Case study · Healthcare
Less typing, more time with patients
Clinicians spend hours writing letters and summaries after consultations. This is how we built an AI assistant that drafts those documents from consultation notes for the clinician to review and sign, with patient data kept in a secure environment.
Project overview
Clinician-reviewed AI drafting.
- Industry
- Healthcare
- Client
- Confidential
- Solution
- Documentation assistant
- Platform
- Azure OpenAI (private)
- Timeline
- 10–16 weeks
- Team
- 4 people
- Documents
- Letters, discharge summaries
- Review
- Clinician signs every document
- Data
- Stays in the client's tenant
Expected business outcomes
What the client set out to achieve
Measured in a clinical pilot with clinician feedback on every draft.
-
30–50%
Less time writing letters
Clinicians edit a draft instead of starting from blank.
Target range; measured in a pilot with timing
-
100%
Drafts reviewed by a clinician
Nothing is sent without clinician review and sign-off.
Enforced by the workflow
-
Tracked
Draft quality ratings
Clinicians rate drafts, feeding improvements.
Reviewed with the clinical lead
Business challenges
Where the client started
Documentation eats into clinical time.
Letters
Written or dictated after each consultation.
Backlog
Letters wait days for typing and checking.
Consistency
Structure varies by clinician.
Burnout
Admin after clinic hours is a top complaint.
Caution
Clinical leaders are rightly careful about AI and patient safety.
Proposed solution
What the assistant covers
Drafting only, with clinicians in control.
-
Draft generation
Letters and summaries drafted from notes using the clinic's templates.
-
Review workflow
Clinicians edit, approve and sign before anything is sent.
-
Safety checks
Drafts flag missing information and never invent medications or results.
-
Data protection
Processing in the client's secure Azure tenant with no data used for training.
-
EHR integration
Notes read from and signed letters saved to the practice system.
-
Quality feedback
Clinician ratings and edits monitored by the clinical lead.
Project requirements
What the client asked us for
The request was to reduce documentation time without compromising safety. The assistant drafts; the clinician decides.
A clinical safety review and data protection assessment come before the pilot.
What the project had to deliver
- 01Drafts, not decisions
- 02Clinician sign-off on everything
- 03Patient data protected
- 04Quality monitored
Key challenges & considerations
What made it harder, and what we did
-
Clinical safety governance
What we did
Clinical safety governance was completed before any clinician used the assistant.
-
Integration with the practice system
What we did
Notes were written back to the practice system through its supported interface.
-
Specialty-specific templates
What we did
Templates were built per specialty with the clinicians.
-
Clinician trust
What we did
The clinician reviewed and approved every note, and nothing was filed automatically.
What we watched for
- Complete safety governance first
- Keep the clinician in control
- Use specialty templates
- Monitor quality continuously
Development process
Safety, pilot, scale
Governance is complete before any clinician uses it.
-
Governance
Clinical safety and data protection assessments.
-
Templates
Letter templates agreed per specialty.
-
Build
Drafting, review workflow and integration built.
-
Clinical pilot
Pilot with a small group of clinicians, rating every draft.
-
Scale
Rolled out by specialty with ongoing monitoring.
Technology stack
Private generative AI
Models in a secure tenant, integrated with clinical systems.
AI
- Azure OpenAI
- Retrieval-augmented generation
- Prompt and output guardrails
Data
- Azure AI Search
- Document pipelines
- Vector index
Application
- Python / .NET services
- Web or Teams interface
Evaluation
- Test question sets
- Quality and safety metrics
- Usage monitoring
Quality assurance & deployment
Measured against a test set before launch
AI quality was measured, not judged from a demo. A test set of real examples with agreed answers decided whether the system was ready, and the same set was re-run after every change.
- A test set of real examples with agreed answers
- Accuracy thresholds agreed with the business before launch
- Safety checks: refusals, permissions and prompt-injection tests
- A person reviewing low-confidence output
- Cost and response time measured at real volumes
Deployment
- 01Clinical pilotPilot with a small group of clinicians, rating every draft.
- 02ScaleRolled out by specialty with ongoing monitoring.
- 03Pilot, then staged rolloutA small group of real users first, with results reviewed weekly before widening access.
- 04Monitoring and warrantyAccuracy, drift, usage and cost tracked after launch, and six months of free bug fixing.
Final deliverables
Documentation that doesn't follow clinicians home
Clinicians review and sign well-structured drafts instead of writing from scratch, and letters go out sooner.
Every document is still the clinician's, reviewed and signed by them.
- Faster letters
- Clinician sign-off
- Protected data
- Monitored quality
Handed over to the client
- 01Production AI serviceDeployed in the client's own cloud tenant
- 02Source code and modelsIn a repository the client can access
- 03Test set and resultsThe evidence behind the launch decision
- 04MonitoringAccuracy, usage and cost
- 05Six-month warrantyFree bug fixing after delivery
Success metrics
How success is measured
| Measure | Outcome | How it is measured |
|---|---|---|
| Less time writing letters | 30–50% | Target range; measured in a pilot with timing |
| Drafts reviewed by a clinician | 100% | Enforced by the workflow |
| Draft quality ratings | Tracked | Reviewed with the clinical lead |
Verified reviews
What clients say about our custom software work
Verified reviews from our clients on the same kind of work, published on Clutch.
-
Custom Software Development
Web-Based Coupon App Development
The work produced by the team was of exceptional quality.
Jon Thies
CTO, Model Rocket Feb 2022 – Mar 2023
-
Custom Software Development
Mobile & Web Platform Development & Design
Silver Scintilla was responsive, respectful, and enjoyable to work with, quickly addressing any questions or requests.
DeShawn Brown
CEO & Founder, Lithios Jun 2021 – Nov 2022
-
Custom Software Development
Web-Based HRM System Development
Silver Scintilla provided the best solution with scalable functionality, personalized features, and robust security.
Suheb Khan
Director, Collaborative Insight Technologies Jul 2020 – Jun 2022
Free discovery call
Have an idea? Let's turn it into AI-powered software.
Book a free discovery call with our experts. Share your idea and we will help you shape the scope, timeline and budget, under NDA.
- Free consultation
- NDA before we talk
- Transparent estimate