AI - The Next Big Thing and The Practical Ways It Could Be Applied In Your Business
In analysis / By Mark Flynn / 10 September 2018
AI is going to be the next big thing for just about everything in industry. This is backed up by the likes of the Bank of England stating the UK's economy needs an AI kick-start. We also have McKinsey stating AI enabled solutions will increase workforce productivity by 40% and finally Accenture believe that between 2018 and 2022, banks that invest in AI could boost their revenue by an average of 34% and their employment levels by 14%.
So investing in AI looks like a no-brainer, but if you are thinking of dipping your toe in the AI waters, then it is important to understand three key things: What really is AI, how you might implement AI and very importantly are you ready and able to take full advantage of it.
Firstly, what is AI? There are a lot of different terms being used to describe AI so being clear about what you are talking about is important. As such I’ve laid out the basic terms to make sure there are no misunderstandings.
Chatbots, or conversational AI, are perhaps the most common kind of AI specialising in talking to people. Chatbots work by recognising keywords and phrases and then responding to what has been said. Chatbots are most often used in customer service either through auditory or textual methods to help answer any questions a potential customer has, as well as helping to start forming a bond between the potential customer and your company. An interesting example is the chatbot lawyer, DoNotPay which was used to overturn parking ticket fines but also help refugees claim asylum.
Machine Learning is often characterised as ‘dumb AI’ or “narrow AI” but in my opinion, it is better to think of it as a far more specialised AI that focuses on data mining and prediction. By learning from the data, machine learning can go on to make predictions or notice trends in behaviour. An interesting example of this came from when Netflix launched a competition to find a machine learning program that would allow them to better predict user preferences.
Finally, we have Deep Learning, which are systems inspired by the human brain. Each layer in a deep learning system will either alter the input or be looking for something different in order to gather the most amount of information from what it is looking at. An example of this is the Drive PX system which uses deep learning to recognise hazards on the road and then help the car respond appropriately.
Knowing these three terms and how they can be used means you can now consider the AI solutions available to your business.
Doom-mongers say that AI will replace the workforce, but this just isn’t true. The sensible view of AI suggests that AI will be of far more use in augmenting employees so that they can do their jobs better and faster, we will see in those businesses that get AI right a core set of people supported by AI systems making quicker/better business decisions and delivering first class service. An AI assistant could be used for basic things like helping to create schedules and automating paperwork, to more complicated things such as modelling data or doing research.
AI can also be used to enhance every part of a customer’s journey with your business. A great example of this is Starbucks. Before a store is even built an AI will figure out if a location would be profitable by analysing data such as foot traffic or local demographics. After that those same systems will predict how badly other Starbucks will be affected by the new store opening.
Once the store is open other AI’s will make use of the data provided by the Starbucks reward program and mobile app. Using this data recommendations can be made to customers based on past likes. Starbucks recommendation program, however, is more sophisticated than most because it will alter recommendations based on variants such as the time of day, weather, holidays or other notable events.
AI will have a significant impact on delivering great service to your customers using conversational AI or chatbots. IBM's Project Debater is a great example of the sophistication of future chatbots.
Project Debator understands the context of what someone is saying during a debate, creating its argument in real time. AI assistants based on innovations like Project Debator will deliver exciting opportunities for every business, improving how chatbots interact with people and turn them into really powerful business tools.
So where should you start? Here are 5 uses you should consider for AI in your business:
- Automate anything that is time critical or reducing time spent manually matching payments and invoices
- Analyse all previous deals won/lost to help improve forecasting, accurate profit estimations per deal and more accurate close dates on deals
- Use AI to analyse video interviews that scan facial expressions to check the suitability of a candidate
- Use AI to support staff retention activities, to pull together data from exit interviews, performance management, surveys and discussion threads to build a real-time picture of the health of your business
- 1-2-1 conversation with your customers is difficult, so use AI to process the massive volume of data you hold on them to help create a hyper personalised service for your thousands of customers
Once you understand the power and innovation of AI it is important to ask the question are you ready to add an AI to your business. The answer for many that have already invested in AI is no.
The first question is do you have the appropriate analytics and systems in place. An AI has to be fed with up to date data otherwise it will come to the wrong conclusions. To do that you have to make sure that your data gathering process is as automated as possible. If it takes you a month to gather all the information then your AI is only going to be able to help with last month’s problems.
You also have to structure your data in such a way that the AI can make use of it and give you useful results. The ability to look at customers entire history with your company is much more useful in certain circumstances than the history of a product but the reverse is also true and the ability to view the data in different ways will only increase the AI capability.
Of course, that all assumes that your data is correct and useful. If it’s not then that introduces problems with your AI. Many assume that an AI will be bias free but in truth, an AI is going to have as much bias as the data it’s been fed on.
A rather horrifying example of this is the chatbot, Tay. Tay was designed to mimic a teenage girl on twitter and interact with others, but this backfired. In just 24 hours twitter users managed to train Tay into being a misogynistic foul-mouthed Nazi. As you can see knowing what data you are feeding an AI is incredibly important as it can drastically change their behaviours.
Hopefully looking at these different parts of the AI conversation will help you when considering whether to make use of AI in your business.For more information on anything in this article, please [contact me](mailto:firstname.lastname@example.org).