Internal audit teams across industries (i.e. retail, banking, oil & gas and manufacturing) need appropriate expertise and tools in their journey to introduce Artificial Intelligent (AI) and advanced analytics into their practice.
Most audit departments are overwhelmed with the transaction volumes and number of documents or contracts they need to check. Reality is that small samples of these transactions and documents are being checked, sometimes as small as less than 1-2%. Advanced analytics tools can enable the check of 100% of the transactions or documents that need to be audited. The auditor can focus on handling exceptions, rather than going trough random transactions or documents.
The Passpadu Audit Analytics is an audit management system which incorporates artificial intelligence and machine learning tools to discover and analyse patterns and identify anomalies. It is built on a modular architecture, which includes the following modules:
- Standard Contracts
- Non Standard Contract
- Sharia Research
The system analyses large volumes of data and identifies misstatements or fraud which would otherwise had been remained unnoticed. It does not only save effort for auditors, but most importantly, it increases the efficiency and effectiveness of the audit process. Auditors can now conduct more frequent audits in high-risk areas and enhance the reliability of audit results. The platform becomes a central repository of all the audit related activities such as file management, findings, creation of reports etc and introduces:
- Knowledge database for assistance in the reassessment of past findings
- Enhanced audit input by transforming hard documents into digital forms, ready to be processed and analysing images and video
Overall the system plays a central role in improving the performance of internal audits and gradually enables the implementation of continuous auditing.
The benefits of audit analytics are immediate. It increases operational efficiency by reducing costs and helps quickly identify potential fraud and anomalies. Auditors are equipped to assess potential risks, identify operational inefficiency, and provide insights, overall they can provide a higher level of assurance.
Auditors of any degree of familiarity with advanced statistical techniques and data analytics tools can use audit analytics. The platform creates data models and visualisation of large volumes of data in an intuitive way that hides the complexity. It empowers auditors even if they have limited data analysis knowledge to make the best use of audit analytics.
Enterprises produce thousand of transactions every day and the Internal Auditors have a limited ability to audit even 1% of these transactions. Even if they have insights about certain business areas, which they consider risky, identifying the right data sample to audit is impossible.
The Audit module can handle million of transactions associated with a small number of variables describing each transaction, such as time, location, currency etc. Depending on the type of issue being investigated, the appropriate model is being used.
The output of each algorithm is weighted and the higher risk areas are being identified. Depending on the volume of the risky transactions, a further analysis can narrow down the outliers sample before it is being handed over to the Internal Auditors.
The workflow module can assist in the investigation process, where the auditors analyse each suspicious case in order to verify whether anomalies do happen.
Standard Contracts module
Certain businesses, such as retail banking, produce every day a large volume of contracts based on standard forms (mortgage, car loans etc). It is not uncommon for errors to happen, such as missing or wrong dates, data mismatch (customer name, product name, interest rate), missing signatures or even missing files.
The contracts module checks the documents against the pre agreed rules, i.e. dates, signatures etc. Once the processing is over, the outcome of the module checks are written in a pre-agreed location in the form of a spreadsheet. We expect the quality of the results to depend on the quality of the input data (images) provided.
Non Standard Contracts module
All businesses, from corporate banking to Oil & Gas and manufacturing, produce contracts, which even if they have a few terms in common, include a number of additional terms and clauses. If dealing with government, for example, even the largest corporates are obliged to use the contract templates of the government entities.
There are occasions where these contacts, each comprising tens of pages, need to be reviewed on certain occasions. For example:
- what are the early penalties we have to pay if we close a facility?
- which clauses need to be updated if GDPR is introduced?
The Non Standard Contracts module uses natural language processing (NLP) technology in order to extract information from any document in any language and any format (eg PDF) to extract entities and terms, organise information and present it into a tabular format (eg Excel). The whole process is automatic and there is no need for experienced legal resources to spend a lot of time reviewing many pages.
Sharia Research module
Sharia Research departments of Islamic Banks have to go through thousands of previously issued Sharia Opinions and cross check against Sharia Standards such as AAOIFI and the four prominent Schools of Thought (Hanafi, Maliki, Shafei and Hanbali) prior to issuing an accurate Sharia opinion.
The Sharia Research module uses natural language processing (NLP) technology in order to turn Sharia Opinions into insight, using a unique linguistic text analytics platform to accelerate Sharia Research department ability to make sense of information in any language – Arabic or English.