05 Jan 2021 17:29 IST

Are human auditors going away?

With the rise of AI, employees who can work effectively at the human-machine interface will be in demand

While humanity fights the war against the pandemic, another battle is happening worldwide that will ring in the new. At the heart of the change is Artificial Intelligence (AI), an omnibus term which I use for a thinking machine. It can read and grasp reams of data; it can determine patterns and spot outliers. Unlike automation, it learns from mistakes, and like human beings, with more practice, it becomes better. The rise of AI is now set to transform accounting, and its first cousin — auditing.

Sampling is a dying art

For several decades, auditors swore by sampling in a bid to tackle extensive data. They started with guesstimates, and then moved to both stratified and random sampling. Today, automation and AI helps you audit the entire population, and not just the sample. AI has made ‘reasonable assurance’ irrelevant because you can analyse the whole ledger, identify anomalies based on risk, and lay down materiality limits. Yes, sampling is now a very pre-2020 tool.

Inventory goes the RFID way

Thanks to Radio Frequency Identification ( RFID), companies can alter the way they cost products, value inventory, and verify stock. Costs associated with each inventory item can now be stored, so accountants can shift from the assumptive LIFO and FIFO to the more appropriate ‘specific identification’ of cost flows. AI helps pin down anomalies such ‘warehouse-size vs inventory size’ at a point in time, and ‘inventory vs sales data correlation.’

With RFID tags, physical verification is super fast. It can monitor the movement of assets and spot the current location in fixed assets verification. Where transceivers are inbuilt, tabulation is automatic. Drones are being used in stock count. Manual stock take, once an audit ritual, is headed to become a pre-2020 tool.

Risk assessment, receivables, payables

Risk assessment is a critical function in an audit. When auditors determine a risk strategy for their audit, they base it on the knowledge they have obtained in similar audits. AI now helps auditors with information on what fellow auditors have done in similar industry and similar client size!

Procurement is is often replete with paperwork and multiple file formats, which may not be compatible with each other. With AI, we can track multiple suppliers simultaneously, and with API unstructured data can be processed. AI helps identify which debtors are likely to default and to what extent this helps in provisioning. AI-based tools evaluate debtors, investments, and suppliers, and thus determine both the accounting treatment and provisioning. Machine learning algorithms improve estimates of bad debts, write off of stocks, and estimates of warranties.

RPA and ML algorithms

An area where AI is already fast impacting accounting is in automating entries. Robotic Process Automation (RPA) can simplify the process of data entry for accountants and auditors. RPA, coupled with NLP, can scan invoices, vouchers, and bank statements into accounting transactions within minutes and account for them under appropriate heads. The audit would then be on evaluating the rules rather than the transactions. Transaction compliance analyses are performed by AI to analyse 100 per cent of the transactions. This is used in audits to segregate between business expenses and personal expenses accounted.

Natural language processing

Natural Language Processing (NLP) makes life easy in an audit to analyse unstructured data such as agreements, contracts, emails, and PDF files. NLP tools help read the data in these documents and identify relevant information in an audit for testing. Contract review, revenue recognition, accounting treatment of lease Agreements, invoices, and purchase review, are some that will fall by the way side. JP Morgan sacked 1,000 lawyers once it had the ability of AI to read lease documents.

Closures

Income computation on investments, physical verification of investment, and balance confirmation, are all going to go the AI way. Payment processing, queuing up for seniority in payment based on cash flow, and provision for warranties, are areas that will receive help from AI. With AI, the numbers can come in faster, speeding up quarterly closing and annual audit. Consolidation becomes easier.

Human auditor

Will AI replace accountants and auditors? While it is not that everyone will go out of work, my sense is that 35 per cent of the bench strength will disappear, and maybe they will discover newer things to do. Those who stay back will help in the AI process, check that it works well, and continuously improve it. Auditors will end up talking with computers and AI for speedier and faster audits.

(The author is a CA, an author, teacher, and public speaker.)

Videos

Jargon Express: Standard Deduction

Jargon Express: Section 80C

What recruiters look for