, by default

AI is for the registered manager and directors. A manager can give it to a named member of the office team, but only after reading a safety notice and recording why, and that's kept on their record. Inspectors with read-only access can't use AI at all.

Guidance, not a verdict

Readiness reviews are an internal self-assessment, not an official rating. Safeguarding and notification prompts support — never replace — your own judgement and your regulator's guidance.

Not used for training

We use a leading AI provider under business terms. Your data is used only to answer your request and is not used to train AI models.

Fair use built in

Each plan includes a monthly AI allowance, so costs stay predictable — no surprise bills.

What the AI sees

Readiness reviews use counts and percentages from your records — for example how many checks and courses are in date, how many supervisions are overdue, how many incidents are open — along with your regulator's framework. No names and no written notes.

Incident reviews and complaint letters use what is written in that record, plus a short list of earlier incidents for the same person so patterns can be spotted. The person is shown by first name and initial.

Monthly reports use the figures in your report — never the names of the people you support.

Ask your system looks up only what each question needs, within that person's access. Staff names and dates are used to answer; contact details, addresses and health information are never included.

Feedback action plans use the survey's scores and written comments. If a form is anonymous, no names are collected in the first place.

Everything is processed securely and used only to answer that request.

For the full detail — feature by feature, what is never sent, and suggested wording for your own privacy notices — see AI and your data. You can also switch AI features off completely in Settings.

AI can make mistakes. Loup's AI output is there to support experienced managers — always check it against your own knowledge of your service.

See Loup working for your service

Sign up today, or talk to us first — we're happy to show you around.

Book a demo Book a demo