Rather than consigning employees to the dole queue, AI and machine learning tools could free people from admin tasks and lead to more fulfilling work.
Spare a thought for all those poor cyborgs and androids. When they were merely exterminating the human race, as portrayed in countless sci-fi films, we looked upon them in awe and admiration. But now, faced with the prospect of robots stealing our jobs, we’ve come to loathe them.
The effect that AI and machine learning will have on jobs, however, is likely to be more nuanced than many internet scare stories would have you believe.
“We’ve spent way too much time thinking about jobs that computers are going to take away from people, and not nearly enough time thinking about people and computers,” says Prof Thomas W Malone, founding director of the MIT Center for Collective Intelligence and author of Superminds: The Surprising Power of People and Computers Thinking Together. “That is, the new kind of jobs that will be created when computers do some things so much better, faster and cheaper.”
There is clearly a pressing debate to be had about the future of those of us who might lose our jobs due to AI and then struggle to find quality replacement roles. But Malone believes this issue, although important, will be “far less apocalyptic than many people seem to be expecting”. Rather than consigning human beings to the dole queue en masse, AI and machine learning could automate many of the more mundane tasks employees currently have to perform, leaving them free to pursue more fulfilling work.
Take data entry, for instance. Nobody enjoys keying information from invoices and expense claims into a spreadsheet. It’s dull, time-consuming and utterly unimaginative work. So why not let a machine do it? Unlike humans, they don’t get bored or tired, which means errors are far less likely to creep into the expense reporting process. And businesses could also save money by reducing the labour costs associated with these tiresome tasks.
This isn’t another case of wishful future-gazing; it’s already happening today in many businesses. Machine learning algorithms are capable of processing vast amounts of both structured data (such as spreadsheet fields) and unstructured data (such as emails or images) faster than any human being. They learn through a trial and error process, and the more data they access the more accurate they become – to the point where they are able to draw their own conclusions about what should go where, or why “A” doesn’t match “B”.
Machine learning algorithms developed by SAP Concur, for example, access millions of aggregated and anonymised corporate travel and expense transactions on a daily basis. “In Q3 of 2018, we processed more than $115bn [£87bn] in expenses,” says Rachel van der Merwe, director of product marketing at SAP Concur, “so there’s a lot of volume and spend that’s going through our system. That helps this technology become much more accurate.”
Employees are able to file expense claims on the move through the SAP Concur mobile app, which uses optical character recognition technology and machine learning to automatically turn photographed receipts – including those that feature scribbly handwriting – into a completed expense claim.
Innovations such as these mean that finance teams no longer have to sift through crumpled paper receipts and invoices to verify an employee’s purchases, or chase people whose claims are only half complete.
“In effect, the database processing that was frequently done by teams of people sitting in an office has now actually been pushed to the user,” says Justin Watson, a partner in Deloitte’s consulting practice and head of its robotic and cognitive automation community. “You can reallocate the people who used to be doing what is, in essence, a fairly boring data-processing role … to something much more fulfilling.”
AI-based solutions can have broader benefits for a business too.
The SAP Concur mobile app also processes data from e-receipts and corporate credit card purchases to ensure nothing is missing or duplicated. And current VAT rules are automatically applied to certain expense types, so the business is always able to reclaim it.
Machine learning algorithms can spot non-compliant or potentially fraudulent expense claims before the claimant is reimbursed as well. And if there’s an issue, the claim is bounced back to the employee first, giving them a chance to correct it. A subtle reminder from the app about what is and isn’t acceptable to claim expenses for can also stop well-meaning but unaware employees from accidentally straying beyond the bounds of the company’s travel and expense policies, as well as saving time for accounting teams who would otherwise be tasked with investigating iffy amounts.
Misleading or missing expense claims can also lead to some awkward questions – and a fair bit of anxiety – if HMRC happens to drop by for an audit. Paper documents can easily get lost, and errors in spreadsheets can lead to inconsistencies that may catch the eye of the auditor. But if everything is captured automatically and located on a single platform, such inconsistencies can be more readily dealt with.
Having all your expense data in one place helps teams make more informed budgeting choices, and no one need scramble through mountains of paper receipts or invoices, or trawl through endless spreadsheet cells, to tally up a certain department’s expense claims. Instead, employees can focus on what they do best: making strategic decisions that will benefit the business. So, rather than simply fearing that robots will take our jobs, perhaps it’s time we also acknowledged how they might improve them.
SAP Concur provides intelligent solutions to help your business automate admin tasks such as expenses and invoices – giving you and your business the time and energy to focus on growing, innovating and making fruitful connections. To find out more about how SAP Concur can help you combine efficiency with intelligence, go to concur.co.uk