As the world of accountancy emerges from a break over the summer, we find ourselves heading full tilt into an autumn season of big events. But from a data analytics perspective, what have we learnt so far?
“Unify your AI-powered lake warehouse”
Attending Big Data LDN in September was an opportunity to see and hear what the major players in the wider analytics space are up to. It’s clearly a very crowded market, with data pipelines one of the key talking points. Here at ICAEW we touched on this earlier in the year in the context of client data extraction but actually the principles, and the criticality of being able to access data consistently, applies equally to all organisations. Solutions which make this as easy as possible are going to play a big role in the finance teams of the future, as they increasingly need to draw on more and more disparate sources of financial and non-financial information. Gone are the days when all the reporting data you needed sat in your accounting software!
Unfortunately, the world of analytics is rife with an array of buzzwords, many of which are meaningless to the average business user. Data lake? Data warehouse? Data lakehouse?! The reality is that most organisations just want their data to be accessible and accurate, and don’t really care where or how it is stored (although arguably they should care, if only from a data security perspective). We know that those involved in technology procurement are struggling with overwhelming choice, so it was a shame that many of the vendors at Big Data LDN were saying much the same thing, drowning each other out in a noise of AI-driven sales speak. We hope, therefore, that our upcoming webinar on Microsoft Fabric will start to make sense of some of this complex landscape and make it a little more real and digestible. Those organisations that successfully integrate their key data sources into a single platform for analysis, reporting and insights will have a competitive advantage in the decision-making process. As will the organisations that invest in transferrable data skills, moving them out of data team silos and spreading data skills across the whole company.
Moving on so soon?
It was noted in a session at Big Data LDN run by Eden Smith that “every three months there’s a new piece of tech and we all turn into kids in a sweet shop”. This turned out to be a remarkably accurate assessment of Accountex Summit Manchester, at which the hot topic was not AI, but AML, as if the London version back in May was from a bygone era.
AML checks – and client onboarding in general – are widely considered to be a necessary evil for the accountancy profession, taking up valuable time and resources and being a key source of regulatory risk, so it is perhaps unfair to accuse accountants of getting distracted by the next shiny toy when getting it right (or wrong) can make a huge difference. But data analytics and AI can play a useful role here too. Process mining is a branch of analytics that explores the pathways that users take through any given system, and can be used to identify inefficiencies and outliers. Given how cumbersome AML and KYC processes have become, perhaps it is time for firms to explore process mining tools, while also ensuring from an audit trail perspective that the onboarding process is adhered to consistently. Other opportunities lie in wait too, such as advanced data capture solutions, open banking, OCR technology and yes, AI, to quickly and easily obtain all of the necessary client information in a robust, structured way, while the use of dedicated AML tools can also support good data governance, mitigating the risks around data breaches, a particular concern in the AML process due to the high volume of personal data involved.
One area in which AI did appear to be gaining traction at Accountex was around tax. With the complexities of UK and international tax law patently obvious, the use of AI to support tax advisors and software platforms seems to be a logical move. Several providers in this space are starting to integrate ChatGPT-style engines into their products, trained on HMRC manuals, to enhance their offerings. Indeed, it will be the topic of one of the sessions at the upcoming ICAEW Annual Conference. Using AI tools to analyse provided data, trawl through the content of the manuals and pick out the relevant sections is a simple but effective way to support tax professionals and allow them to spend more time providing meaningful advice.
Human-in-the-loop
This very much ties into a key theme which emerged at the Accountants in IT Conference, but was also frequently mentioned at Big Data LDN - AI has to be about more than just the algorithms. The concept of ‘human-in-the-loop’ came up several times, by which is meant the notion that humans still have a critical role to play in supervising, reviewing, checking and supplementing AI outputs. The AIT conference highlighted the opportunities for AI to use complex, disparate data to drive better decisions in areas such as resourcing, leading to more equitable work allocation and happier staff, but crucially not doing away entirely with resourcing managers who still must, inevitably, have the final say in the work allocations. The latest tools (such as the ones ICAEW are using to enhance it’s Mia chatbot) make this process of oversight much more manageable and accessible to most organisations, as it reduces the burden in training the AI engine, allowing resources to focus more on what matters most – the experience of the end user.