Documents and WorkflowsDocument Intelligence
From letter to evidenced work.
A discharge letter arrives; Reedley identifies diagnoses, medication changes and follow-up, links each to its passage and proposes the work; a clinician reviews, edits and confirms.
- What entersA discharge letter
- What Reedley preparesFindings linked to their passages, and the work they imply
- What the clinician decidesA clinician reviews, edits and confirms
Interactive demonstration
Synthetic patient · pause, step and replay at any point
A discharge summary becomes evidenced work: findings linked to their passages, a clinician’s review, tasks with owners, and an audit trail.
Brian CALDWELL NHS 999 000 0611
Born 09 Feb 1951 · 75 y · Male · As of 27 Oct 2026 16:00
A discharge summary becomes evidenced work: findings linked to their passages, a clinician’s review, tasks with owners, and an audit trail.
Step 1 of 8 · Discharge summary arrives · Continues as you scroll
- Discharge summary arrives
- Identity matched
- Findings identified
- Each linked to its passage
- Compared with the record
- Clinician reviews each item
- Tasks in their workflows
- Auditable outcome
Received 26 Oct 2026 09:14 · MESH (simulated)
Discharge summary · Brian CALDWELL
Respiratory ward · Northfield General Hospital (fictional) · Dr L. HARGRAVE
IdentifiedDischarge summary received via MESH (simulated). Document review item identified. Received is not reviewed.
Routed by a practice rule“A discharge summary goes to a prescriber for review within 1 working day of receipt.”Practice routing rule · draft · unsigned
Evidence foundPatient identity matched
| Matched on | Discharge summary | Practice record |
|---|---|---|
| Name | Brian CALDWELL | Brian CALDWELL |
| Date of birth | 09 Feb 1951 | 09 Feb 1951 |
| NHS number | 999 000 0611 | 999 000 0611 |
Discharge summary header · dated 23 Oct 2026
Identity matched on name, date of birth and NHS number. Identity is confirmed before anything is extracted.
Opened alongside · Medication record · as at 26 Oct 2026
- Metformin 500mg tablets · 1 g twice daily
- Gliclazide 80mg tablets · 80 mg twice daily
- Atorvastatin 20mg tablets · 20 mg at night
- Ramipril 5mg capsules · 5 mg once daily
Problem list
- Type 2 diabetes mellitus recorded 3 Apr 2012
- Essential hypertension recorded 29 Sep 2015
11 findings from one discharge summary: 2 diagnoses, 4 medication changes, 4 follow-up requirements and 1 conditional instruction for a clinician.
2 Diagnosis4 Medication change4 Follow-up requirement1 Conditional · for a clinician
You are the reviewer: accept, edit or reject each proposal. Nothing is chosen for you. The review recorded by K. AHMED follows in the next step.
Suggested concept · synthetic terminology in the development build · confirm against the NHS SNOMED CT browser before launch
Diagnosis
Community-acquired pneumonia (right lower lobe)
Discharge summary ¶1 · dated 23 Oct 2026
In the record: Not on the problem list
Suggested concept: Community acquired pneumonia · 385093006
Proposed: Coding proposal · Admin team (summarising) · due 2 Nov
Clinician options: Accept · Edit · Reject. The choice is the clinician’s.
No decision recorded · for K. AHMED (synthetic clinician)
Diagnosis
Atrial fibrillation · new diagnosis this admission
Discharge summary ¶1 · dated 23 Oct 2026
In the record: Not on the problem list
Suggested concept: Atrial fibrillation · 49436004
Proposed: Coding proposal · Admin team (summarising) · due 2 Nov
Clinician options: Accept · Edit · Reject. The choice is the clinician’s.
No decision recorded · for K. AHMED (synthetic clinician)
Medication change
Apixaban 5 mg twice daily · started
Discharge summary ¶3 · dated 23 Oct 2026
In the record: Not on the medication recordMissing from the record
Dose as written in the letter. Reedley does not assess doses against the product information.
Proposed: Medication reconciliation · K. AHMED, clinical pharmacist (prescriber) · due 28 Oct
Clinician options: Accept · Edit · Reject. The choice is the clinician’s.
