Deploying an artificial intelligence (AI) voice agent across emerging markets should not feel like a punishment. Yet, for global technology giants and local developers alike, hitting the wall of fragmented regulatory frameworks, physical office requirements, and convoluted KYC processes kills innovation before it reaches the market. Joshua Barimon Firima, founder and chief executive officer of KrosAI, describes this reality with chilling accuracy: “It’s like walking on glass, wearing concrete shoes. Any small thing, it breaks, and you’re gone.”
The global technology ecosystem is rapidly entering the agentic era, where AI is moving from text-based chatbots to autonomous voice agents handling customer support, outreach, and operations. However, the adoption curve in emerging markets is continuously choked by a lack of deployment infrastructure.
“When global companies look at this market, they see 54 fragmented markets they have to get licences in,” Firima explains. “They have to get regular approvals, they have to set up a team, so in the long run, they tend to just focus on already established markets.”
Firima experienced this bottleneck firsthand while trying to test a previous product. “It took us like four months before we could get a provider that supports our use case,” he says. “Then another like one month to do the integration work. That’s like five productive months gone just simply because we want to deploy AI agents.”

KrosAI was born from this frustration. Relaunched in January 2026, the platform provides a single API that handles both technical and regulatory complexities in just 15 minutes. Users complete KYC once and gain the ability to deploy AI agents on local numbers across multiple regions without establishing a physical corporate presence in each one.
Building the Stripe for telecom
To understand KrosAI’s business model is to look at the evolution of financial technology a decade ago. Before banking-as-a-service providers and unified payment APIs, cross-border transactions in Africa were a logistical nightmare. Stripe and its local equivalents changed that by abstracting the complexity. Firima positions its platform as a merchant of record, effectively building the ‘Stripe for telecom’.
“If you register on Stripe, they give you different payment methods. That’s more like a merchant of record,” Firima notes. “We do so on our own end. You register on our platform on KrosAI, you KYC once, and you can deploy and interact with customers across different markets.”
This infrastructure creates massive customer retention. “Once you find one that just works, it’s extremely hard to leave because you’ve already embedded it into your day-to-day workflow, and pulling it out causes more damage than replacing it,” Firima says.
Despite being less than a year old, KrosAI has processed hundreds of thousands of minutes of AI-powered phone calls. The platform boasts over 750 registered business users, with more than 60 actively paying for the service, a solid 10% conversion rate that highlights deep market need. Their enterprise client roster already includes notable names such as the Kenyan social enterprise Ushahidi, AethexAI, and the United Nations-backed Clean Cooking Alliance.
KrosAI is engineering for the agentic era
Firima’s background as a two-time founder heavily influences KrosAI’s developer-first approach. His previous venture, DFYHero, automated website building for local businesses and taught him the strong business value of compressing technical barriers into easy-to-use tooling.
This philosophy is evident in KrosAI’s aggressive product shipping cycle. The startup is rolling out features specifically designed for autonomous AI, including an ‘adaptive phone book’. Much like a human retains a contact list regardless of their network provider, this feature allows an AI agent to dynamically recognise and pull contact details across different providers and numbers, solving a critical identity problem for enterprise users.
“We are rebuilding it from the ground up for AI agents,” Firima states.
Furthermore, recognising that more than half of the code on the internet is now AI-generated, KrosAI has shipped a Model Context Protocol (MCP) server. This allows coding agents to seamlessly interact with KrosAI’s APIs, meeting developers directly within their automated workflows.
“If you want to build a durable company, it doesn’t always have to be like building for clout or building for investors,” Firima says. “We follow our customers, and our customers dictate our road map.”
KrosAI’s trajectory hasn’t been without its scars. Late last year, Firima experienced a severe co-founder fallout that forced a hard reset. He had to rebuild the company, the product, and the customer base entirely from scratch.
“That has been one of the darkest periods in my life where I had to face my fears,” Firima recalls. “But that period really tested me, and I’m glad I didn’t fold, and I was able to come out of that and just kept building, which is what I know how to do the most.”

That resilience is now powering an ambitious expansion strategy. Currently live in Brazil, South Africa, Ghana, Kenya, Nigeria, the US, and the UK, KrosAI is targeting 30 to 35 markets within the next 12 months.
Backed by about $135,000 in early funding from Right Side Capital via MARL Accelerator and the INSEAD AI Venture Lab Program Abu Dhabi.
Firima describes their current financial standing carefully: “We are not literally bootstrapping. We call it Seed-strapping because we have raised a little bit of money.” He is preparing to initiate a pre-seed fundraising round in the coming weeks.
His ultimate five-year vision is to secure direct telecom licences across key regions. “Our goal is to make it possible that from day one, you could deploy voice agents across 100 different markets, spanning Africa, the GCC, Latin America, and Southeast Asia,” Firima outlines.
For a founder who started his first business at 16 and has never drawn a traditional salary, surviving the rigorous regulatory climate of Nigeria was the ultimate training ground. If KrosAI can successfully build the rails for AI voice agents in the Global South, they will not just be participating in the agentic era; they will be the infrastructure that powers it.