We’re not even close to reaching the pinnacle of the Artificial Intelligence hype cycle, yet not a day goes by when there isn’t an AI story in the news and already billions have been attributed to its value to business.
In April 2018 (long enough ago to be considered the equivalent of a dozen ‘technology’ years), Gartner predicted that the AI industry would be worth $1.2 trillion over the calendar year. It’s mind boggling to both the average person on the street and industry alike, as the media talks about AI as though it can potentially be applied to almost anything, yet the reality is that much of it isn’t really AI at all. Just a technology-led new way of doing something.
In business, it’s all about familiarity. That is, understanding what AI is capable of doing and whether it can help you do something better or more efficiently. If it doesn’t, then why bother?
For example (and bear with us, as it’s about as far away from the exciting world of robotics and autonomous cars as it’s possible to be), take the processes in the average office environment. There are multitudinous ways in which AI can analyse data for business. It can provide insights into future trends, rather than taking a backwards look to inform decision making. It can manifest itself in the automation of repetitive tasks (“taking the robot out of human”) and mimic human behaviour to collect and extract knowledge, recognise patterns and adapt over time
But once you’ve recognised an opportunity where AI can give you rewarding results, how do you go about realising it? Well… people.
The reality of AI is a world away from ‘the robots are coming’ scenarios that we see in the media. There are literally thousands, if not hundreds of thousands, of potential applications that could quietly improve our daily experiences. If you’re looking to better your business through AI, then once you’ve identified your particular flavour of Artificial Intelligence, you need the know-how to understand where it fits in and further understand how to implement it. In type, it looks straightforward, but does the solution you want exceed the capabilities of your organisation?
It’s not rootless negativity. For example, if a deep learning solution has the potential to transform your business, then the data you use must be squeaky clean – ‘garbage in, garbage out’, as the old adage goes. But internal data can, quite incidentally, be spread out and serve many different purposes, across many different teams, departments and even business groups – none of which can be easily combined. Is this a cost you’re willing to foot? Can you afford the investment? And is your accuracy or your legal compliance guaranteed?
In a nutshell: Think before you leap.
The reality of AI is a world away from ‘the robots are coming’ scenarios that we see in the media.
As tempting as it might be to metaphorically sweep your hand across the desk – despatching everything old into the trash and making way for a brave new world – drama has no place in business. That’s not to say that you should move to AI at a glacial pace. ‘Considered evolution’ might describe the best approach, where you look at AI as a natural part of growth and development of your business.
A retailer, as an example, might begin with intelligent ‘triage’ in their call centres, screening customers and directing them to the right contact centre worker. Or chatbots could deal with customer queries, using AI to provide responses. At the most sophisticated level, Ocado uses deep learning to detect fraudulent activity on its website. This flexibility is hugely appealing to businesses as it plays to the aspiration of the ‘agile organisation’. But box-ticking aside, it also gives them the chance to prove value, collect feedback, and respond.
In a nutshell: Be cautious (but not slow) in your curiosity.
It’s all thoroughly exciting, though isn't it? You’ve taken the time, identified the business need, been pragmatic and considered, formed cross-business taskforces to work on implementation and invested wisely in the tools required to deliver. Don’t crack out the bunting just yet.
AI doesn’t have an end date.
So, who is going to keep thinking and innovating, manoeuvring your AI strategy into the right areas, at the right time? Those stalwarts of business intelligence, McKinsey predicted that the United States alone could face a 50–60% gap between supply and demand for deep analytic talent. When market growth outpaces the availability of expertise, then there’s a serious problem. It’s only cultural change that can address this shortfall – through the time-honoured tradition of investment in people, training and development.
In a nutshell: Deep learning isn’t just for algorithms.
You can’t cherry pick Artificial Intelligence. You either adopt it technologically, culturally and foundationally, or you simply don’t. Embracing new technology has never been an easy ride, but it’s certainly proven to be an exciting one, even when you plan meticulously, expect the unexpected.
Amara’s Law: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
In a nutshell: That’s not a bad thing.