Maximising IoT business benefits with AI

The full benefits of enterprise IoT cannot be realised unless the business leverages on artificial intelligence (AI) to drive insights and action. But all that sounds very elaborate and expensive. Does it make sense for the bottom line? Why are such levels of sophistication necessary and if so, how much time and money will it cost?

In a nutshell, the reason why AI is required in IoT is because connecting many devices together generates a huge amount of data very quickly. Furthermore, this data could come in various forms such as text, audio, pictures or videos. It could also be ‘unstructured’ (not organised into suitable formats for analysis, for e.g. email content). To digest and analyse the mountain of data with a human operator is inefficient and also, in many cases, impossible.

Take for example, Singtel’s partnership with Mobike, a bike-share company. Its fleet of seven million bikes are equipped with smart IoT locks which generates over 30TB of data daily!

Given that data is real-time and continuously accumulating in IoT systems, the only practical way to utilise it for optimisation and prediction is to leverage AI – specifically, using machine learning models.

IoT devices at the edge of a network are like the five senses of the human body. They perceive, collect and transmit information to a central server for storage, processing and decision making. AI is the brain that runs this ‘body’. Without a good ‘brain’, no matter how well designed or extensive the ‘body’ is, the full potential of the system can never be achieved.

Therefore implementing IoT without progressing to being a cognitive enterprise is the equivalent of staying unevolved as a basic, instinctive organism. You will not thrive or even survive against those that have evolved to become much stronger and better.

The Frost & Sullivan IoT Actualisation Quotient study which surveyed businesses across Australia, Hong Kong and Singapore backs up this reasoning. Enterprises that utilise AI in their IoT systems achieved an average positive impact of 16.8% across eight business performance metrics, as opposed to just 8.6% for ‘Connected’ enterprises and ‘11%’ for ‘Data-First’ enterprises.

There are three key ways we see businesses changing once they reached the cognitive stage in their IoT maturity:

• Predictive and adaptive processes. Before AI came along, machine-to-machine (M2M) connections made linear decisions i.e. if event ‘A’ (a breakdown) occurs than the response would be ‘Z’ (something on the dashboard lights up to alert the operator). Systems were reactionary and ‘after the fact’. With modern AI, machines can predict events before they occur and adapt to new inputs and variables as they are introduced into the system. The implication of this for enterprises adopting cognitive IoT is that business processes must be redesigned along predictive and adaptive responses as well instead of reactionary. Problems should be anticipated and fixed before it happens, and products and processes are continuously rather than routinely optimised as operating conditions change. The emphasis would be on knowledge workers who can design, monitor and fine-tune IoT systems to yield maximum productivity.
• Enterprises will move towards cloud computing. AI requires a large amount of computing power and expertise to run and manage. This is something smaller enterprises may not be able to invest in and maintain exclusively for themselves. However, they can leverage the affordable and easily scalable cloud computing and Software-as-a-Service (SaaS) solutions out there to integrate AI into their IoT systems. Except for very large and/or highly security conscious enterprises, it makes better sense for enterprise IoT systems to leverage on cloud and SaaS solutions to drive it, so that initial investment is smaller and recurring costs scale up in line with the business.
• Products and services will be redesigned. Businesses will change their product research and development, manufacturing processes, as well as customer service and support in the IoT age. This is because IoT will enable them to be linked directly to the customer and gather continuous feedback and performance data. For example, home appliance manufacturers will no longer just sell their product once and provide warranty services when customers contact them with an issue. With IoT enabled products, they can monitor the usage of their products, monetise that consumer behaviour data, push software upgrades, service fault complaints remotely, and develop cloud-based services to supplement or complement their product. Such product redesign will be driven by AI as businesses increasingly compete to make their products as ‘humanistic’ as possible to make it user friendly.

How much (more) investment is needed to become a cognitive enterprise?

There is no simple way to gauge the cost of implementing or upgrading to an AI driven IoT system since it depends on the size of the organisation, its legacy systems, and the extent of the implementation across the organisation.

The good news is today, automated machine learning tools, out of the box sensor kits, SaaS based cloud computing services, API-based solutions etc. make it easy, cheap and scalable for even small businesses to integrate AI and IoT into their operations.

Every business must invest in technology to stay competitive. The only question is, how much and how soon. The answers depend on each enterprise’s circumstance, but the first step is always to initiate a conversation with a competent technology partner to help you make that evaluation.

Speak to our IoT solutions team today.

For more information on Singtel IoT, visit www.singtel.com/iot.