The missing line should record whether each farmer received safe, understandable guidance before using a funded input, and whether uncertain or dangerous questions reached a qualified person. Without that evidence, a programme can prove delivery while remaining unable to prove safe use.
In January 1986, engineers at Morton Thiokol faced a similar gap between what had been supplied and what people understood before acting. The space shuttle Challenger was scheduled to launch from Kennedy Space Center in Florida during unusually cold conditions. Engineer Roger Boisjoly and colleagues warned that low temperatures could compromise the solid rocket boosters’ O-ring seals.
The warning existed. The people, hardware and launch procedures were all accounted for. Yet the decision process failed to turn technical concern into a safe outcome.
Challenger launched on 28 January 1986 and broke apart 73 seconds later, killing all seven crew members. The presidential commission led by William P. Rogers concluded that the O-ring failure caused the accident and examined the decision-making failures surrounding the launch. The findings remain available in NASA’s Rogers Commission archive.
Delivery records stop too early
A cocoa programme officer may have a complete ledger: farmer enrolled, input purchased, quantity distributed, signature collected. Those records answer an important financial question. They show where programme funds went.
They do not show what happened in the minutes before a farmer opened the container.
Did the farmer hear the dosage in Asante Twi? Did the response explain protective equipment, re-entry time and the pre-harvest interval? Could the farmer understand the voice? If the guidance was missing or uncertain, did someone tell the farmer to wait and contact an extension officer?
A signed distribution sheet cannot answer those questions. Neither can a photograph of bags arriving at a community handover.
This is the mechanism the Challenger story exposes. An organisation may document every physical component and still miss the evidence connecting a warning to the decision that follows. In cocoa programmes, the stakes are different, but the reporting flaw has the same shape: possession is recorded more carefully than safe action.
The line belongs in the programme design
The additional line should be defined before distribution begins. Retrofitting it after a reporting deadline usually produces anecdotes, incomplete call records or a survey asking farmers what they remember weeks later.
For each guidance interaction, the programme needs a small, auditable record:
- The farmer’s question was answered from reviewed content or correctly escalated.
- The response was delivered in language the farmer could understand.
- Mandatory pesticide safety information was included when relevant.
- The farmer confirmed understanding, or the interaction recorded that comprehension was uncertain.
- A named extension officer received questions the system could not answer safely.
- The programme measured how long those escalations remained unresolved.
These records should protect privacy. They do not require a permanent archive of every voice message. Consent, audio retention, phone-number storage and deletion rules must be established before the pilot, with responsibilities agreed under Ghana’s data-protection requirements.
They also require honest failure states. AgriVoice currently withholds three chemical guidance blocks because dosage, re-entry and pre-harvest information still needs approval from a Ghanaian agronomist. Silence or escalation is safer than completing the report with fluent, unverified advice. Kofi’s pesticide question and the six gates behind a safe answer shows why each gate matters.
Measure understanding, not message delivery
A platform log can confirm that audio was sent. That does not establish that the farmer understood it.
Neuralis proposes a two-week AgriVoice pilot with 20 to 50 Asante-Twi-speaking cocoa farmers. The scorecard separates technical delivery from useful communication. At least 80 percent of participants should report understanding the response. At least 70 percent of questions should be answered or correctly escalated. Any unsafe pesticide instruction reaching a farmer is a stop or redesign signal.
Repeat use matters too. If farmers ask one question during a demonstration and never return, the programme has measured novelty. If at least 30 percent of activated farmers ask another question in the second week, that begins to show practical value.
The same discipline applies to human escalation. A report should name the extension officer responsible and measure response time. “Referred to an expert” means little when no person owns the queue.
Add the evidence before funding the next input
Climate-finance reporting becomes more credible when it follows the farmer beyond receipt. Add a field for guidance status beside the fields for input, quantity and value. Define acceptable evidence. Test it with real farmer questions, including ambiguous requests, code-switching, poor speech recognition and pesticide cases.
Then inspect the failures. A confident wrong answer matters more than a polished completion rate.
The Rogers Commission did more than identify a damaged seal. It documented how information moved, how concerns were assessed and how the launch decision was made. Cocoa programmes need the same depth of inquiry before claiming that funded inputs produced safer or more resilient farming.
For the next distribution, require four outcomes in the record: reviewed guidance delivered, understanding checked, uncertainty escalated and escalation resolved. That one line turns an inventory trail into evidence about what the farmer was actually prepared to do.
Comments
No comments yet.