Company · The Akoni Global Ltd
Built in Oyo State, for stores across Nigeria.
The Akoni Global Ltd is a company registered in Nigeria, building Akoni Loop: AI agents that keep online-store listings current with the market, in the language buyers use.
Why Akoni exists.
Most online stores write a listing once, when the product goes up, and rarely touch it again. The market around it moves every day: what buyers search for, what competitors charge, what reviews say, what is in stock.
Large sellers pay people to keep up. Small stores cannot. Akoni Loop does that work for them, in the language their buyers use, and shows its working for every change.
Rewritten by hand, once
At listing time, and rarely again.
The market moves
Search demand, prices, reviews, stock and season.
Rewritten on every signal
Versioned, measured, and reversible in one click.
At a glance.
Done so far.
From registration to a specified product and a team building it.
Company incorporated
The Akoni Global Ltd registered with the Corporate Affairs Commission, Nigeria.
Product and architecture specified
Five-signal loop, decision log, guardrails and measurement defined; five-stage system architected on AWS; GPU path mapped to NVIDIA tools.
Clients, team and first models
SEO growth work with online sellers; open-weight models fine-tuned on NVIDIA H100 GPUs; founder and CTO building; akonilabs.com live.
Team.
Small on purpose: a founder who writes code, and a CTO who owns the system.
Adeyinka AbdulAzeez Babatunde
Founder · The Akoni GlobalOyo State, Nigeria
Leads the company and the product, and writes code. Started Akoni so small stores get the kind of listing optimisation only large sellers could afford.
Firas Lilouke
CTO · The Akoni GlobalOwns the architecture: the agents, the store connectors and the rewrite models.
Work with us.
The team stays small until early access opens. These are the next people we expect to need.
- Build and maintain Shopify and WooCommerce connectors with versioned publish and rollback.
- Own evaluation: click-through per version, assistant citations, native review results.
- Production Python or TypeScript, and hands-on work with language models.
- Review Yoruba, Hausa and Igbo product copy before it is first published.
- Flag the words buyers actually search with, and the ones they never would.
- Your corrections become training data for the market-tuned models.
Contact.
One inbox, read by a person.