AI powered import for handwritten attendance sheets

Mireaux’s Web QMS helps clients in ISO and API heavy industries run audits end to end: agendas, meetings, findings, actions, verification, and records. One slow spot remained: recording who actually attended opening, closing, and other audit meetings. Auditors often pass around a paper sign‑in sheet. Later someone has to type each name into the system. It is boring, error‑prone, and eats time the audit team could spend on real issues. We set out to make paper sign‑ins as fast and reliable as the existing digital template import, without forcing auditors to change a habit that works for them in the room.

Handwritten sheets introduce several problems:

• Messy or rushed handwriting and varied formats (first name only, initials, nicknames).
• Typos and misreads when retyping into the app.
• Risk of adding the wrong person or missing an attendee.
• Duplicate entries across multi‑page sheets.
• No easy way to catch if the count of names is off versus the expected participants.

We needed a focused, trustworthy way to map scribbles to real employee records already stored in Web QMS. Plain OCR alone was not enough: it might read "J. Ramíez" instead of the registered "Juan Ramírez" or confuse similar surnames.

We needed a bulk import that was fast, transparent, and safe, so users would trust the result without line‑by‑line checking.

We built an AI powered import flow inside the Web QMS Audits app:

1. User scans or snaps the handwritten attendance sheet(s) and uploads the images/PDF.
2. OCR extracts raw text blocks.
3. Our matching algorithm feeds the company’s employee roster and context (expected meeting type, location, known attendees if scheduled) into an AI model.
4. The model normalises each raw name (expanding initials, fixing casing and accents) and scores matches against official employee records.
5. Low‑confidence or ambiguous matches are flagged for a quick one‑click confirm or select from shortlist; clear matches auto‑resolve.
6. The system checks for duplicates, missing expected roles, and overall count before final import.
7. Confirmed attendees are written into the Audit attendance record with an audit trail linking back to the source scan.

The result: what used to be tedious manual entry becomes a guided review that usually finishes in a couple of minutes. Data quality improves (fewer spelling errors, consistent naming), and teams gain a clear trace from each digital attendance record back to the original sheet for compliance. Both paper and digital templates now flow through one consistent process, removing the last paper friction in meeting records.