Committee Process Automation: From Emails to Controlled Decisions
- ukrsedo
- 12 minutes ago
- 3 min read
Most committee processes still rely on email.
Submissions arrive in different formats, agendas are rebuilt manually, attendance is guessed, conflicts of interest are tracked in people’s heads, and voting outcomes are reconstructed after the fact. The decisions may be valid — but the process is fragile, opaque, and painful to audit.
This solution fixes that.
What this automation solution is about?
This is a SharePoint + Power Automate–based committee administration solution that structures how committee submissions are:
registered
scheduled
attended
voted on
and formally closed
No AI aids. Full audit trail.
The system does not “decide”. It enforces discipline around how decisions are taken.
What the solution covers
The solution governs four essential parts of any committee lifecycle:
Entry registration and tracking: Submissions are formally registered, status-controlled, and traceable end-to-end.
Agenda formation: Each meeting has a single, dated, approved agenda — no moving targets.
Attendance and conflict-of-interest control: Voting eligibility is determined for each meeting and for each agenda item.
Voting and decision recording: Quorum rules are enforced, votes are counted correctly, and outcomes are recorded for each submission.
Voting can happen before, during, or after the meeting — including controlled closure when time runs out, and the agenda isn’t empty.
Flow-level design
Flow 1 — Entry Registration & Status Control
Purpose: Control the lifecycle of committee submissions.
Trigger: SharePoint item created or modified.
What it does
Registers new submissions
Detects and controls status changes (e.g. New → Scheduled)
Sends acknowledgements and scheduling notifications
Output: A single, authoritative list of committee entries with a clear, auditable status history.
Flow 2 — Meeting Agenda Formation
Purpose: Define what is officially on the table for a specific meeting.
Trigger: Manual run by Secretariat or Manager.
What it does
Identifies the next scheduled committee date
Selects all submissions planned for that date
Orders items by priority
Generates a structured agenda
Distributes it for managerial approval
Archives the approved version
Output: One dated agenda representing the official scope of the meeting.
Flow 3 — Attendance & Conflict-of-Interest Recording
Purpose: Define who is allowed to vote.
Trigger: Manual run with a meeting date.
What it does
Identifies voting members from the committee roster
Sends attendance confirmations
Records Present / Absent responses
Creates or updates attendance records
Records conflict-of-interest abstentions per agenda item
Output: A formal list of present and eligible voters for the meeting.
Flow 4 — Voting & Decision Processing
Purpose: Execute voting and record outcomes correctly.
Trigger: Manual run with:
meeting date
voting phase (pre / in / post-meeting)
run mode (vote or out-of-time close)
What it does
Selects agenda items still requiring a vote
Determines eligible voters per item
Sends approve/reject requests
Counts votes, abstentions, and non-responses
Applies quorum rules (count or percentage)
Handles ties via Chair vote where required
Writes final decisions back to each submission
Defers unvoted items when a meeting is closed
Output: A final decision status per agenda item, with full supporting vote artefacts.
How the flows work together
Each flow does one job — and only one:
Flow 1 controls what exists and when it moves
Flow 2 defines what is on the table
Flow 3 defines who can vote
Flow 4 defines what decision is taken
This separation is intentional. It makes the system auditable, testable, and governable.
What can be improved next?
The core is intentionally conservative. Enhancements can be layered without breaking governance:
AI-assisted submission critique: Flow 1 can be extended to provide structured feedback against corporate or departmental standards.
Better user experience: Adaptive Cards can replace standard emails for cleaner interaction.
Automated protocol amendments: A separate flow already exists to convert Teams meeting transcripts (VTT files) into protocol updates and action lists.
Final thought
You do not need AI to optimise processes.
Most organisational inefficiency comes from undefined states, unclear ownership, and uncontrolled transitions — not from a lack of intelligence. Those problems are solved with basic RPA and workflow design, not machine learning.
AI can enhance a stable process. It cannot fix a broken one.
Start with RPA. Add AI only after the process is under control.






