Structured data: accountant killer, insurer thriller, Mike Daly

MTD. Making Tax Digital.

Sounds so benign... so modern. Yet for many small accountancy firms in the UK, this is a death knell.

What makes us use accountants? Mainly, it's because they convert our invoices, receipts, spending and allowances from an apparent jumble for mere humans into a form that HMRC uses to calculate the correct donations we must make to HM Treasury and UK plc. For the last 40 years, much day-to-day accountancy has been the alchemy of applying complex rule sets and decision trees to a mix of structured and unstructured data (for structured, think "I've entered it all on a spreadsheet for you"; for unstructured, think "Here's a bundle of my taxi receipts...")

With MTD, everything in accountancy will need to acquire the structure it previously lacked. This shift will, in turn, mean that Xero (most likely but could be others) will kill off the high street accountant within a decade. A well researched and coded SaaS won't stress about the size or complexity of the rule set or the thousand-branched decision trees that underpin tax calculations. So long as the rules are clearly expressible in logical/arithmetical form, then structured data can be processed into tax returns with huge and brutal efficiency.

We all grew up hearing about 'creative accounting' but no-one has yet been able to explain to me where the creativity lies in taking multiple alternate courses of action and doing a calculation as to which untrodden path of the decision tree complies with 'the rules' while delivering the best economic outcome. The creative part is usually just the use of non-standard routes through the tree to end up with a lower tax bill where the leaves emerge!

The three areas where human accountants have held dominion are: 1) the methodical application of structure to unstructured data (ie entering your receipts into a spreadsheet) 2) knowledge of all the rules that most of us would sooner eat our own hair than get familiar with; and 3) applying said rules - and the decision forks they sometimes throw up - to the results.

With MTD, once all tax data becomes structured so there are no barriers to robo-accountant unless you need the human interaction and reassurance of Fred Figures ACA and the bill that accompanies his soft skills!

MCD- Making Claims Digital

In the insurance world, however, the contrast could not be more stark. Most of the 'juice' of claims is locked up in unstructured data and for decades insurers have only been able to interrogate what that they have manually moved from printed letter or email into spreadsheet or reserving software.

The UK claims disruptor 360Globalnet - who have built a powerful arm of their no-code system to turn unstructured into structured data - have a great example. A cyclist knocked off his bike by a driver sustains a broken hand. In the structured data, this is the broad synopsis of the claim and a reserve consistent with such injuries is applied to the case. However, when the unstructured data pertinent to this case is retrieved and analysed the emails and letters point to a very different outcome. The cyclist is a top surgeon who has damaged his operating hand and will require over six months of rehabilitation and even then might not regain some of the fine motor skills required to cut it (apologies!) in this profession.

In financial terms, this information has moved the cost of the claim from the thousands to the hundreds of thousands... and maybe beyond depending on the outcome.

For every insurer, early knowledge and action invariably mitigates cost and hassle. The best medical attention applied fastest usually has a massive effect on long term prognoses and costs. The body adapts around damage quickly because evolution has designed us for homeostasis... but that adaption is often quick, dirty and effective rather than ideal (think scars, bone spurs and snot!) So it's always best to set the body on the path for resolution rather than adaptation.

By extracting actionable business intelligence out of their millions of PDF, JPGs, text files, office files and emails (where data is available but only one document at a time), insurers can see the previously unseen. They can analyse everything about a claim, not just the figures that a data entry clerk has tapped into a shared spreadsheet.

Digging deeper, they can use re-structured data to test hunches, find and tackle fraud, spot emerging trends, mitigate claims costs, improve reserving, enhance staff productivity and slash their expense ratios using a light sabre rather than a spoon.

It will change the dynamics of the industry for sure. Meanwhile, for many accountants, however, I fear that - with MTD - the light sabre is double ended and twirling towards them with frightening ferocity.

NB; published with consent of author