Case Study
SafeFlow
Voice AI Absence Reporting
We built SafeFlow for a 1,200-pupil secondary school: a production absence-reporting platform that replaces voicemail with structured voice AI capture, verifies every report with the registered guardian, and writes approved absences straight into the MIS.
"Parents love it. They call at 6am before work, or whenever suits them. And the SMS confirmation gives them peace of mind that we've received the message."
Attendance Officer, West Midlands Secondary
The Challenge
- 1Two staff members juggling 40-50 calls in a 90-minute morning window
- 2Parents frustrated by engaged tones during peak times
- 3Manual data entry causing ~12% error rate in absence records
- 4Registers couldn't close until all voicemails were processed (10:30am)
- 5No verification that caller was actually the child's guardian
- 6Double-entry required: spreadsheets first, then MIS system
The Solution
- Voice AI system capturing child name, year group, and absence reason 24/7
- Automatic matching against school's pupil directory
- SMS confirmation sent to registered guardian with secure reporting link
- Confidence indicators flagging calls needing human review
- Direct MIS integration writing approved absences to registers automatically
- Fraud detection flagging unauthorised absence reports before they reach the register
01 / Results
Measurable Impact
Measured after 3 months in production at a West Midlands secondary school.
| Metric | Before | After |
|---|---|---|
| Missed calls | 15-20/day | 0 |
| Avg call duration | 3-4 mins | 90 seconds |
| Data entry errors | ~12% | <1% |
| Register close time | 10:30am | 9:15am |
| Staff hours saved | n/a | 8 hrs/week |
02 / Platform
Key Features
Voice AI Capture
Parents call the same number they always have. SafeFlow captures child name, year group, and absence reason as structured data in 90 seconds, down from 3-4 minutes per manual call.
Safeguarding Built-In
Every report triggers an SMS to the registered guardian with a secure link to flag calls they didn't make. This caught fraud attempts in the first term of deployment.
Staff Dashboard
Purpose-built review queue with fraud alerts, match confidence scores, and one-click approval, plus analytics on trends, completion rates, and notification delivery.
03 / Dashboard
Staff Interface
A custom dashboard built around how attendance staff actually work: review queue, reporting, and analytics in one place.
Review Queue
Fraud alerts and high-priority reports surfaced first for immediate action.
Reports
Search, filter, and export all absence reports with match confidence.
Analytics
Report trends, confidence distribution, and notification performance.
04 / Technical Build
Safeguarding Infrastructure
Engineered for compliance and security first, then shaped around real school workflows.
Pupil Directory Matching
Fuzzy matching against the school's pupil database with a confidence score on every report. "No Match" entries are never auto-approved; they route to staff for review.
Fraud Detection System
Deterministic rules flag withheld caller IDs, unusual call times, and unverified callers. A priority queue surfaces high-risk reports for immediate staff attention.
MIS Integration
Direct write-back to SIMS, Arbor, and Bromcom. Approved absences land in registers automatically, eliminating spreadsheet double-entry entirely.
Fraud Prevention
Every SMS confirmation includes a secure link for guardians to flag a call they didn't make. Within the first term, this closed-loop verification caught two instances of older siblings attempting to report themselves "sick".
"That's a safeguarding feature we didn't know we needed. It closes a vulnerability in our attendance process."
Deputy Head