While most artificial intelligence tools stop at providing information, an 18-year-old Nigerian student has built something different a platform that actually completes tasks on websites instead of just explaining how to do them.
Obinna Chimdi, who studies mathematics and computer science at the University of Port Harcourt (UNIPORT), spent six months developing ChatATP after struggling to find anyone willing to collaborate on his vision. His platform bridges the gap between conversational AI and real-world action.
“If you tell ChatGPT to book a flight for you, it will respond by saying it does not have the capacity to do that,” Chimdi explained. But ChatATP works differently—users simply type their request, and the system executes the task directly on the website.
The young developer’s journey started at 16 when he began teaching himself programming using his father’s mobile phone, completing his first project within three months. His inspiration? Mark Zuckerberg’s early success story.
“He was my idol,” Chimdi said. “I loved automations too. I wanted to have my own AI company.”
That drive led him to create Wall Street, a business-focused social networking platform. But ChatATP represents something far more ambitious—an attempt to solve what he sees as a fundamental limitation in current AI tools.
Chimdi’s frustration grew from watching powerful language models stop halfway through tasks: “If you ask them to analyse data and send the report to someone by email, they stop halfway. You still need to copy, paste, or do the rest yourself”.
His solution involves toolkits—small connectors that link websites with language models like ChatGPT, Gemini, or Claude. These toolkits translate between human intent and website functions, whether that’s checking flight availability, sending emails, or pulling university timetables.
The breakthrough came with his Agents2 protocol, which Chimdi describes as “HTTP for AI agents”—a universal standard that defines how AI models and toolkits communicate, much like HTTP standardized how browsers fetch web pages.
But turning technical innovation into market success presents steep challenges. So far, ChatATP has attracted just three users: his neighbour and two classmates. Expanding would require compute resources he can’t afford, and raising funding with such limited traction seems unlikely.
The competition looks daunting. Open-source projects like AutoGPT and LangGraph already enable developers to build autonomous agents and execute complex workflows, while newer entrants such as Lumio AI offer multi-model workspaces where users can switch between different AI systems. These competitors boast strong developer communities and significant financial backing.
ChatATP requires technical knowledge to use—developers must configure API keys from major language model providers, install domain-specific toolkits, and select which model powers their agent. That complexity might limit adoption compared to more user-friendly alternatives.
Still, Chimdi believes his Agents2 protocol offers something distinctive. “I wanted to build something that shows AI can go beyond answering questions,” he said. “AI should be able to connect to the real world and actually do things.”
Whether ChatATP evolves beyond proof-of-concept into a viable product depends on Chimdi’s ability to attract users and resources in a field where innovation moves quickly and funding matters enormously.
Source: newsghana.com.gh
