Artificial Intelligence (AI) is all about machines, obviously. Except it’s not. In truth, discussions surrounding AI may often centre around how competent, intuitive and contextually aware the machine brains we are building have become. But really, AI is all about us―the humans—and how it can make our lives better.
There was a time, perhaps even inside the current decade, when AI tools and functions were still associated with the fanciful ‘talking computers’ that featured in many 1980s movies. It wasn’t that long ago that we considered AI as something of a ‘toy’ and its application in mission-critical enterprise applications was still somewhat laughable. Of course, now we take talking computers completely seriously. So much so that we’re equally focused on the proficiency of computer speech recognition.
Application of AI
But as far as we have come, we still need to look at the real world use cases of AI and ask how it can help us make our lives better. If we’re not applying AI to our human work experiences to examine and analyse where it can make those experiences greater, then what are we doing here in the first place?
The truth is, many enterprises large and small have been struggling with finding the appropriate use cases for new and emerging AI technologies. Companies need to find the workflows inside their business models that can benefit from AI. Only then can they start to architect towards turning those operational throughputs into truly digital workflows.
So how do we define AI-enabled digital workflow Nirvana and how do we get there? Typically, the process starts with a technology audit and a process of assessment, quantification and qualification running throughout the IT stack in question. Individual business units will need to step back and identify their work problems and challenges as they look for the elements of their workflows that can be digitised.
Everybody across every line of business function will be involved―we need to crowdsource and collaborate to identify strategic areas of business operations that still exist as predominantly manual, accurately measurable and fundamentally repetitive.
These are the parts of business that represent liquid gold, i.e. once we tap the seam, we can channel these functions into AI-driven services that subsequently run as digital workflows. Individuals are liberated from drudgery, productivity is increased and employees have a greater experience—a new virtuous circle is established.
Think about a typical office. When people leave the company, we need to manage who has a key fob for access to the car park. This is a perfect example of the type of job that has typically been performed manually through the use of a spreadsheet. This is time consuming, error-prone and obviously creates security issues.
But it’s also (I hope obviously) a perfect example of the type of task we can evolve to become a digital workflow driven by intelligence stemming from AI. Our analytics engine should know that an employee is leaving the firm and so reports, alerts, emails and perhaps even mobile device management, to cancel the key fob, can all happen automatically.
If we can make all those things smarter and more intuitive, then we can build better experiences faster. Uber hasn’t actually done anything fundamental to change the way taxis work or drive. It has changed the digital workflow that governs the ability to book and pay for the service. The list of services-centric examples in this space is growing every day.
Automating a bad process doesn’t make it good
In the technology industry, we are often bad when it comes to decommissioning things. Think about how many business processes probably exist today that firms need to eradicate and get rid of.
There’s no point in applying AI to these aspects of the business. As we know, automating a bad process doesn’t make it a good process; it just makes it an automated bad process! So, this re-engineering is actually an opportunity to clean out your cupboard and stop doing the things that you no longer need to do.
An example that came out of a recent hackathon, we conducted, is a tool to help with filing of patents. One of our hackathon teams used AI and ML to trawl the web for all registered patents using word recognition. They wanted to identify connected words to see if a new invention already existed in some form already. This would have been costly manual work, that may have been handed over to a specialist (in this case, a patent lawyer), but now we can digitise these aspects of the business.
The human factor baseline
We as humans now need to engineer the existence of AI into our own mindsets and consider how it can help us work differently. This includes knowing what things we don’t need to worry about anymore. For example, we don’t take a map out with us these days, because we use a smartphone—so what else can we stop doing?
As we move down the more humanised road to AI, we will find that AI itself gets smarter as it learns our behavioural patterns, penchants and preferences. We must still be able to apply an element of human judgement where and when we want to, but that’s already part of the current development process as we learn to apply AI in balance when and where it makes sense.
The future of AI is smarter, and it is also more human. The end result is more digital at the core, but more human on the surface. If that still sounds like a paradox, then it shouldn’t. We’re at a crucial point of fusion between people and machines and it’s going to be a great experience.