There’s a lot of scaremongering about the idea of artificial intelligence (AI) taking our jobs – but I think we need to consider where it fits well and that there are some things it’s not so good at.
We need to understand the nuances. AI, for some, is being seen as the latest digital solution that will transform everything. The idea that you can just implement AI and the world will be a much better place is not the full picture. Like all things, if your data is poor, your results will be poor; rubbish in, rubbish out.
AI is really good at some things, but less good at others and it absolutely needs to be managed by a person to get the most out of it. We need to understand how it’s working and then tweak and refine it. We need to have transparency, so we understand the decisions it is making and can challenge them. Another consideration is the need to look out for biased data sets, because whatever AI uses as its source data will shape the results on that bias.
AI is everywhere! For example, it is really expensive to launch a new drug into the health market: AI adoption can shorten the time this takes and play a factor in deciding which drugs might be more successful before undertaking expensive trials.
Banks are using it for fraud protection; energy companies use it to monitor customers’ behaviour; Network Rail and others use it for the diagnosis of where faults may occur to enable preventative maintenance, saving staff time.
Olay are using it to help ladies pick the best facial products for them. So, if you take a photo and answer a few questions it will recommend the best product for you. There’s an interesting discussion on how much the marketing team should get involved to shape the results. At present there is no input from marketing. The deciding factors are primarily driven by what you say rather than how you look! AI is also being used by sports teams to monitor how players are moving in training, which can predict injuries and assist talent-spotting.
The key approach used by Aylesbury Vale is more about automation than just AI; I mention this because it is a matter of scale and timing. We’ve been automating and promoting self-service to our customers for a long time, and AI is just the next iteration of that. We are also using AI through Amazon’s Alexa to answer customer queries in their home.
In addition, we use AI software in our call centre: it takes the text of incoming webchat messages, or emails, and helps the agent by suggesting possible responses. It’s not ‘instead of people’ – it’s AI software assisting the agent working alongside them. The benefit is that agents then have more time to manage the more complex enquiries.
AI is great at dealing with lots of data, and much quicker than we are at crunching through that data to find patterns. It does this in a standard way and learns by working alongside our staff.
For me, AI is about augmented intelligence: augmenting humans to make decision-making quicker.
This is based on Maryvonne Hassall’s talk at Socitm’s President’s Conference in Glasgow on 8 May 2018