Automating Audit
From Clay Styli to Autonomous Digital Ledger Recording
Over the decades we have witnessed a drastic shift in the way the accounting profession records, processes and reports a transaction. Some research indicates that the book keeping in 2600 BCE used to happen with the help of styli on small slabs of clay. Now the digitalized systems not only processes and reports the transaction recorded by the accountant but it can also record the transaction by itself and eliminates the entire manual intervention.
In this everchanging world the auditor’s had learnt a lot about system automations and how to perform an effective and efficient audit in an automated environment. But it is not the end of great technological learnings, rather it is the beginning for the next gen learning where an auditor needs to learn how to automate the audit process itself.
Core Dividends of Audit Automation:
The automation can provide greater efficiency, reliability, accuracy, flexibility, timeliness of information, expand capacity, boost quality, enable greater audit coverage and can even provide cost advantage over the long run. The modern machines not only match humans but really do even more than a human. The talented workforce can be directed towards more complex and judgemental areas and excel in their core activity. Hence as an auditor why don’t we take the advantage of automating the audit itself?
Artificial Intelligence & Robotic Process Automation: The Technological Enablers
A technology which enables the machines to perform tasks of repetitive nature which are generally performed with human intelligence (i.e., in simple words the computing system is enabled with human intelligence). The best examples which we are most familiar with are search recommendations by Google & Amazon, self-driving cars, etc. This technology is so powerful that it enables the systems to even compose music, analyse human voice, capable of talking back (Siri, Alexa, Google Voice Assistant), identify the images and take informed decisions which are rule based. Yes, the AI can make decisions on its own & even capable of expressing opinion on financial information.
• Robotic Process Automation (RPA)
• Cloud Computing
• Natural Language Generation (NLG)
• Blockchain
• Intelligent Content Recognition (ICR)
• Intelligent Chatbots (Virtual Assistants)
• Predictive Modelling
One needs to have a good understanding about these new gen tools as they may become equally important as a technical standard to perform audits.
Practical Audit Use-Cases & CARO Automation
Following are eight illustrative audit workflows ripe for full automation:
I. Automated Newsfeeds & Market Intelligence:
Use automated newsfeeds to pull information about entity (market data, regulatory filings, financial and non-financial news articles).
II. Algorithmic Materiality Determination:
Materiality can be determined by pulling out the numbers from last audited financials and applying the percentages as per the firm’s methodologies.
III. Rule-Based Compliance Checking under CARO 2020:
The auditor can use this technology to automate compliance checking processes which are rule based. Illustrative areas under Companies (Auditor’s Report) Order (CARO):
- Title Deeds of Immovable Properties (Clause 3(i)(c)): If a company holds a large portfolio of immovable properties, it requires an auditor to manually examine each title deed. With RPA, optical verification is executed in fractions of a second: the system parses all scanned title deeds, compares entity ownership with land records, and outputs exception registers.
- Undisputed Statutory Dues Deposited (Clause 3(vii)(a)): RPA reads monthly and state-wise returns and automatically maps them against electronic tax challans w.r.t. TDS, GST, PF, PT, ESI, etc. The script detects payment delays and pinpoints statutory dues overdue for more than six months from the date they became payable with 100% accuracy.
IV. Complete Journal Entry Testing & Visual Analytics:
It can read through each and every entry passed throughout the period under consideration by the clients and provide most valuable pictorial representations of huge transaction volumes. This adds value not only to auditors in identifying audit risks, but also provides valuable inputs to clients regarding business trends.
V. 100% Population Testing & Sampling Risk Elimination:
Examines the full population of data and identifies anomalies and patterns useful in risk identification. Checking the full population eliminates sampling risk, leading to superior audit quality and reducing overall audit risk to the maximum extent possible.
VI. Independent Calculations & Auditor Point Estimates:
Runs independent re-computations for depreciation, amortisation, interest schedules, etc., delivering an objective auditor point estimate for audit evaluations.
VII. Drone-Assisted Physical Stock Verification:
Unmanned Aerial Vehicles (drones) deployed across large warehouses, yards, and remote project sites to execute high-speed volumetric counts and physical inventory verification.
VIII. Continuous Real-Time Auditing:
Auditing transformed into an instantaneous, real-time activity powered by AI tools that continuously analyse, test, and flag anomalies the moment a transaction occurs.
Factors, Motivations & Practical Limitations of Automation
Factors Affecting Automation:
- Identifying the areas susceptible for automation
- Technical feasibility
- Cost of development, implementation and sustainability
- Adoptive work culture
- Time taking process
- Government Regulations
- Confidentiality considerations
- Benefit from automation realized over the long run
Motivations for Audit Automation:
Modernisation at client organisations leaves no option for auditors but to emerge as tech-savvy firms providing enhanced real-time assurance. Eliminating routine tasks opens new revenue streams.
Automated Data Ingestion: The tedious initial task of gathering unstructured data from unstructured sources is streamlined by predetermining required schedules and auto-triggering scheduled requests for recurring audits.
Operational Limitations & Security Roadblocks:
Cost, adoptability, and confidentiality of information obtained remain key limiting factors. Real-time automated testing requires continuous access to sensitive client servers, which clients are often unwilling to link to audit firm platforms.
Will AI Replace Auditors? Research Insights, Co-bots & The Luddite Fallacy
Historical Context: Will the Luddite Fallacy Come True This Time?
A Luddite describes people who oppose new technology. In the 19th century, English textile workers were displaced by automated power looms producing cheaper cloth. Unemployed workers launched campaigns to destroy machines. While displaced workers suffered in the short term, over time society benefitted immensely from cheaper clothes, lower prices, and massive new employment opportunities across expanding economic sectors.
Global Empirical Research on Automation & Financial Job Losses:
From “Robots” to “Co-Bots” (Collaborative Bots):
New age technologies are best viewed as “co-bots” rather than “robots”; they will not steal our jobs, but will work alongside us as co-employees. Historically, only physical muscle power was mechanised; today, the human brain itself is being mechanised.
Conclusion: Professional Scepticism Remains Irreplaceable
The automated world equips the new generation auditor with the capacity to test 100% of the data population. This elevates stakeholder expectations from traditional reasonable assurance to near-absolute assurance.
While routine tasks are readily automated, predetermined rule-based systems can be bypassed by fraudsters employing sophisticated techniques. Therefore, human professional scepticism, deep experience, and contextual judgement remain irreplaceable, particularly as novel fraud methodologies evolve even faster in an automated environment.