Nigerian startup Decide launches DAX-1, its first in-house AI model for spreadsheet editing

Nigerian startup Decide has taken its first step toward building its own AI. The company, which makes AI tools for spreadsheet automation, has introduced DAX-1, a specialised model designed to edit spreadsheets faster and more cheaply than the larger models it previously relied on.

According to Techpoint Africa, Decide had been using a mix of models from OpenAI, Anthropic and Google to power different parts of its product. DAX-1 handles tasks such as fixing formulas, cleaning duplicates, correcting spreadsheet errors and formatting workbooks.

Instead of training a model from scratch, Decide started with an existing open-source model from Qwen, Alibaba's AI research arm, and trained it specifically for spreadsheet editing. "Our goal is to be a critical infrastructure for spreadsheet automation operations in general on the internet," co-founder Abiodun Adetona said.

Adetona argues that general-purpose models fall short on cost. "For some repetitive tasks that require doing them over and over again, they can get very expensive quickly." The company is not trying to build the next GPT or Claude — it is making a much narrower bet.

According to Decide, its base model achieved 9.17% accuracy on the company's benchmark before additional training; DAX-1 reached 91.7% after the work was completed. The production version is designed for deterministic spreadsheet editing, producing executable edits rather than suggestions about what a user should do.

Internal testing found DAX-1 performed better than several larger models on its specialised tasks while being significantly cheaper to run — five times faster than Anthropic's Haiku on the task it was designed for and 97% cheaper. Decide also tested against models from OpenAI, Google and DeepSeek, with larger models performing strongly but at a higher cost for this particular type of work.

The company acknowledges the benchmark was conducted internally and has published its methodology and evaluation approach so others can verify the results. Earlier, its agent scored 82.5% on SpreadsheetBench Verified, placing it fourth on the leaderboard at the time of its evaluation.

Adetona compares the economics to the Jevons paradox, arguing that lower costs could make spreadsheet automation easier to use at scale and create more demand rather than reducing what the company earns. DAX-1 does not mean Decide is abandoning external providers — it will continue to use other models where they make sense.