No decision recorded · for K. AHMED (synthetic clinician)
Medication change
Bisoprolol 2.5 mg once daily · started
Discharge summary ¶3 · dated 23 Oct 2026
In the record: Not on the medication recordMissing from the record
Proposed: Medication reconciliation · K. AHMED, clinical pharmacist (prescriber) · due 28 Oct
Clinician options: Accept · Edit · Reject. The choice is the clinician’s.
No decision recorded · for K. AHMED (synthetic clinician)
Medication change
Gliclazide 80 mg → 40 mg twice daily
Discharge summary ¶3 · dated 23 Oct 2026
In the record: Gliclazide 80mg tablets · 80 mg twice dailyDose differs medication record · as at 26 Oct 2026
Proposed: Medication reconciliation · K. AHMED, clinical pharmacist (prescriber) · due 28 Oct
Clinician options: Accept · Edit · Reject. The choice is the clinician’s.
No decision recorded · for K. AHMED (synthetic clinician)
Medication change
Amoxicillin 5-day course · last dose 24 Oct
Discharge summary ¶3 · dated 23 Oct 2026
In the record: Not on the medication record
Information only
Information only: the course ended on 24 Oct, before the letter was received. No proposal is made.
Proposed: Investigation · Admin team · due 6 Nov
Clinician options: Accept · Edit · Reject. The choice is the clinician’s.
No decision recorded · for K. AHMED (synthetic clinician)
Proposed: Investigation request · Dr S. PATEL, GP · due 27 NovRequest made by 27 Nov so the film can be taken around 4 Dec, 6 weeks after discharge
Clinician options: Accept · Edit · Reject. The choice is the clinician’s.
No decision recorded · for K. AHMED (synthetic clinician)
Follow-up requirement
Actions for GP box: U&E and FBC; chest X-ray
Proposed: Investigation · Admin team · due 6 Nov
Clinician options: Accept · Edit · Reject. The choice is the clinician’s.
No decision recorded · for K. AHMED (synthetic clinician)
Conditional · for a clinician
Gliclazide: possible further reduction
Clinician review required
Routed to a clinician as a decision · no medication task, no Accept button · Illustrative
To Dr S. PATEL, GP · 26 Oct 11:32
Conditional wording is flagged for a clinician to review and is never treated as an instruction. Routing it to a GP as a decision, with no Accept button, is Illustrative · intended workflow.
Follow-up requirement
Echocardiogram · arranged by the hospital
Information only
Information only: the hospital arranges it; no practice task.
Your review
0 of 8 decided · 1 routed to a clinician
Work in its workflows
As of 26 Oct 2026 11:33
Document review recorded. Document review item closed · completed. The work it created continues separately.
Medicines
Reconcile the medication record with the discharge summary: apixaban, bisoprolol, gliclazide
Assigned · due 28 Oct
From the discharge summary ¶3
Coding
Add atrial fibrillation to the problem list
Assigned · due 2 Nov
From the discharge summary ¶1
Record community-acquired pneumonia as a past episode (edited)
Assigned · due 2 Nov
From the discharge summary ¶1
Results and monitoring
U&E and FBC 2 weeks after discharge
Assigned · due 6 Nov
From the discharge summary ¶4
Request a repeat chest X-ray for around 4 Dec
Assigned · due 27 Nov
From the discharge summary ¶4
Clinician decisions
Decision on the conditional gliclazide wording
Routed to a clinician · awaiting GP decision · Illustrative
From the discharge summary ¶4
Auditable outcome
As of 27 Oct 2026 16:00
Reconcile the medication record with the discharge summary: apixaban, bisoprolol, gliclazide
Assigned · due 28 OctClosed · completed
Add atrial fibrillation to the problem list
Assigned · due 2 NovClosed · completed
Record community-acquired pneumonia as a past episode (edited)
Assigned · due 2 Nov
U&E and FBC 2 weeks after discharge
Assigned · due 6 NovActioned · booked 5 Nov · attendance not yet confirmed
Request a repeat chest X-ray for around 4 Dec
Assigned · due 27 Nov
Decision on the conditional gliclazide wording
Routed to a clinician · awaiting GP decision · Illustrative
Event history · K. AHMED and Reedley
- 26 Oct 09:14ReedleyDischarge summary received via MESH (simulated). Document review item identified. Received is not reviewed.
