Natural Language Queries: Ask Inventra Anything in English or Pidgin

The Interface Problem

Most enterprise software assumes its users want to interact with tables, filters, and dropdowns. For a data analyst or accountant, that works. For the owner of a 12-branch distribution business who is on the road visiting clients and needs to know if their Ikeja warehouse is about to run out of a key SKU — it does not work. They need to just ask.

Natural Language, Built for Nigeria

Inventra's NLP query engine — launched in February 2026 — allows any user to type or speak a question in plain English or Nigerian Pidgin and get an instant, accurate answer from their inventory data. The engine is trained to understand the way Nigerian business owners actually speak about their operations:

• →"Wetin dey remain for Apapa store?" → shows current stock levels at Apapa location
• →"Which item dey run low?" → surfaces all items below reorder threshold
• →"How much we sell last week for Abuja?" → returns revenue breakdown for Abuja branch
• →"Which supplier give us wahala this month?" → flags suppliers with delayed deliveries

Language is the most natural interface. When business owners stop thinking about how to use the tool and start just talking to it, everything changes.

How We Built Pidgin Support

Nigerian Pidgin is not a dialect of English — it has its own grammatical structures, vocabulary, and regional variations between Lagos, Port Harcourt, and Benin City. We partnered with a team of Nigerian linguists to build a Pidgin corpus specific to business and commerce contexts, then fine-tuned a base language model on this data alongside anonymized query patterns from our beta users. The result is a model that handles code-switching between Pidgin and English naturally, the way real users actually type.

What's Coming Next

Hausa language support is in development for Q2 2026, targeting our expanding user base in Kano, Kaduna, and Abuja. We are also building voice query support for hands-free use during warehouse walks — say your question out loud while doing a shelf audit, get the answer through your earpiece. If there are other languages or query patterns you want to see supported, write to product@inventra.ai.