When a water quality disaster wiped out nearly a third of her mother’s fish pond, pharmacy graduate Ukachi Benita didn’t just offer sympathy. Armed with YouTube tutorials, Claude, and Google’s Gemma AI, she built Aquamanne: a low-cost hardware tool that translates real-time water sensor data into plain-language alerts in Yoruba, Igbo, and Hausa. What started as a ₦30,000 hackathon build to solve a family crisis is now turning into a full-fledged commercial startup, backed by an Oxford lecturer and an embedded engineer, with fish farmers across Nigeria begging to buy it, writes JOHN ADOYI.
By the time Ukachi Benita began asking her mother why her fish kept dying, the losses had become familiar. Fish would die in batches, sometimes leaving thousands unsold. Her mother, who had run the farm for more than a decade, had come to see the losses as one of the risks of fish farming.
Then, in March 2026, more than 3,000 fish died in a pond that held about 10,000.
“It was so bad,” Benita said. “My mom was really hurt.”
Benita had grown up around the farm and was used to hearing about losses, but this one troubled her. Her mother had already spent money on feed and other costs of raising the fish, only to lose a large part of the production. When Benita asked what had happened, her mother suspected the water.
The answer stayed with her.
“You have already spent a lot of money on feeding and other processes and it gets to a point, they will just die,” Benita said. “You cannot sell dead fishes.”
Benita began researching fish mortality and learned that poor water quality can be one of its major causes. Yet despite knowing that water could affect the fish, her mother had no way to continuously monitor what was happening inside the pond.
When a Google Gemma hackathon came around, Benita decided to build one.
The problem was bigger than her mother’s farm
Benita’s mother had long suspected that something in the pond was going wrong. Sometimes she blamed the feed. Other times, she blamed the water. But knowing that water could be responsible for fish deaths was different from knowing when conditions in the pond were becoming dangerous.
Water quality plays a central role in fish health. The Food and Agriculture Organisation (FAO) identifies poor water conditions as one of the factors that can contribute to stress, disease and mortality during fish production. Changes in dissolved oxygen, pH and temperature can all affect whether fish survive.
The problem has also been documented among fish farmers in Lagos. A 2018 study of 72 small-scale fish farmers in two local government areas found that 76.4% identified water pollution as a major cause of disease on their farms. The researchers also identified poor access to good water sources as one of the challenges farmers faced in managing fish health.
Sesan, a catfish farmer in Bariga, Lagos, has seen the consequences himself. He has farmed fish for 12 years and manages 10 ponds with about 8,000 fish. Twice a week, he checks the water’s pH and alkalinity with a pH meter. Even with those checks, he once lost about 30% of his stock after oxygen depletion caused by water pollution.
“The fish started dying immediately after the oxygen depletion,” Sesan said.
He changed the pond water to increase the oxygen level and stop more fish from dying. A system that continuously monitored water conditions, he said, would help farmers respond before a problem became a major loss.
Goodluck Juwan, a fish farmer in Sango Ota, Ogun State, learned that lesson even earlier.
He began farming fish while he was in secondary school in Abeokuta, starting with 500 fish in 2017. They all died.
“I did not know anything like that,” Juwan said, recalling that he was unaware farmers could use tools to check the pH and quality of pond water.
He later learned through agricultural science classes that poor water conditions could reduce oxygen levels and kill fish. He began changing the water regularly and followed advice from books and Google. For a while, that was enough.
Then, in 2021, he lost 5,000 catfish at once.
“I was already preparing for Christmas and how much I was going to sell, and till this day, I don’t know what happened,” Juwan said. “I cried because catfish then, the ones I had were not going to go for less than ₦5,000 for one.”
He cleaned the pond and managed to save about 2,000 fish. Afterwards, he bought a device that measures temperature and pH and began checking the readings regularly. He says it has helped him avoid similar losses.
Juwan now manages ponds with about 10,000 fish. So when he heard about Benita’s idea of a system that could monitor water conditions and tell farmers what action to take, he was interested.
“If there is a tool like that, I want to have it,” he said. “A lot of fish farmers would pay to have it. I will pay for it because there is no way, as a fish farmer, you will not lose money. You have catfish that is expensive to feed and, at the end of the day, it dies. Your money is gone.”
For Benita, conversations like these suggested that the problem extended beyond her mother’s farm. Farmers were already checking their water in different ways, but there was still a gap between knowing that water quality mattered and knowing quickly enough when conditions were changing to act before fish died.
So, during the Google Gemma hackathon, she began building Aquamanne, a fish-pond monitoring tool designed to track water conditions and alert farmers when something changed.
How Aquamanne works
Aquamanne started as a ₦30,000 prototype that Benita built in a week.
During the Google Gemma hackathon, she assembled the first version of the fish-pond monitoring system using an ESP32 microcontroller, jumper wires and a handful of sensors she could source in Nigeria. Some of the sensors she wanted were unavailable locally, so she built the prototype with the parts she could find.
