Y'all are building bots, right?


Bots have been a thing for months. Slack lit a proper fire under it with their bot fund. Then Facebook poured gasoline on it yesterday at F8. If they weren’t mainstream before yesterday, they are now. So, my people, which of you are building bots, and for what? Don’t worry, nobody is going to steal your bot idea.

I can already see them. Kongabot. Hotelbot. Ozombot…


lmaooo! … “Ozombot” sounds funny :smile:


This is a very valid concern since Nigerians always think others are out to steal my ideas. In this case, my people have no fear. You can take ‘my idea’ for free.

Any ways this idea is simple. Build a shopping assistant on messenger primarily targeted at Nigerians. Why would you win?

  1. If you’re platform agnostic - e.g you’re not tied to Konga (kongabot might be on it’s way) or Jumia etc, you can get your customers the best prices by sourcing the best deals on items. You can also proactively recommend items. e.g ‘Kemi, I thought you would like this shoe on sale on Fashpa…since you bought this on Konga’.

  2. Nigerians love shopping and FB boasts ~18million Nigerians. FB is also perfect for the age demographics which is young and upwardly mobile.

  3. Definitely choose FB over other platforms. Especially choose FB over Slack. People go on Slack to be productive (get work done) but go to FB to kill time (perfect for shopping). The added benefit is that since FB owns whatsapp, there’s likely hope of not needing to rebuild for it in the future.

  4. Finally (this is probably the most important point), tye contextual way the westerners shop can be dramatically different from the rest of the world. The key thing is machine learning (A.I) capabilities of bots depends a lot on the learning. A lot of assumptions are made on how we say, what we say, what we mean etc. that there’s absolutely no universal way of getting it close to right. See below screenshot as an example from Operator on how a customer will shop:

You can read more about Operator’s post below. Also think this might be of interest to @Yinka and @xolubi since we’ve had many a discussion about this in the past.


CNN’s messenger bot. Like Qz’s app, except that it is actually interactive. No natural language processing abilities detected though, so it’s still pretty dumb.


Yeah makes sense that it appears dumb as they’ve built it for a global audience. So everything about the way it interacts will actually be generic in nature.

That’s why if anyone in Nigeria has the serious chops to build for the local market, that’s an home run right there.


KongaBot and JumiaBot would be pretty lame, considering Facebook has made it clear that the Send/Receive API must not be used to send marketing or promotional messages, such as sale or product announcements, brand advertising, branded content, newsletters or the up-selling or cross-selling of products or service.

It seems not to have value for e-commerce at the moment.


KongaBot and JumiaBot would be pretty lame, considering Facebook has made it clear that the Send/Receive API must not be used to send marketing or promotional messages, such as sale or product announcements, brand advertising, branded content, newsletters or the up-selling or cross-selling of products or service.

It seems not to have value for e-commerce at the moment. What a restrcicted experience for your unique audience.


Hmmn, these are a lot of words. But are you sure about what you’re saying? From everything I’ve read, FB is offering the exact opposite of what they mean.

Can you point out where you got your info from?


Right on the Facebook Developers Page. Check the footer.


Facebook bots platform is based on an opt-in (at this time) model where the user must first engage the bot. There is nothing wrong in asking a KongaBot or JumiaBot to send you a list of items, you browse and make purchase without leaving messenger. I am pretty sure facebook would release a universal payment service for bots.


Yeah @MrASulaiman the way you’ve lifted that passage makes it seem like Facebook is against commerce especially as you said ‘KongaBot and JumiaBot would be very lame’. Whereas the whole point of F8 is to drive ecommerce…sell sell sell!

Of course Facebook wouldn’t allow any one to use their platform to spam their 700million users on messenger. That’s asking to be shot down like a bad rash!

See a good article from Mike Isaac which kind of shows how the likes of Jumia and Konga can use bots in a way that’s not lame.


The next big thing won’t have a website. Or an app. Oh, God I’m grateful to be alive in this era. To have actually experienced everything move from point A to B. In less than 15 years.


Hmmmm. So I can see brands will be offerred the ability to include richer content such as images, links and call-to-action buttons in the message thread.

It won’t be cheap, either. Facebook will eventually open up ‘Sponsored Messages’ and ‘Click to Message’ news feed ads that prompt conversation with bots.

I didn’t say it won’t be possible or it won’t happen, I used lame because I don’t see how the bot will add more value. Bots are indeed an exciting new addition to the Messenger platform that we should embrace.

However, just like the article you pointed out mentioned,

making each transaction into a personal conversation may not be what people need when they just want to book a flight or order food.

I don’t buy having a drawn-out conversation with a bot to make a purchase. If it were Voice like Siri, hell yeah. Until then, it’ll be LAME.

Using the Bot API for Sales and Product Campaign would be near reality, atleast to instigate discounts and new products on shelve. One can clearly block Sponsored Messages via the User Controls.


Hmm, how?

I think sponsored messages will come embedded as recommendation when a user request for a product or service.


An excerpt from the Messenger Product Overview

We’ve updated Messenger to provide people with even more ways to control the conversation and make sure they can reach the businesses they care about.

<img src="/uploads/default/original/2X/3/306bdbf445768570c9bfd04c8ec78c4c7a0cd7e6.png" width=“282” height="50


Well, very good that you’re now checking this out properly as it shows your initial conclusion (like below) wasn’t exactly correct.


Yes boss @PapaOlabode! Still it doesn’t have value at the moment. It probably will after the bots are improved further as promised. This is all beta yeah.

Patiently watching. XYZBot coming soon. LOL


“Stealing” this idea. :innocent:


@celestocalculus was championing Telegram thanks to bots, I know he has a few.


Just seeing this @PapaOlabode.

The idea is indeed simple. But it is also a little tricky. Here are a few things I’ve come to realize (causing me to slow things down) .

  • It is more successful in countries where there are a lot of established third party merchants in ecommerce. There are countless ecommerce companies in US, UK etc. Here in Nigeria, you don’t have as many good options.

  • Payments is also an issue. You can let the user pay themselves after they’ve decided on a product or you charge them first and make the purchase yourself. The latter means double processing fees which will eat into affiliate commission if you even have any.

  • With Nigerians getting so used to pay on delivery, you need a special partnership with the ecommerce companies if you want to be making purchases for customers.

  • When there are issues with order deliveries, Nigerians will be on your case. They don’t want to hear that your TOS indemnifies you.

  • The best monetisation strategy is affiliate commissions but only few ecommerce companies have affiliate platforms. So to offer more options to users, you need to partner with other ecommerce companies.

  • Unless it’s a bot, you need to start with a few chat assistants to give users short response times and unlimited internet for browsing products (This might not be an issue depending on where you are)

  • If you are using a bot, well you’re going to need to code plenty pidgin in dia cos Nigerians cn lyk 2 spoil English.
    Also depending on how the bot works, there can be hiccups. If it’s an API thing, well good luck with that in Nigeria. But if you’re using a scraper setup, it may or may not be easier.

It’s a very doable idea. Just not as easy to implement as it seems.