Nigerian startup Decide has ranked fourth globally for accuracy in spreadsheet-focused artificial intelligence tools, according to results from SpreadsheetBench, a widely used benchmark for evaluating AI performance on real-world spreadsheet tasks.
The ranking places Decide alongside major global players, including Microsoft, OpenAI and Anthropic.
SpreadsheetBench is designed to test how well large language model-powered agents handle complex spreadsheet environments. Unlike simpler table-based benchmarks, it evaluates tasks involving formulas, multi-sheet workflows, spatial layouts and multi-step reasoning.
According to founder Abiodun Adetona, a former Flutterwave developer, Decide achieved an 82.5% accuracy rate, solving 330 of 400 verified tasks. Only three AI agents — Nobie Agent, Shortcut.ai and Qingqiu Agent — ranked higher in the evaluation.
SpreadsheetBench measures practical spreadsheet capabilities, including writing formulas, cleaning messy datasets, reasoning across sheets and handling complex Excel workflows.
Adetona said the result reflects sustained investment in applied research and product development grounded in real-world business data use cases.
“This milestone highlights how focused engineering and domain-specific AI development can deliver frontier-level performance outside large research organizations,” he said. “By concentrating on practical business data problems and building systems grounded in real user environments, smaller teams can meaningfully advance applied AI.”
Founded to address the time professionals spend cleaning data and debugging formulas, Decide has grown to more than 3,000 users, including paying customers. Adetona said the user base signals the startup’s potential to scale further.
He described the ranking as a foundation for continued development in intelligent spreadsheet automation and analytics, as African startups increasingly compete on the global AI stage.
