How Artificial Intelligence (AI) Will Improve The Audit Process
How artificial intelligence (AI) will improve the audit process was the subject of a recent article in the Journal of Accountancy. Our team member, Alex Frank, CPA, offers this summary of the article, and what it means for the future of our business and the way we will be able to use AI to positively impact our clients:
Summary of Article
- AI can and will remove the need for audit sampling
- Using Optical character recognition (OCR) the AI can quickly scan large documents (such as contracts, leases, debt agreements) and extract relevant terms or phrases OR the AI can flag an area of the contract that needs further examination
- At first the AI will make many mistakes that the auditor will have to correct manually, but the AI will learn from those mistakes and make them less and less until there are little to no errors made, thereby improving accuracy and efficiency
- The AI can also ask the auditor “what should I look for” and the auditor can set specific parameters to look for
- AI can sort transactions based on the transactions risk based on knowledge obtained from client information and firm risk analysis
- Allows the auditor to focus professional skepticism on a specific area as opposed to trying to apply it to all areas at an equal level
- The AI can increase interest in auditing samples
- Sampling will have a reason for why transactions are selected as opposed to haphazard or random sampling that has no real risk included in the sampling process.
- The AI will select high risk transactions as a sample and will help teach auditors what is considered high risk and allow them to focus more time to apply judgement to analyze the data in front of them
- As the AI continues to learn, it will begin to identify patterns and therefore be able to flag any anomalies in the data to help focus the auditor’s attention to riskier areas
- AI will eventually lead to validation of 100% of client transactions in an almost real time basis
- Auditors will be able to implement their own AI into a client system
- The AI will analyze every transaction and journal entry and determine if processes are followed and will flag any transactions made erroneously or incorrectly or transactions that are risky in nature.
- The AI will flag these transactions to alert the auditor as to what transactions need to be review as the transaction happens
- This will create a shift from auditing as of the balance sheet date (historical data) to a form of continuous assurance that will take place as the transactions occur
How AI can apply to BTCPA
- BTCPA has begun to implement some forms of technology (data analytics) in several of our audit processes.
- This is a good first step into the world of automating the audit process but is several steps away from AI
- AI has potential to help optimize some of our more uniform clients
- Single audits (nonprofits and governments)
- Search for compliance across entire populations of data
- Delaware Charter Schools all utilize the same chart of accounts and all data is maintained in the same accounting system
- Pennsylvania School Districts
- they all share a common chart of accounts, so AI could analyze them all the same way
- Small government clients
- Share a common chart of accounts and all have relatively similar operational costs and revenues
- Challenges to navigate:
- Finding an AI technology that we can use and is easier to use from a staff perspective
- Or learning how to create one
- Finding a best practice for how to implement the technology
- Training the AI to fit our client base (time consuming)
- How would the implementation be viewed from a peer review standpoint?
- Finding an AI technology that we can use and is easier to use from a staff perspective
- Single audits (nonprofits and governments)
What the Future looks like – and Why we remain committed to learning more
- Allows for 100% transaction review of clients
- Can eliminate the “routine-ness” of the annual audit process and allow for increased risk analysis
- The annual audit approach would be able to tailor an approach on a per client basis and create a more meaningful analysis of the client data. Therefore, giving the client a more meaningful report more tailored to their activity during the year.
- AI will eliminate the need to scan through checks or GL detail for samples or transaction details.
- Instead AI will be able to tell you exactly what transactions we are testing or need to be examined more closely without the monotony.
- The samples will have MEANING now
- No more haphazard or random sampling
- Since AI will automate the sampling process we can spend more time examining data and applying professional skepticism to the most important areas of the engagement.
- AI will be able to compile client data into a uniform “structure” which will allow us to easily drill down into data and easily determine:
- Transaction amounts
- Exact accounts effected by a JE with out having to scan through a GL
- Quickly determine if the client has a lease we do not have documented
- Determine if the client has incorrectly coded a revenue or expense
- Determine where issues exist that would cause the clients equity to not agree to the PY FS
- No more comparing the CY with the PY and figuring out the difference manually. The AI would be able to compare information it already has on the client, find any discrepancies on its own and notify the auditor what needs to be fixed
- If AI is implemented at the client and operates in the background, we would essentially eliminate “traditional audits”
- Audits would not be done after year end examining “old” data
- Instead, audits would be conducted by the AI on a continuous basis
- Whenever there is a transaction that the AI identifies as an issue, it will notify the auditors and they would be able to test the transactions as they occur
- With the impending implementation of Blockchain, AI will become a reality sooner rather than later.
- AI is already prevalent in our everyday live whether it be Siri, Alexa, or a YouTube recommended video, AI is ever growing and evolving.
- AI will seem scary at first, but once implemented the AI will streamline almost every step of the audit process
- The future of Auditing looks different than anything we have ever seen, but it also looks much more efficient, and will allow for a more tailored audit experience for our clients.
- Transaction amounts
Barbacane, Thornton & Company is a highly regarded, regional certified public accounting and consulting firm specializing in auditing and tax services for government agencies and nonprofits.