Decide, a burgeoning AI startup based in Nigeria, has garnered global recognition by being ranked as the fourth most accurate AI agent for spreadsheet tasks. This significant achievement was determined by SpreadsheetBench, a highly respected benchmark designed to evaluate AI performance across practical, real-world spreadsheet challenges. The ranking places Decide alongside well-established and well-funded global AI enterprises, a remarkable feat considering the company only launched publicly a few months ago.

Founded by Abiodun Adetona, a former developer at Flutterwave, Decide has demonstrated impressive growth since its inception. Within just 24 days of its public launch, the platform attracted 1,000 users. Adetona has since confirmed that the user base has expanded to over 3,000, including a segment of paid subscribers. Despite operating with a lean team of only three employees and having secured no external funding, Decide has outperformed a vast majority of AI agents globally, with only three—Nobie Agent, Shortcut.ai, and Qingqiu Agent—ranking higher. This is particularly noteworthy when considering that Shortcut.ai was developed by Fundamental Research Labs, a company that has raised $30 million in funding, and Qingqiu Agent boasts over 5,000 employees and an estimated market capitalization of $5 billion.

The integrity of Decide's ranking is rooted in the rigorous methodology of SpreadsheetBench. Developed by researchers from Tsinghua University and Renmin University of China, this benchmark was unveiled at NeurIPS 2024, one of the world's premier artificial intelligence conferences. As Adetona explained, SpreadsheetBench distinguishes itself by evaluating AI agents on actual Excel problems sourced from online forums, where users seek assistance with complex spreadsheet issues, rather than relying on artificial or simplified tests. The 'Verified' subset, comprising 400 meticulously curated tasks used for Decide’s evaluation, ensures consistent and comparable results across various agents, making it a reliable indicator of an AI system's efficacy in real-world professional environments.

Decide achieved an impressive 82.5% accuracy score on SpreadsheetBench, successfully resolving 330 out of the 400 verified tasks. This high accuracy demonstrates its proficiency in handling a diverse range of practical spreadsheet operations, including writing complex formulas, cleaning messy datasets, working seamlessly across multiple sheets, and navigating intricate Excel workflows with precision.

The startup itself was born out of a shared frustration with the substantial amount of time professionals dedicate to manual data cleaning, debugging formulas, and navigating between spreadsheet tabs. Decide addresses these pain points by operating on a principle of deep understanding: it comprehends the inherent structure of a spreadsheet, directly executes necessary changes, and lucidly explains its actions in plain language, moving beyond mere suggestions to active problem-solving.

Beyond its immediate product utility, Decide's global ranking significantly contributes to a broader discourse surrounding AI development in Africa. Techpoint Africa's assessment of notable African AI products in 2025 suggests that the continent's AI landscape has been characterized by an emphasis on solving immediate, practical problems with limited resources, rather than prioritizing large-scale expansion. Decide's exceptional performance, however, prompts a critical question: should African AI startups begin to conceptualize and build for a global scale? This aspiration, while promising, is often tempered by the significant capital requirements necessary for achieving global reach, a factor that could potentially impede many talented African founders from realizing their full potential.