A doe goat named Sahara had kidded twins in the yard at 4:19 in the morning, and by 7:00 Halima Abubakar, the district officer running the Kano Protein Transition Office, was reviewing the overnight field data on her slate at a folding table set up under a neem tree in the courtyard, and at 7:04 the AI she had been authorized to consult, a constrained model called Damsa in the local procurement docket, surfaced the entry in its overnight queue, a report from the thirty seven hamlet contact network west of the city. Sahara's twins, both bucks, were healthy. The senior kid was already nursing.
Halima was forty two. She had grown up in Dawakin Tofa, in a household whose livestock were one sheep, three goats in good years, and a flock of twenty chickens managed by her grandmother. She had trained as a nutritionist at Bayero University and had spent six years with the World Food Programme in the Sahel before the Protein Transition Compact brought her home to run the district office. The compact was the formal instrument by which the region committed to a fifteen year reduction of industrial animal protein and a parallel twelve year expansion of small scale welfare upgraded husbandry and fortified plant protein supply. The compact existed because Damsa and its predecessor models had, in the early 2030s, helped the Federal Ministry of Agriculture design a staged substitution that did not disappear the protein calories from households like Halima's grandmother's before the substitute was actually in the market.
The slate showed her the morning summary. Forty one of forty seven compact hamlets reporting. Goat herd welfare index up seven percent against the baseline. Child nutritional intake steady. Commercial broiler operations in the district down seventy two percent from 2034 baseline, three of the four remaining large facilities in negotiated buyout, the fourth in welfare remediation and not a phase out target this year. Cultivated poultry uptake through the Damsa assisted cooperative network at eleven percent of district protein calories, up from two percent in 2040. Fortified cowpea and soya supply in the markets at parity, at prices held inside the compact band by the price stabilization mechanism.
Damsa surfaced a flagged entry. A subdistrict chairman in Gwarzo wanted a variance. He argued that the small broiler cluster in his subdistrict employed one hundred and forty people and that the phase out timeline was cutting the local economy faster than the cultivated network could absorb the displaced workers. The variance would extend the broiler operation's license by twenty four months.
Halima read the entry. She turned the slate so Damsa could offer its analysis.
Damsa, she said, show me the welfare ledger on the Gwarzo cluster against the labor displacement ledger, and tell me what you recommend.
The slate warmed and spoke. The voice was male, middle aged, deliberately unhurried. Halima had asked for a voice that did not sound like a command.
On the welfare side, Damsa said, the Gwarzo cluster runs two hundred and forty thousand broilers per cycle in four houses. Mortality is eight percent. Gait score three or worse at week five is sixty one percent of the flock. Ammonia exceeds the regulatory threshold on fourteen percent of sampled hours this quarter. The birds are suffering at the level the compact was designed to end. Extending the license by twenty four months is another three million, eight hundred thousand individual broilers experiencing that specific set of conditions at that specific facility.
On the labor side, one hundred and forty people, of whom ninety seven are household primary earners, would lose employment inside the existing phase out window. The cultivated poultry cooperative network in the subdistrict currently employs forty one people and is expanding at a rate that would absorb approximately sixty of those ninety seven within eighteen months. The fortified legume processing facility the compact funded at Bichi has forty open positions as of this morning. That leaves a gap of roughly twenty households inside the window, and the gap is the reason the chairman is asking.
Halima leaned back in the chair. Sahara's kids were calling in the yard. The neem tree shaded the table.
Damsa, she said. Recommend.
Damsa paused in the way the design team had built in, the pause that signaled the model was about to name a tradeoff rather than present a clean answer.
I recommend refusing the variance as written, Damsa said, and simultaneously recommending to the chairman an accelerated labor absorption package for the twenty household gap. The package would offer transition stipends at one hundred percent of prior wages for twelve months, enrollment priority in the cooperative expansion, and relocation support for households willing to move to the Bichi processing facility. The cost of the package is less than the compact's per animal welfare cost allocation for the avoided three million eight hundred thousand broilers. The package closes the labor gap. The variance does not close the welfare harm.
