4 recommendations to get your organisation moving beyond the AI tipping point
Do you remember that day a few years back when your boss, maybe even the boss of your boss entered your office asking why you had not already developed a mobile app for the millions of smartphone early adopters? Probably the same person who believed it was pointless, with limited business impact a few months ago. Well that's the tipping point for technology: the moment in time when your energy moves from convincing everyone that it's worth doing it, to trying to catch up with the demand. We are getting there for AI, and if you don't want to rely only on sheer luck it takes organization and focus. To get you up there successfully I recommend that you set yourself on the journey of doing 4 things.
1 - Industrialize the development of bots
Bots and virtual assistants are a great way of starting to use AI. They are now easier to build, as they start entering our daily lives we accept them better, and they can bring significant operational gains. The challenges are very much in designing great experience, and using AI to manage the cognitive aspects and the ability to automatically get better as people use them. We are in the same situation as mobile applications 10 years ago: exploration will end extremely rapidly, to move to the industrialization phase. That's why you need to get organized: build bot factories now. If you haven’t really started to work on bots, employee centric applications are a great way to ignite (e.g. IT desks & internal service desks).
2 - Have AI preprocess large volumes of data
AI is getting to be very very good at analysing image, voice or text. Google, for instance, reports that they can understand speech better than humans. The more structured the content the better the results are. Spotting a ship in a satellite image, extracting facts from technical or legal documents, or automatically classifiying millions of users’ mail requests are examples of typical use cases where accuracy is outstanding. Algorithms are now scalable, rational substitutes to the man days you allocate to low level data processing. Look for semi-structured images or formal texts to leverage as a very good starting point to ignite this opportunity. If you can choose, don't start with the analysis of potato images or comments from blogs.
3 - Explore the potential of human - machine collaboration
AI opens up opportunity to rethink and redesign systems that solve inefficiencies. Sometimes humans are responsible for these inefficiencies, but the answer is not to remove humans from the equation, but to focus on what they are good at: expert rules’ definition, creative processes, result interpretation, global systems supervision... The answer, and a fundamental benefit from AI, lies in the capacity to design and deploy cognitive systems, in which humans and algorithms successfully collaborate, and that improve over time. If you are looking at igniting this collaboration, look at cognitive searches within knowledge databases or supervised semantic market watch as good first application domains.
4 - Build trust and confidence in algorithms
No driver feels safe the first time he lets its hands off the wheel to let a car drive by itself. It takes time, it takes confidence, but we won't get the full benefits of driverless cars if we don't work in making this happen. In the same way companies have to create the conditions to hand over the daily run of some of their key business processes to algorithms. Software automation and process robotics is inevitable, but unless you are a total AI believer, start by focusing on well controlled actions, with high volume, where humans do mechanical and repetitive tasks (e.g. simple back office processes heavy in copy-paste across applications)
As these 4 opportunities show, it makes sense for enterprises to start embracing the potential of AI where it answers well-known pain points. But that cannot be the target. To get the full potential of AI, it will take much more time, effort and commitment. Enterprises need to work out an ambitious plan, to do more than small changes in the way they work. They need a global AI transformation plan, in which people are as important as technologies. Do you also have the feeling that companies need an AI plan?