I’m sure you’re with me when I say we’ve all been seeing and hearing a lot about Artificial Intelligence (AI) - machine learning, chatbots, voice skills and automation at the moment.  Recent research* suggests that 86% of large organisations surveyed were in the mid to late stages of developing AI solutions - up from 65% the previous year.  Not surprisingly, the same research found that 57% of these organisations cited competitive advantage as the driving force behind their investment in AI technologies, up from 16% percent the year before.

So, if AI is the driving force behind organisations developing competitive advantage, where do you start?

Here at DiUS we’re seeing a lot of organisations (large, medium and small) really curious about the application of these types of technologies - keen to understand what it could mean for their customers, their business  strategy, competitors, and in some cases the future of their jobs. We’ve always got an eye on the new tech, but we also temper that with the practical - is it actually useful, is it ready, and does it suit the problem at hand.

Understanding the technology is important, but just as important in our view is understanding the problems or opportunities the technology has the potential to solve.

The temptation to rush into spending time developing a solution to a problem that may not actually exist, or build a competitive edge when there is no edge to be gained- is a sure-fire way to fail at AI implementation. Getting it wrong could have some significant impacts on your customers, and ongoing implications for your brand, and reputation. The technology is brilliant, and it can solve a range of problems - as well as deliver businesses a significant platform to develop competitive advantage, but there are some key considerations for businesses embarking on the AI journey.

How’s your data?

For a machine to learn - or for something to become intelligent - you need to be able to feed it the right information and data.  Gathering, storing and being able to access the right data is the key to being able to develop AI solutions. New technologies such as IoT and the use of sensors (such as cameras), or voice recognition devices are also emerging quickly - and the amount of data being collected from these sources, as well as the more traditional sources is enormous.

Having an understanding of the infrastructure and software required to collect, store and quickly access the data needed to feed the AI system is important. Knowing what you have and how to access it means you can objectively assess what you may need to do to get into a better position to leverage AI in your organisation.

A business can then plan and effect changes to business systems, infrastructure, strategy, platform or ways of working - and use agile or experimental innovation to determine what the best way to progress might be.

When it comes to the storage side - cloud platforms enable businesses to store large volumes of data, more easily accessible and more economically than ever before - and this is more evident than ever before in Australia with 31% of Australian businesses adopting cloud services (up from 19%) in just one year.

What’s the problem (or opportunity)?

Diving into a technology that’s still evolving can be a master stroke, but it comes with its own set of risks. Seeing an opportunity to leverage a competitive advantage or grab a new segment of the market, or having a well defined problem that can be solved using the technology is a must. It can be easy to get sucked into the early adopter buzz, but if it feels like a solution looking for a problem then it’s probably better to wait for the right time to strike.

Starting small and using innovation techniques - product and design thinking, lean UX, service and human centred design methods - to identify, assess and validate problems or ideas is critical - but the temptation to overlook this stage is ever present.

Thinking about an AI roadmap - and considering the types of problems and solutions and how they will eventually link up is important too.  The application of what’s described as narrow AI, or AI developed to solve specific problems, means that having an understanding and a vision on how these potentially multiple  narrow AI solutions will hang together, is critical for future success.

Companies like amaysim are starting slowly, with a well defined hypothesis in mind.  In fact they engaged DiUS recently to develop a deliberately selected, low-risk use-case - a skill for the Amazon Alexa voice recognition platform - to expand their customer service channels, and use it in a light-touch way to better understand if their customers would embrace this type of technology.

Who’s the end user and what will they think?

Similarly, it’s important to think about how the technology will impact or support your customers, and whether or not they will even be up for using the latest chatbot for example.  User research helps with understanding the customer view, and this combined with a good understanding of the problem or opportunity should at least get you to a position of knowing whether if you build something it will be a) useful, and b) used.

There is also the need to ensure that this new technology or AI tool fits into your end-to-end service offering.  Again, there no point building a smart system unless it’s integrated into the customer experience properly, and continues to be sustainable from a data integration point of view.

Prove it can deliver value

Taking the next step, and deciding how or whether to proceed with an AI idea can be the really challenging piece in the puzzle though. There is still a perceived element of risk within some levels of management and boards for example.

Proving the value of a concept, that it can actually be built, and could work, especially with a new technology, requires someone to take a bit of a risk, some time and some resources - and in most cases help from people who have done it before.

Getting such a proof of value delivered provides not only the evidence required to demonstrate that a project can succeed, but can also provide a team or business the confidence and the experience to realise that they can innovate quickly and in a repeatable way.

Again, the challenge for some organisations will be their ability to deploy technology quickly as a minimum viable product (MVP), and build on that capability over time. This could require a rethink of how new software and technology is delivered in a business - and a move to more agile business practices.

nib Health Insurance is a good example of a business that has embraced the need to develop agility within their business, with results that speak for themselves. After undergoing a major technology transformation, they are now able to focus their efforts on innovation, and getting new products and features to market - that are already delivering value - quickly and effectively.

We think the next twelve months will be the year of AI and IoT in Australia.  We’ve gone from seeing businesses struggling to identify how they could use the technology, to businesses actively identifying opportunities to realise the benefits. The real message though is that it’s not too late to start, and you haven’t missed the boat. In fact if you’re actively discussing adopting AI in your organisation, well - you’ve already overcome the largest barrier!

_______

* Infosys - Leadership in the Age of AI - January 2018