It’s not exactly conventional wisdom to pivot from technology that literally saves lives to crypto, then to technology that handles customer complaints, in between publicly clashing with one of Nigeria’s prominent fintech firms over who the rightful owner of a certain brand name is. But for Moore Dagogo-Hart and his team at Cognito Systems, it is the logical next step.
Dagogo-Hart, 26, is a Nigerian tech entrepreneur and poet on the side, perhaps best-known as the Co-Founder and CTO of Zap Africa, a non-custodial crypto exchange at the centre of the aforementioned name dispute that kicked up a fuss earlier this year.
His latest venture, as Founder and CEO of deep-tech company Cognito Systems, is a bet that the same artificial intelligence (AI) principles used to detect emotional distress in suicide prevention helplines can be used to improve customer service in Africa.
The new platform is called Martha AI. It arrives as customer service in the region faces a crisis. Nigeria’s 2024 Customer Satisfaction Index sits at 67%, and South Africa’s telecoms sentiment score is a dismal –25%. Dagogo-Hart believes the problem is not a lack of technology, but a lack of understanding.
“The through-line has always been human context,” he tells WT. “Every project I’ve led has been about bridging the gap between how people think and how digital systems respond.”
First suicide hotlines, now support desks
The foundation for Martha AI was laid years ago at Syx Labs, the team’s previous company. There, they built systems for mental health helplines in Canada. That technology could pick up signals of fear, isolation, or hopelessness in text messages and alert human responders before a situation turned fatal.
That same core technology is now being retooled, this time, amid a global AI boom. Dagogo-Hart argues the challenge is similar. “Someone in crisis might say ‘I’m tired’ when they mean ‘I’m in danger.’ A frustrated customer might say ‘this is fine’ when they mean ‘I’m about to cancel,’” he explains.
The early evidence comes from their own operations. They tested Martha AI on thousands of customer support conversations from their crypto platform, Zap Africa. The result was a dramatic reduction in average response time, from 30 hours to under 30 seconds.
“We are not selling empathy as a feature,” Dagogo-Hart states. “We are selling resolution as an outcome.”
AI with Empathy
In a continent with thousands of languages and dialects, Dagogo-Hart argues that the key to understanding is less vocabulary and more rhythm, code-switching, and cultural nuance. Global customer service tools from companies like Zendesk or Intercom are often built for English-first markets and stable internet connections.
Martha AI takes a different approach. It uses OpenAI’s powerful models as a base but fine-tunes them on a diet of African conversations. It learns from how people actually talk on platforms like WhatsApp, where messages blend English, Pidgin, and local languages in a single breath.
“Global tools assume stable infrastructure,” says Dagogo-Hart. “African customer support happens on WhatsApp, over 2G connections, often in voice notes. Martha is built for that reality.”
His commercial thesis is that cheaper, basic bots already exist and often make problems worse by frustrating customers and forcing escalations. “Our commercial thesis isn’t ’empathy is nice to have.’ It’s ’empathy reduces cost-to-serve because it deflects tickets that basic bots escalate,’” he says.
For him, the business case for Martha AI is not about selling a feeling, but a concrete financial result. The return on investment, he says, comes from fewer repeated questions, shorter call times, and higher customer retention.
The platform operates on a subscription model, with tiers starting at around USD 20.00 a month. The pricing is a direct challenge to the high cost of traditional support. Dagogo-Hart highlights that hiring just three human agents in Nigeria can cost a company roughly NGN 600 K monthly.
“Even our premium package doesn’t come close to that cost,” he says, positioning Martha AI as a tool that can drastically cut support overhead while maintaining service quality.
He is also clear about his target audience and the competitive landscape. His focus is on African small and mid-sized businesses who are overlooked by global giants. “We’re not competing to be the global CX platform,” he explains.
While he acknowledges that companies like OpenAI or Zendesk could theoretically enter the market, he believes his window of opportunity is their inability to adapt. They would have to completely rebuild for African communication patterns and price for local budgets, a gap his team is built to fill.
Balancing Two Worlds and Learning from Stumbles
The company is currently powering two major brands: the established Zap Africa in crypto and the nascent Martha AI. This is not a new situation for the team. Zap Africa itself began as a product within Syx Labs before spinning out.
“They’re both meant to coexist long term, because they’re smaller parts of a bigger system I’m building,” Dagogo-Hart explains. The ecosystem is designed to be interconnected. Zap Africa provides Martha AI with a real-world testing ground and payment infrastructure.
But the journey has not been without public learning experiences. The founders were previously involved in a trademark dispute with the fintech giant Paystack over the name “Zap.” It was a headline-grabbing episode that highlighted the growing pains in Nigeria’s tech scene.
Dagogo-Hart acknowledges the lesson. “That chapter was a learning curve for us, and I think for the wider ecosystem too,” he says. “It showed how important governance and clear IP frameworks are.”
He notes that Zap Africa has since built more robust legal foundations, including securing an EU VASP license for its crypto operations. For Martha AI, they are aligning with Nigerian data protection laws from the start.
The race is on
The financial runway for this ambitious AI pivot is a mix of bootstrap and angel investment. Dagogo-Hart says Cognito Systems has generated about USD 120 K in revenue and has one angel investor who has committed USD 50 K. They are in talks for more funding.
The team operates like a “Black dev-ops unit,” a term Dagogo-Hart says reflects their focus on efficiency and precision. “We strive to deliver the best results with the least input, building systems that are compact yet powerful,” he explains.
So, what will define success or failure for this unusual bet? For Dagogo-Hart, two metrics matter most in the next 12 to 18 months. “ARR tells me how fast what we’re doing is growing,” he says. “And sentiment accuracy is also key. The whole point of building Martha AI is to solve real problems.”
He is clear on when he would consider pivoting. It would only be if someone else built a better solution for the African market. “I believe it’s a problem that truly needs solving,” he says.
For now, the team that once built technology to listen for despair in Canada is now betting that teaching machines to listen for frustration in Lagos is not just a business, but the next chapter in their mission to bridge the gap between people and the digital systems that serve them.