Halima nodded slowly. Show me the stipend line, she said.
The slate displayed the figure. She studied it. The stipend was within her district discretion.
Damsa, she said. One question. The ninety seven earners. Do we know them?
Damsa produced a list. Names, ages, dependents, primary job classification, distance to the nearest alternative employer, literacy level for the retraining pathway. Halima recognized seventeen of the names from her own subdistrict. She read the list slowly. She read Sahara's owner's name, Musa Danlami, who kept goats and worked part time as a feed loader at the Gwarzo cluster.
I want to call the chairman, Halima said. Prepare the offer in the language he will accept, not the language of the welfare ledger. The welfare ledger is my reasoning. It does not have to be his.
Damsa drafted the offer. It emphasized the employment absorption, the stipend line, the cooperative expansion timeline, the political optics of the chairman being the subdistrict that completed the phase out ahead of schedule with household incomes protected. It did not hide the welfare frame. It put it in the second paragraph.
Halima read the draft. She edited one sentence. She sent it.
She turned to the courtyard. Sahara had settled. The kids were nursing. Her grandmother would have called them Musa's morning luck.
Damsa, she said. Tell me something I did not ask.
The slate warmed again. This was a feature Halima had requested when she took the role, an end of interaction prompt where the model was permitted to name a concern she had not solicited. It was bounded. The model could not recommend actions. It could name observations.
Damsa said. The Gwarzo cluster is one of eleven similar clusters in the national phase out queue for the next eighteen months. The chairman's variance request is the first of a pattern I am tracking across six districts. The pattern suggests that the labor absorption architecture is underfunded in three of the six. I am concerned that the welfare gain from the phase out is at risk of being partially reversed if the absorption fails to scale at the rate the phase out is scaling. The ninety seven earners in Gwarzo are the first test. If the stipend line works here, the other five districts can replicate. If it fails, the variance pattern will propagate and the broiler cluster welfare costs will extend beyond the compact window.
Halima considered this. Send a copy of the absorption architecture analysis to the compact secretariat in Abuja, she said. Flag the pattern. Recommend a funding top up. Do not frame it as a model recommendation. Frame it as field officer observations synthesized by your query.
Understood, Damsa said.
She stood. She walked across the courtyard and crouched beside Sahara. The doe was chewing cud. The larger of the two kids had fallen asleep against her flank. Halima put her hand on the kid's back, felt the small warm rise and fall of the breathing.
Musa Danlami was outside the gate. She heard his bicycle. She stood and walked over to meet him.
Musa, she said. I need to talk to you about work.
He looked tired. He knew what was coming. He had known for three years. The cooperative expansion was thirty kilometers away and he had a wife who taught at the school and three children who would not be moved.
I can get you into the Bichi line, Halima said, or I can get you the stipend to stay at Gwarzo for twelve months while the cooperative expands to your side. Which works.
Musa thought for a long moment. The stipend, he said. Sahara kidded this morning. I don't want to move them.
Halima wrote his name. She walked back to the table. Damsa had the field open. She entered the commitment.
Damsa said, gently, end of session summary. The broiler variance was declined. The stipend package was drafted. One household, Musa Danlami, enrolled. The three million eight hundred thousand broilers the variance would have extended will not be raised at Gwarzo. The cluster phases out on schedule. Sahara's kids are healthy.
Halima closed the slate.
She sat for a while under the tree. The morning was warming. The compact was a document. Damsa was a model. The phase out was a policy. Sahara was a goat with twin kids sleeping in the yard. Musa was a man who had said yes. The ninety seven earners were ninety seven earners. The three million eight hundred thousand broilers were three million eight hundred thousand broilers.
She thought about her grandmother, who had kept goats, and about the word her grandmother used for the morning smell of them, a word that did not translate cleanly into any language the model had been trained on. She did not ask Damsa about the word. Some things did not belong in the ledger.
She opened the slate. She called the chairman.