Every hospital letter.
Every action.
Every patient safe.
Care Continuo AI reads incoming hospital letters, extracts every follow-up action, and presents them for GP approval - all inside SystmOne, in the screen you already use.
GP correspondence is a patient safety issue hiding in plain sight
Every unread, misrouted, or delayed hospital letter is a potential gap in a patient's care. And with volumes growing year on year, the status quo isn't sustainable.
Administrative overload
Practices manually process each letter: reading, interpreting, coding, and routing entirely by hand. Admin staff without clinical training must identify what needs acting on, creating backlogs and errors.
Cognitive burden
GPs must reconcile unstructured prose against medication lists, pathology results, and referral history under time pressure. Important actions buried five paragraphs into a two-page letter are easily missed.
Patient safety risk
Blood tests, medication changes, and recalls hidden in long letters can be delayed or missed. A delayed INR check or missed two-week-wait flag has direct patient harm potential.
The AI reads the letter. You make the decisions.
Care Continuo AI extracts actionable tasks from each letter and presents them for your review inside the SystmOne screen you already use to file correspondence. Nothing is actioned without your explicit approval.
Letter arrives in SystmOne
Hospital letters land in your Documents Inbox as normal. Nothing changes in your existing workflow up to this point.
AI extracts every actionable task
Our model reads the letter and identifies blood tests, medication changes, referrals, and recalls - each with a confidence score so you know how certain we are.
You approve or reject each task
A new panel in the Scan Action screen shows each suggestion. You approve, reject (with optional reason), or undo - one click each. Referrals open a pre-drafted letter for review.
Governance gate releases - letter filed
The "Complete" option is locked until every suggestion is reviewed. Once all actions are decided, click OK and the letter files to the patient record with a full audit trail.
Everything a GP practice needs to process correspondence safely at speed
Instant task extraction
Blood tests, medication changes, referrals, and recalls identified automatically from any hospital letter, each with a confidence score in green (≥90%) or amber (<90%).
Pre-drafted referral letters
Referral tasks open a fully editable letter pre-filled from the hospital correspondence - destination, priority, clinical summary and all. Review, edit, approve or reject before it's sent.
Delegate to the right person
Send tasks to any team member - nurse, pharmacist, reception, ANP - immediately or on a scheduled date. Set recurring reminders for monitoring recalls at 2 weeks, monthly, or 3-monthly intervals.
Designed to feel like SystmOne - because it is
Native chrome, native fonts, native buttons. Care Continuo AI doesn't look bolted on because it isn't - it's woven into the screen GPs already use for every letter.
Built for NHS clinical safety standards from day one
Not an afterthought. Every design decision in Care Continuo AI is built around human oversight, auditability, and the governance expectations of DCB0129 and DCB0160.
Human-in-the-loop, always
The AI never actions anything automatically. Every suggestion requires an explicit Approve or Reject decision by a named staff member before it enters any workflow.
Mandatory governance gate
The "Complete" option is disabled until every AI suggestion has been reviewed. It is physically impossible to file a letter with unreviewed suggestions still pending.
Referrals require deliberate review
Referral tasks cannot be bulk-approved. Each opens a full review modal with the AI-drafted letter. The GP must read, potentially edit, and consciously approve before anything is sent.
Full audit trail
Every approve and reject decision is timestamped and attributed to the reviewing clinician. Rejection reasons are recorded. This audit trail is preserved in the patient record.
Confidence-scored suggestions
Every suggestion shows its confidence level. Tasks below 90% are flagged in amber, prompting extra scrutiny. The score is always visible before any decision is made.
Administrative AI - not diagnostic
Care Continuo AI extracts follow-up tasks from letters. It does not interpret clinical results or generate diagnoses. This distinction is deliberate and is central to our regulatory positioning.