In the last article, we shared some common misconceptions between the meaning of Artificial Intelligence (AI) and Machine Learning (ML), so you can decipher for yourself if the salesperson is speaking the truth when he tries to sell you a product, claiming it can do AI/ML.

Now that we have established the correct definition of the terms, we will be sharing some examples of how we have infused the use of Artificial Intelligence (AI) and Machine Learning (ML) into our mobile apps to enrich the experience of it.

Infusing AI/ML in our work

“Any sufficiently advanced technology is indistinguishable from magic.”

Clarke’s Third Law, Arthur C. Clarke

Our company, Originally US, is founded by two like-minded individuals (myself and our CTO, Torin Nguyen) with strong backgrounds in technology. Both of us started programming in our secondary school days, due to our long-standing interest in coding. I, for one, learnt how to operate the command line and touch type when I was 4 years old.

We founded Originally US to ensure that we can continue to pursue our passion for programming and technology without having to rely on other corporations to pursue that passion. Hence, Originally US is a perfect vehicle for us to pursue our passion while being able to feed ourselves and our family.

With this kind of mindset, the culture we set in the company is one of exploration and innovation. We often pursue new, interesting technologies and encourage our developers to experiment and innovate during office hours, even if what they work on does not have immediate relevance with work being done for clients at the time. This ensures our company is constantly at the forefront of technology.

While we call ourselves a mobile application consultancy, we have begun infusing the use of Artificial Intelligence technology and concepts in our work as far back as 5 years ago. Our aim for dabbling in Artificial Intelligence is to further improve our software capabilities, further provide more value-add to our clients, and most importantly, to satisfy our own intellectual curiosity.

While we cannot go into too many details due to the confidential nature of some of the implementations, here are some brief examples of the significant role AI/ML play in enhancing our mobile apps’ capabilities.

AI/ML in SG BusLeh

SG BusLeh is one of the most popular apps we have ever created. With more than 9.6 million app launches every month, it is one of the most popular apps in Singapore.

Because of the app’s popularity, we are in a unique position to have big data that can be used to benefit our users and give us an edge over competing apps.

From having an in-depth understanding of our users, we realised that one of the most important features (apart from looking at bus arrival ETA) is to let commuters know in advance how crowded a bus would be when it arrives at the user’s bus-stop.

While Land Transport Authority (LTA) already provides a crowd level estimate (in three distinct states, “seating available”, “standing available” and “limited standing”), these are semi-realtime values that are updated once per minute that bring about inaccuracy in reflecting real-life situations.

Allow me to illustrate the issues with using these values to determine the bus crowd level. For example, a commuter may be waiting for a bus that is 6 stops away. By looking at real-time values, the bus may still have “seating available”. However, this does not mean that the bus would still have seating by the time it arrives at the commuter’s stop. Therefore, looking at the real-time crowd level data may not value add to commuter’s decision-making for their commute.

We recognised that crowd level data would only be useful to commuters if it can predict how crowded the arriving bus will be when it arrives at the commuter’s stop, instead of only reflecting the crowd level 6 stops before. Hence, we decided to build a Machine Learning model around the bus stop crowd level for every bus-stop and bus service in Singapore and use it to predict how crowded each bus will be when it arrives at the bus-stop.

With Machine Learning, we are able to model and predict the crowd level at all bus-stops in Singapore for all bus services 24 hours a day, 7 days a week. This feature provides much needed information for our users, allowing them to travel and commute smarter.

AI/ML for Price Kaki

Price Kaki is a community-contributed price comparison app that we have developed for the Consumer Association of Singapore, supported by the Ministry of Trade & Industry.

While the key problem we are trying to solve here is to help lower to lower-middle income families make better purchase decisions for their daily necessities, the main feature of the app is its reliance on the help of the community (“kakis”) to help capture contribute the latest price and promotion data in the app.

With a large influx of User Generated Content (UGC), moderating and administering these contributions prove to be a daunting manual task for our client.

To help ease our client’s workload, we built Artificial Intelligence models that can evaluate each user submission and determine if a submission should be automatically approved or deferred to a human administrator.

At the same time, we are doing further Machine Learning work to train our model on the millions of user submissions we have collected so far to further refine our processes and automatic evaluation.

AI/ML for AIA Singapore

We have been working with AIA Singapore since 2016 to enhance the productivity of their financial advisors through digital transformation. Recently, we have also begun expanding the project to collaborate with regional AIA offices as well, including Malaysia, Thailand and Indonesia.

With the vast amount of data we have collected on how financial advisors use our solution when talking to their customers, we are also able to build up a Machine Learning model that predicts and recommends the most relevant products and documents to a financial advisor based on their activities.

Closing Thoughts

AI/ML enables new capabilities that can augment existing solutions, allowing solutions to provide even greater value-add and capabilities to the users. It is a wonderful tool that greatly enhances any mobile app.

However, many vendors/developers have abused AI/ML to position their solutions in a much more sophisticated light than they actually are.

If you are procuring any AI/ML related solutions or services, it is important to conduct your own due diligence on the vendors/developers’ claims first.

At Originally US, we do not embark on building a mobile app unless we want it to be successful, because your success is also our success.

If you’re looking for a trusted mobile app development company to build your app, let us discuss how we can help you! Chat with us on Whatsapp today.