Bookkeeping AI for small shop owners in Africa.

Designing and Building for how people behave, not how we want them to behave.

Some Back Story

Micro, small and medium-sized businesses are presumed to make up 90% or more of all businesses in sub-Saharan Africa. This comprises local estate shops, kiosks, small hotels “kibanda-ski” and various other small corner store businesses. They are the backbone of most African economies.

The one thing they all have in common is that even with the major advances in technology, bookkeeping is still done manually using pen and paper.

Back in 2020 while a friend and I were attempting to build a simple record keeping mobile app for these businesses, I decided to do a bit of research first. I visited about 15 shops in my area and inquired as to how they did record keeping, stock keeping and profit/loss tracking, the answer I got was mostly the same.

For record keeping, they all had a book and everything was done manually, mostly. One or two had cracked versions of QuickBooks but they would still record on a book and transfer to the software later after closing shop.

For stock keeping and profit/loss tracking, they would all do a count of whatever was in the store in the morning and subtract that from whatever they had the previous morning and that would be what they would have sold the previous day. It’s really quite hectic and there are a lot of loopholes in this method as far as theft is concerned, but it worked for them.

After talking to these business owners, I realized that an app was going to be a hard sell. Typing on a phone would be cumbersome and it would be a huge shift from what they were already used to doing on a daily basis, writing. We would have to offer immense value for which we had none, if we were ever going to convince them to make a shift. Honestly, writing was just much easier and they had been doing it since childhood.

We wanted access to the sales data and insight as to how products flow into and out of these shops but what were we going to give them in return? A learning curve on an app that they would abandon after a week or two. Granted this data would be gold if we could get it and a number of startups in East and West Africa have tried or are trying to get this data through various means but mostly using a mobile app. Some such as Sabi Cash, Bumpa, Kippa, OZE, Leja, RejaReja, Patika and most recently Bamba have all had a go at innovating for this sector of the market, as to whether they have been successful or not, yet remains unknown. There are also a host of other independently built apps in the play store that help with inventory and store management for small businesses

The one thing that remains true is that to this day, these businesses still use pen and paper for record keeping. Is there a way to digitize these records? Yes. How? By designing for how these shop owners already work on a daily basis, building for how they behave, not how we would want them to behave.

Below are two case studies.

1. Using AI to scan the written down records:

AI has been getting better at accomplishing many tasks over the years and one is scanning handwriting. There are currently a number of AI tools in the market that can convert handwriting to text/a digital document. This technology can be leveraged to help with the digitizing of handwritten records for small shop owners in sub-Saharan Africa.

One can look at how different shop owners keep their records in the books that they use and train the AI using this data to recognize the different bookkeeping patterns more so in terms of how tables are structured so that once a scan is done the app then organizes this records properly and comes up with the required arithmetic and insights for the shop owner.

If the methods used by the shop owners are too different, say one uses tables, another just writes down records in no defined manner, one uses dates, another doesn’t etc. A defined way of record keeping can be agreed upon and the shop owners urged to use this method when writing down records so that the AI can easily convert them to digital records. One can even go as far as providing books that have been printed on already and will make both record keeping, scanning and conversion easier.

Having digital records can really help with providing proper analytics and predictions for these small shop owners and I’m sure even they would welcome a digital solution. They are already willing to try.

The methods that are currently available (mobile apps) can be quite cumbersome to use, especially for repetitive data entry tasks such as record keeping. Writing on the other hand is second nature for most humans, technology that works with this in mind could be a winner for all parties involved.

A value addition service would be to use the data collected digitally to give insight to the shop owners on the products that consumers buy most in their area from the different shops that use their AI infused app. How many chapatis are sold in that area per day, soda, mandazi, maize flour? Which brands sell the most? At what times and days do people buy the most? How much do barber shops make in a month in that area? To mention a few.

This is data that can also help those who are looking to start a small business of their own. It could help provide insight as to which small business to venture into in a particular area

If you already have a record/bookkeeping app for small businesses, you could also look into using AI that can complement the day to day habits of these small shop owners.

A custom note taking tablet for small shop owners in Africa:

This one is a bit of stretch especially because hardware can be an expensive and time consuming endeavor but, hear me out.

Sometime last year I came across a device called remarkable, touted as the only tablet on the market that feels like paper. They have engineered this tablet so that writing on it feels exactly like writing on a piece of paper. It is designed specifically to make note taking on an electronic device, feel as it would while using a pen to write on paper.

I checked out their website, watched a few product reviews on YouTube and I was sold. In terms of design, the product looks simple, minimal and functional, just like a good notebook would. It’s also pretty sleek.

It was definitely not designed for bookkeeping at a local Kenyan corner store but the functionality built into it would serve this demographic quite well. Yes, it even converts handwritten notes to typed text, battery lasts about 2 weeks, works like a book thus no distraction from WhatsApp notifications, offers proper organization of work, unlimited pages among many other great features.

All that data can be saved to a cloud and used later for analysis and insight building.

A tablet for note taking and bookkeeping, designed for small shop owners in Africa. That would be something, in the meantime, we keep building and innovating.