Listen to the demonstration

Nigerian English narration · Microsoft Abeo · 12 chapters with synchronised captions.

Audio credits

Walkthrough: Microsoft Abeo, English (Nigeria), en-NG-AbeoNeural. Staged incident report: Microsoft Ezinne, English (Nigeria), en-NG-EzinneNeural. Both recordings use AI-generated speech from Microsoft’s online read-aloud service, produced with the edge-tts client. PollCommand supplied the scripts and assembled the chapter recordings, captions and playback.

The recordings are saved audio files. Playback does not use browser speech synthesis. These voice credits identify the technology; they do not imply endorsement by Microsoft or a human speaker.

Staged incident report

Aisha is a fictional character. Her recording begins with an audible disclosure: “Staged demonstration. Synthetic voice, not a real field recording.” The matching image is an AI-generated reconstruction. Neither is real election evidence.

Download staged report · Read its transcript

Transcript

One national picture. Named responsibility in every state. Where are reports overdue, which evidence needs review, and who must act? Start with clearly labelled example figures across eight states; the other entries are not included in this demo. Then open three competing Kaduna cases and see local action, escalation and reviewed outcomes.

Drill from the state into the local government area, ward and polling unit. In Kaduna, Salisu oversees the state, Hadiza the local government area, and Musa the ward. Aisha is the proposed replacement agent. The national directory is accessible, but a place in the directory does not mean an agent is appointed.

Organisation begins weeks before polling day. Track appointments, training attendance, contact drills and daily coordinator briefs. In the final days, check deployment acknowledgements and outstanding gaps. Here, the previous agent withdrew; Aisha is prepared, but her deployment still needs a recorded local decision.

Aisha reports her readiness. Musa, the ward coordinator, proposes her replacement deployment. Hadiza, the LGA coordinator, approves within delegated authority. Aisha acknowledges, and Musa reviews the updated roster. Salisu, the state coordinator, includes the outcome in his daily brief. National command sees the decision; overdue or exceptional cases are escalated, while each coordinator acknowledges responsibility.

Review the named agent, account details and amount. Confirm a five thousand naira test payment, then compare the matching command and agent receipt. No money moves. The account name is simulated, not genuine bank verification. Live transfers remain disabled.

Reach the field through PollCommand, WhatsApp sharing or verified Direct SMS. An explicit response is recorded separately from delivery. WhatsApp sharing still requires the sender to choose and confirm the recipient. This public demonstration sends no external messages.

On election morning, Aisha reports that staff are present but materials have not arrived. Her unit has not opened. The timestamp, unit and sender stay attached to the report. Another unit has not reported; its status remains unknown.

Musa follows up locally. Hadiza coordinates wider support and escalates the exceptional delay through Salisu. Aisha reports materials received and voting begun at eight thirty-five. Musa records a separate review. National command sees the outcome without becoming the approval desk for routine ward work.

Inspect one coherent incident pack. Aisha’s staged voice report, the matching photo reconstruction, Musa’s follow-up and the reviewer’s decision all concern the same unit and delay. These are demonstration reconstructions, not real election evidence. A file received is not a claim proved.

After counting, Aisha submits the sample result document. It contains three hundred valid votes and six rejected ballots. The original entry incorrectly says three hundred and six valid votes. Submission means received for review, not accepted as an official result.

Zainab flags the difference. Aisha corrects the transcription to three hundred, and Zainab reviews it independently. The original, correction and source remain visible. Now compare three reviewed units: three hundred, two hundred and forty, and one hundred and eighty. Their subtotal is seven hundred and twenty.

The ward and local government summaries say seven hundred and twenty-six. Hold the roll-up, assign reconciliation, retain both versions, and review the corrected seven hundred and twenty once. Then propose a fourteen-day supervised pilot: two states, four local government areas, eight wards and forty-eight units. Measure timely reports, delegated decisions, evidence, delivery and recovery before considering national scale.