The system was designed to monitor three water conditions that matter for fish survival: pH, dissolved oxygen and temperature. Sensors placed in the pond collect readings and send them through the ESP32 microcontroller to Google’s Gemma model, which Benita ran locally during the hackathon. The model interprets the readings before the information is sent to an app on the farmer’s phone.
Rather than displaying raw numbers alone, Aquamanne is designed to explain what those readings mean. If oxygen levels fall or the pH moves outside a healthy range, the system tells the farmer that the pond’s conditions are changing and suggests steps that could reduce the risk to the fish.
That layer of interpretation is where Benita says AI adds value. A conventional monitoring device can report readings, she said, but Aquamanne is designed to turn those readings into plain language and eventually communicate them through text or voice in African languages.
“The main thing AI does is the language barrier,” Benita said. “A normal sensor might just give you the readings, but the conversion to local languages is not there, the voice capability as well is not there.”
During the hackathon, the prototype supported Hausa, Igbo, Yoruba and Swahili. Benita says the team is now working to improve its language capabilities because the model used during the hackathon did not perform equally well across Nigerian and other African languages.
Building the prototype meant learning hardware almost from scratch. A pharmacist by training who studied at Nnamdi Azikiwe University in Awka, Anambra State, Benita had little experience with embedded systems before the hackathon. She spent days watching YouTube tutorials to understand how the components worked and relied on AI tools, particularly Claude, whenever she got stuck.
“To be honest, it wasn’t very easy because I am not very technical,” she said. “I had to do a whole lot of research building it.”
The Aquamanne Benita demonstrated at the hackathon is also different from the product she is now trying to build. The prototype worked offline through Bluetooth, with her laptop acting as the host for the local AI model. That made it suitable for a demonstration on her mother’s farm, but not something that could be left running beside a pond indefinitely.
That limitation has shaped the next version. Benita and her two teammates are redesigning Aquamanne to move beyond a hackathon demo by making the hardware smaller and more durable, reducing its dependence on electricity, supporting remote monitoring for farmers away from their ponds, and improving its language capabilities before taking it to market.
From a hackathon project to something bigger
Image Source: Ukachi Benita
Benita did not expect the prototype to attract much attention beyond the hackathon. But after she shared Aquamanne on X and LinkedIn, fish farmers began reaching out to her with problems that sounded familiar.
One of them was a farmer who lives in Lagos but has a fish farm in Owerri. Benita says the people managing his farm called him to report that all the fish had died and attributed the loss to the water. He wanted Aquamanne to help him keep an eye on the pond from a distance.
“He was asking me if it was possible for me to develop the tool so that he could be monitoring at least the water situation from anywhere he is,” Benita says.
The request made Benita reconsider what Aquamanne could become. What she had built for the hackathon was a simple demonstration that worked on her mother’s farm. Now, she was seeing a need for a system that other farmers could use beyond a single pond.
“That really got me thinking. So, I have actually started the process of public production. I am in the process with my team. We are trying to refine it because it is no longer a hackathon project. It is now something bigger for public use,” Benita says.
Farmers were not the only people responding to the project. After seeing her posts, two people with technical expertise reached out to help her develop it further. One is a Nigerian professor, while the other is an embedded systems engineer. Together with Benita, they now form a three-person team working on Aquamanne.
“The first person is a Nigerian professor that is in Oxford. He was just really interested in what I built. And the next person is an embedded system engineer. He is really good with hardware. So, we are in constant communication on how to properly develop it so that it will be better for the farmers,” she says.
Refining Aquamanne for the market
Now that Benita has a team backing her and has seen that the problem extends beyond the farmers in her mother’s circle, she and her teammates are refining Aquamanne into a product that can be used by farmers beyond the hackathon.
The team is working on the hardware, including making the device more compact and durable, while reducing its dependence on electricity. They are also considering the availability of the sensors they need and how easily a farmer can install the device.
“We don’t want something so big,” Benita says. “We don’t want something that the farmer cannot easily install.”
Cost is another consideration as they decide which hardware to use. Benita says the team wants the device to be reliable without making it too expensive for farmers.
“We want to use the best of hardware systems, but we don’t want it to be at the cost of the farmers,” she says.
The team is also refining the software. The hackathon version could interpret the water readings, but farmers could not yet ask the system questions. That feature will be included in the next version, while the team continues working to improve its performance across more African languages.
They are also thinking about how farmers will interact with the finished product. Benita expects many users will not be highly technical, so the team wants to make Aquamanne simple enough to use without requiring specialised knowledge.
“We are trying to make sure that there’s no technical barrier to the product,” she says.
Benita says they are looking to have a functional product by the end of the year. For her, the excitement is not just about turning a hackathon project into a commercial product, but about the possibility of helping fish farmers avoid the losses that first inspired her to build Aquamanne.
“I’m actually looking forward to how much relief we can give the farmers with our tool,” she says.
True scale demands moving beyond surface-level integrations to robust execution. We’ve filtered the noise out of Moonshot 2026, optimising the conference strictly for high-calibre connections between startup founders, global financial operators, enterprise leaders and individuals rewiring Africa’s technical frameworks. Get 20% off Early Bird tickets for a limited time.
