Why data skills matter in an AI world

Imagine a typical morning, not so far removed from the digital age we now live in. You wake up, and your smart speaker gives you the latest news and weather. You don’t have to worry about things running low in the fridge for breakfast – it’ll update your next grocery order automatically. You get in your autonomous vehicle which whisks you to work, and when you arrive, your computer has already sorted, summarised and prioritised your emails for you. Artificial intelligence is taking care of all the little things, so you can focus on what matters.

Now imagine this. You wake up and your smart speaker is giving you the weather forecast for Boston, Massachusetts. But you live in Boston, Lincolnshire. Not a great start. For some reason, your fridge is full of semi-skimmed milk, which isn’t much use given your dairy allergy. The vehicle drives you safely to work – or it would do, if it hadn’t got stuck in those roadworks on the M1. And when you get there, your boss is shouting at you, because your computer had moved their critical email to junk.

The second scenario might seem like an AI nightmare. But what if I told you that the algorithms behind each of these situations were identical? More often than not, when AI goes wrong, it isn’t the logic that’s the problem, it’s the data that drives it. That smart speaker knows you live in Boston, but doesn’t have enough information to know which one. No-one told your fridge you had a dairy allergy, or your car that there were roadworks. It’s a cliché, but bad data means bad AI. You can have the most powerful algorithms in the world, but if the data they use is inadequate, then so will be their outputs.

The fundamental challenge that we have in the AI era is that there is just too much data. On the one hand, AI can be helpful in trying to sift through it all. But on the other, it isn’t necessarily very good at knowing what’s important on any given day. This is where data skills come into play. To be able to get the most out of AI, it has to be trained, and it has to be given context. Both of these things involve data, and involve knowing that, for example, someone living in the UK is more likely to want a weather forecast for Boston, Lincolnshire. They involve making sure that the latest information about roadworks is available for route planning, and that if AI is going to make decisions about what to add to a grocery shop, it needs details of allergies, as well as important context such as “milk is a dairy product”. It’s down to humans to provide all of that information in a structured and meaningful way.

According to the recent research on AI conducted by Chartered Accountants Worldwide, 79% of respondents agreed that accountants are playing an increasingly important role as ‘data guardians’. This recognises the crucial role that accountants play, in their organisations or for their clients, in making sure that financial data (and often non-financial data too) is available to be acted upon on a timely basis. Accountants are very much seen as the custodians of the data which typically sits at the heart of a business, so to have that responsibility requires the ability to know what you’re dealing with.

To take another analogy – working with data is, in many ways, a bit like being a musician (says the musician-turned-data specialist). Music, written down, is largely pointless. ‘Reading’ a piece of music won’t give you the experience of a Taylor Swift concert or the London Symphony Orchestra live on stage. You have to take the dots on the page, interpret them and create your own story out of them. It is the performance of the music – its presentation if you will – that gives the music meaning. This is true of data too – being able to work with data is a skill, and for that data to have purpose it is necessary to both understand how it is constructed, and be able to turn raw, unfiltered information into something of value.

Before the AI revolution, the question I was most often asked was “should accountants become data scientists?” My answer, without fail, was “no”, but I was adamant that they should understand what can be achieved with data, and arguably that’s becoming more important than ever. Data skills are, in essence, about having enough of an understanding of data to be able to use it effectively. There are lots of parts to this, but they broadly boil down to the following:

• Core data literacy – understanding at a fundamental level what data is, and how it is structured, stored and processed.
• Data governance and management – appreciating the meaning of data quality, integrity and security – in other words, knowing what data you’ve got, where it is, who’s got access to it, and what’s needed to make use of it.
• Data analysis – the ability to use widely available tools to transform data into formats that enable interpretation and exploration.
• Data presentation – using data to effectively tell stories, develop a deeper understanding, and ultimately make decisions.

Developing these skills doesn’t have to be a daunting task. And it’s no coincidence that our Analytics in Power BI Certificate for Finance Professionals is structured in a way that addresses each of these elements in turn, starting with establishing the foundations of working with data, and developing a crucial understanding of what can be achieved with data. All built around a tool that many finance professionals recognise as having huge potential in their roles: Microsoft Power BI.

If you’re worried about using Power BI, you needn’t be. Helpfully, Power BI is very accessible for most finance professionals. If you’re used to using Microsoft Excel (and let’s be honest, most accountants are!), there’s a lot in Power BI that will feel reassuringly familiar, even some of its most advanced features. It also has a low-code interface, meaning that you can do most of what you need to without having to write complex algorithms.

If you’ve ever used Power Query in Excel, you’re already on the road to using Power BI, as the process to import and transform data is exactly the same. And even if you haven’t used Power Query before (and if not, take a look at this ICAEW Insights article which gives a taste of what you’re missing), the ability to perform calculations, to pivot and to derive meaning from data in Power BI is designed to be intuitive. It’s all about taking the sort of work that many accountants are used to performing in Excel and turbocharging it.

Building data skills starts from a place of curiosity – wanting to understand more about the data that you’re working with on a daily basis. What is the data telling you – or perhaps more pertinently, what could it be telling you? What frustrates you about it? What excites you about it? What happens if you dig a little deeper, or look at it from a different perspective? And how do other people use that data – are there things they wish it could tell them?

Don’t be afraid to start asking questions of your data – it might surprise you with what it reveals. So next time you open some data in Excel (or your tool of choice), ask yourself – “what does it mean to me?”.