Key ideas and takeaways from a recent TED Talk by Andrew Ng on practical use of AI for small business. Andrew shows how machine learning systems can be used for businesses growth and development in small and midsized companies.
AI skills are as essential as literacy
Andrew starts with a powerful mental metaphor, comparing contemporary state of AI systems to that of literacy back in the day.
When I think about the rise of AI, I’m reminded by the rise of literacy. A few hundred years ago, many people in society thought that maybe not everyone needed to be able to read and write. Fortunately, it was since figured out that we can build a much richer society if lots of people can read and write.
The idea is that knowing how to use and build an AI system is an essential skill. Akin to the ability to use internet today. It is possible to live without this skill but it sure is much better to have it mastered.
Andrew also points out how today AI is primarily utilised by large tech corporations and hence AI systems are tuned specifically for their needs. The reason behind it is that building and honing AI systems is expensive and thus smaller business usually don’t have the financial justification to bare the costs of having an AI team.
AI systems are good at spotting patterns when given access to the right data
AI does not need much data
One interesting revelation is about data. In his Pizza shop example he shows that even modest amount of sales data generated by a single Pizza shop is enough for an AI system to generate helpful advice. Advice which can yield the business additional couple of thousand dollars a year. His message essentially is – by using help from AI any business can generate extra revenue to minimum cost.
Having more data does help, but contrary to the hype, AI can often work just fine even on modest amounts of data
AI ideas for small business
Typical auto mechanic, small retailer or school don’t use AI in their line work today, however they very well may in the future. Below are some business application ideas from Andrew himself, for different types of small businesses
Demand forecasting
Demand forecasting for a T-shirt company, based on a popularity of memes on the internet. This information can be used as a print to potentially drive sales.
Product placement
Front-of-store manager can take pictures of what the store looks like and AI can recommend where to place products. The change in placement can improve sales.
Supply chain
AI can recommend a buyer whether they should pay certain price or should keep looking for a cheaper option.
Quality control
Automated and semi-automated inspection for offline goods, where AI acts as an inspector identifying defects. A baker for instance can check for the quality of the cakes they’re making. An organic farmer can use AI in the baking business to check the quality of the vegetables. Furniture maker can use AI for business to check the quality of the wood they’re using.
Long tail of AI for business
According to professor Ng AI is not that widely used outside the internet and tech sectors in other industries. Surprisingly even in large companies like pharmaceuticals, car manufacturing and hospitals. This is the so-called long tail problem of AI.
Allegedly, the single most valuable AI system right now is the one that decides what Ads to show people on the internet. Followed by a web search engine and a shopping recommendation system. ¯\_(ツ)_/¯
No code AI platforms for business
There are now emerging new AI development platforms that shift focus from writing code to having the users provide data. The latter of course is much easier for a lot of people to do.
These are essentially supervised learning platforms. A tool called LandingLens allows users to identify discolouration and tears in the fabric. AI then learns from that data to do the same. This naturally reduces the amount of work that needs to be done by human inspector.
We can start to empower every accountant, every store manager, every buyer and every quality inspector to build their own AI systems.
Final thought
The approach described above essentially creates added business value by having existing employees train their own AI replacements. Hopefully later becoming their human supervisors. Although most likely transitioning to a completely different role altogether.