A Nigerian artificial intelligence startup, Decide, has ranked fourth globally for spreadsheet task accuracy, outperforming several well-funded competitors despite operating without external investment.
The ranking, released by SpreadsheetBench, a widely recognised benchmark for assessing AI performance on real-world spreadsheet problems, places Decide alongside some of the world’s most resourced AI agents.
Competing With Global Giants
Only three AI agents ranked higher than Decide: Nobie Agent, Shortcut.ai, and Qingqiu Agent. The comparison underscores Decide’s achievement, as some of its competitors operate with significant financial and human resources. Shortcut.ai, for instance, was developed by Fundamental Research Labs, which has raised $30 million, while Qingqiu Agent is backed by a company with over 5,000 employees and an estimated $5 billion market valuation.
Founded by former Flutterwave developer Abiodun Adetona, Decide launched publicly just a few months ago. The startup recorded 1,000 users within 24 days of launch and has since grown to over 3,000 users, including paying customers. The company currently operates with three employees and no venture funding.
How Decide Performed
According to Adetona, SpreadsheetBench evaluates how well AI agents handle practical spreadsheet workflows, including writing formulas, cleaning messy data, working across multiple sheets, and reasoning through complex Excel tasks.
Decide achieved an 82.5 percent accuracy score, successfully solving 330 out of 400 verified tasks in the benchmark.
SpreadsheetBench was developed by researchers from Tsinghua University and Renmin University of China and was introduced at NeurIPS 2024, one of the world’s leading artificial intelligence conferences.
“Rather than testing AI on artificial or simplified problems, SpreadsheetBench is built from real Excel questions sourced from online forums where users seek help with complex spreadsheet issues,” Adetona explained.
The “Verified” subset used in Decide’s evaluation consists of 400 carefully curated tasks, ensuring consistency and comparability across AI agents, and making the benchmark a trusted indicator of real workplace performance.
Built to Solve Real Problems
Decide was created out of frustration with the amount of time professionals spend cleaning data, debugging formulas, and navigating between spreadsheets. Unlike tools that merely suggest actions, Decide analyses the structure of a spreadsheet, executes changes directly, and explains its actions in plain language.
A Signal for African AI
Beyond product performance, Decide’s global ranking highlights a broader shift in Africa’s AI ecosystem. According to Techpoint Africa’s list of notable African AI products of 2025, innovation on the continent has been driven more by practical problem-solving than by scale or funding.
However, Decide’s success also raises a larger question: can African AI startups compete at global scale without significant capital? While limited funding remains a challenge, Decide’s performance suggests that technical focus, clarity of purpose, and execution can still deliver world-class results—at least in the early stages.
As African AI founders continue to build with constrained resources, Decide’s rise offers a compelling example of how global impact can emerge from local innovation.