- 26 Oct 09:14ReedleyIdentity matched on name, date of birth and NHS number.
- 26 Oct 09:15Reedley11 findings identified, each linked to its paragraph; 3 compared with the medication record.
- 26 Oct 11:05K. AHMEDReview opened.
- 26 Oct 11:27K. AHMED7 accepted (2 with edits), 1 rejected as a duplicate, 2 information only, 1 routed to a clinician.
- 26 Oct 11:32K. AHMEDDocument review recorded. Document review item closed · completed. The work it created continues separately.
- 26 Oct 11:33K. AHMED5 work items assigned with owners and due dates; 1 decision routed to Dr S. PATEL.
- 26 Oct 15:40Reedley (Simulated)Medication record re-read: the changes are on the record. Illustrative: verifying and closing a reconciliation item is not in the development build (it waits for a design decision). Issued is not taken.
- 27 Oct 10:12Reedley (Simulated)Problem list re-read: atrial fibrillation coded; item closed · completed.
- 27 Oct 11:05Practice nurseBlood test booked for 5 Nov (reported). Attendance not yet confirmed.
2 work items closed after verification (Simulated) · 4 still open, each with an owner and the reason it is open.
Discharge summary
Respiratory ward · Northfield General Hospital (fictional)
Dr L. HARGRAVE
Patient: Brian CALDWELL · born 09 Feb 1951 · NHS 999 000 0611 · admitted 19 Oct 2026 · discharged 23 Oct 2026
Dr L. HARGRAVE
Diagnoses: 1. Community-acquired pneumonia (right lower lobe). 2. Atrial fibrillation, new diagnosis this admission, rate controlled on discharge.
Dr L. HARGRAVE
Admitted on 19 October 2026 with three days of productive cough, fever and breathlessness. Chest X-ray showed right lower lobe consolidation. Treated with intravenous then oral antibiotics. Atrial fibrillation was noted on admission and persisted; CHA2DS2-VASc score 4. Creatinine 112 µmol/L on admission and 96 µmol/L on discharge. Weight 74 kg on admission. Several capillary blood glucose readings below 4 mmol/L were recorded on the ward.
Dr L. HARGRAVE
Medication changes: apixaban 5 mg twice daily started. Bisoprolol 2.5 mg once daily started. Gliclazide reduced from 80 mg twice daily to 40 mg twice daily. Amoxicillin 500 mg three times daily to complete a 5-day course (last dose 24 October 2026). All other medicines unchanged.
Dr L. HARGRAVE
Requests to the GP: please check urea and electrolytes and full blood count in 2 weeks. Please arrange a repeat chest X-ray in 6 weeks to confirm resolution. Please consider reducing gliclazide further if home blood glucose readings are frequently below 5 mmol/L.
Dr L. HARGRAVE
Follow-up: an outpatient echocardiogram has been requested by the ward; the hospital will send the appointment.
Dr L. HARGRAVE
Actions for GP: U&E and FBC in 2 weeks; chest X-ray in 6 weeks.
Concept suggestions use synthetic terminology. Record re-reads are simulated; Reedley does not change the medication record.
In the development build · synthetic data · Scripted reconstruction: the development build itself is not running on this page.
- Identity confirmed before anything is extracted.
- Every finding quoted from the exact passage it came from.
- Confirmed work lands in the queue the practice’s draft routing rule names, with an audit record.
What it does not do · what exists today
- There is no live document intake: letters in the development build are synthetic.
- The demonstration uses pre-written outputs on synthetic letters. In the current build the findings are pre-written; the review, comparison and proposal steps that follow are implemented and tested on synthetic data. A language model has been run once, outside the product screen, on synthetic development documents through a development harness; the intended production provider has not been used, and the run is not evidence of accuracy.
- Extraction accuracy will be measured before any claim is made about it.
Every patient shown is synthetic. No integration with a clinical system is claimed, and no clinician has yet used Reedley in practice. Safety and governance
See it with your own work in mind.
Tell us your role and the workflows you care about, and we will walk you through them on synthetic data.