• Post author:
  • Post category:AI World
  • Post last modified:February 14, 2026
  • Reading time:4 mins read

Inside Decide: the 3-person Nigerian AI ranked 4th for spreadsheets

What Changed and Why It Matters

A three-person Nigerian startup, Decide, ranked fourth globally on a public leaderboard measuring AI agent accuracy for spreadsheet tasks. Multiple outlets across tech media and social channels reported the result.

Why it matters: spreadsheet work is one of the most universal workflows in business. Accuracy isn’t a nice-to-have—it’s the whole game. A lean team hitting top-tier results signals a shift: in agentic AI, precision and tight scoping can outperform size and spend.

Accuracy is the real moat for AI agents. Distribution amplifies it. Scale comes last.

The Actual Move

Here’s what happened, pulled across all reports we reviewed:

  • Decide placed fourth worldwide for spreadsheet task accuracy on a public agent leaderboard focused on spreadsheet operations and task completion.
  • Coverage highlights that the team is just three people and, per one report, has taken zero outside funding.
  • The ranking places Decide alongside much larger, well-funded AI groups, and the result was widely shared across African tech channels (media, LinkedIn, Facebook, Instagram, and X).

A small Nigerian team beat most of the field on a global spreadsheet benchmark—without traditional venture funding.

The Why Behind the Move

Zoom out and the pattern becomes obvious: agentic AI is moving from demos to dependable task execution. Spreadsheets are the perfect proving ground—structured domains, clear success metrics, and immediate business value.

• Model

  • The likely approach: a tightly-scoped agent that plans, verifies, and executes spreadsheet actions with strict tool use.

  • Constrained domains reduce hallucinations and let evaluation be unambiguous (did the cell change correctly?).

• Traction

  • A top-5 global ranking is a credible signal of product readiness, not just model flash.

  • Broad social amplification suggests strong founder-market fit and early community pull.

• Valuation / Funding

  • Reports point to zero external funding. That’s capital-efficient learning and high leverage per dollar.

  • It keeps optionality open for timing a raise around real adoption.

• Distribution

  • Public leaderboards are underrated go-to-market. They compress proof, narrative, and reach into one artifact.

  • Next natural channels: plug-ins for Excel and Google Sheets, and bottoms-up teams that live in CSVs.

• Partnerships & Ecosystem Fit

  • Spreadsheet ecosystems are mature. Integrations with BI tools, accounting stacks, and workflow automation platforms can compound quickly.

  • Accuracy-first agents slot cleanly into finance, ops, analytics, and revops.

• Timing

  • 2026 is the year reliability becomes the differentiator. Benchmarked accuracy travels faster than model size as a buying trigger.

  • Enterprises want agents that do one painful job perfectly, not ten jobs poorly.

• Competitive Dynamics

  • Incumbents can bundle agents into office suites. But bundling often lags on depth and accuracy for niche tasks.

  • Specialist agents can win with speed, quality, and trust—if they stay focused.

• Strategic Risks

  • Benchmark overfitting: great scores may not generalize to messy, real data.

  • Integration friction: adoption dies if setup, permissions, or data privacy are unclear.

  • Sustainably defending accuracy as bigger players target the niche.

What Builders Should Notice

  • Pick a painful, universal workflow—and own it with ruthless focus.
  • Benchmarks can be distribution. Public proof compresses sales cycles.
  • Accuracy is the wedge. Trust compounds faster than model size.
  • Fundraising is optional when the product tells the story.
  • Integrations are strategy. Meet users where they already work.

Buildloop reflection

The winners won’t ship the most—they’ll ship the most correct.

Sources

Tags: , , , , , , ,