I agree with your points especially the 2 above. This is why I’ve said that machine learning needs learning of the right dataset to be effective. Which invariably means there’s an inherent advantage for a Nigerian founder.
Talking about the logistics issues and challenges, it’s clear that this is where the real work will happen. But I think it’s actually more of a business model question before working out the logistics of implementing it. For instance, let’s assume the bot technology was built (which caters to the unique lingo and language attributes of Nigerians). Then the following could be likely approaches;
Man-in-the- middle/affiliate: Where the bot sources the best deals and the purchase is directly between customer and ecommerce store. Eg Kemi wants a bag and not checks 10 stores to find her Gucci in Jumia. Kemi is happy she’s gotten a good bargain, pays Jumia and they ship. Jumia pays you commission.
Own the whole relationship: In this case, Kemi pays you. You buy from Jumia and ship to Kemi. It’s assumed that you will do this as able to make more money, than just affiliate fees. Refunds and returns has to go through you.
New Konga/Jumia: If you flip this around and will say you have the technology and enough dataset to know what Nigerians are buying. Then perhaps it makes sense to purchase the Gucci bags, and when Kemi needs you sell directly to her.
And @Yinka no one said it will be easy to implement…