Out With The Old, In With Kona
Corruption Risk is still top of mind with regulators. Digitizing the approach is proving cost-effective and reliable.
“Compliance monitoring over third-party payments is now fundamental to an effective corporate compliance program. Not since BrewRIGHT have I seen a platform that measurably predicts potentially improper payments, circumvention of controls and corrupt payments as effectively as Kona AI. Their platform is AI-driven, using proven models tested in the field and validated with both academia and compliance professionals. It is also easy-to-use, intuitive and rapidly deployable to support both investigations and proactive compliance monitoring programs.”
– Matt Galvin, Former Head of Compliance at Anheuser-Bush InBev, now a Research Fellow at Harvard Business School
In June 2021, the Biden Administration issued a briefing document establishing the fight against corruption as a core U.S. national security interest. Add to it the war in Ukraine, high inflation, supply chain issues and a depressed stock market and you have the perfect storm for fraud, corruption, waste and abuse.
For years, the Association of Certified Fraud Examiners (ACFE) Report to the Nations on Occupation Fraud suggests that companies lose up to 5% of their revenues annual to fraud and corruption, at an average lost of $1.783 million per case (Source). But this number continues to remain at 5% every year – why aren’t we showing improvement? The answer may lie in how we prevent and detect fraud, corruption and circumvention of controls. The time is now to rethink traditional rules-based tests and manual processes that generate too many false positives, and worse…false negatives.
The challenge with current compliance technology tools is they focus solely on the due diligence and pre-approval workflows during the vendor or third-party onboarding process. Leading tools like Navex or Process Unity among many others focus on pre-approval workflow, due diligence screening, vendor questionnaires, training and whistle blower case management among several other features supporting the third parties. But once the contract with the third party is signed, it’s never looked at again – unless it becomes in scope for an internal audit, or worse, subject to a whistleblower allegation that requires investigation. Fraud and corruption happen after the contract is signed with the third-party – where money is exchanged, invoices are created and cash is paid out. It just makes sense that legal and compliance professionals include payment and transaction monitoring into their compliance program.
In fact, the U.S. DOJ states in their June 2020 Guidance specifically calls out ongoing monitoring activities when it asks, “Does the company engage in risk management of third parties throughout the lifespan of the relationship, or primarily during the onboarding process?”
What companies need is an AI driven platform built for investigators, compliance and internal audit professionals that finds hidden money faster, with measurable improvements in your anti-fraud, anti-corruption and corporate compliance program in line with regulator expectations – particular the U.S. Department of Justice and the U.S. Security & Exchange Commission.
Kona AI brings years of investigative experience from the Big 4 firms, as well as leading academic institutions from MIT and Harvard Business School in the development of our proprietary library of behavioral algorithms and predictive models vetted from decades of cumulative industry experience. Nobody, and I mean nobody, can empower you to cost-effectively rip through procure-to-pay, order-to-cash or travel & entertainment as well as Kona AI.
Kona AI is not a software company. We are an AI-driven company that uses software to easily and intuitively deliver vetted and reliable predictive models that statistically measure risky payments via our Vendor 360 Module, risky sales via our Customer 360 Module and risky travel & entertainment expenses via our Employee 360 Module.
Forensic Data Analytics Maturity
In thinking about how investigations and the use of data analytics and compliance monitoring are done in your organization, take a look at this maturity model and see where you are spending most of your time.
It’s unlikely that you are still relying on “the obsolete way” as most companies have given up on sample selections when it comes to fraud and corruption. Sampling is ideal when testing a process, as in a financial audit, when you are testing a mathematical calculations to see if the process works effectively. If the math formula is wrong, taking a sample and manually recalculating the formula will identify any issues or errors in the process. Statistical sampling in an audit is also effective when outcomes are indeed random. But fraud is not random! Fraud is deceptive and the perpetrator is intending to hide their activities. Simply put, random sampling is not effective when preventing and detecting fraud. Indeed, you have to be very lucky to stumble across it with a sampling approach.
The next ways is “the old way”. While still currently in use by most Big 4 consulting firms and the boutiques, the old way which has been around since 2010 when I started doing it was manually extracting and reconstructing the financials so that 100% of the data could be analyzed. This requires a tremendous amount of data manipulation work (and consulting hours) to prepare the data for analysis and manually scripting (programming) the testing algorithms. Budgets typically reflect 80% of the work being data extraction, transformation and loading (i.e., Data Prep) into an analysis platform like Tableau, Qlik, Spotfire, Domo, PowerBI. This often leaves little time and budget for the actual value-added part which is the analysis of the results or vetting out false positives using predictive modeling algorithms.
Finally, there is the Kona AI way. Our approach is AI-Driven where using scripting and unified data models that connect to leading ERP systems such as SAP and Oracle in an automated fashion. What use to take 4 – 6 weeks to extract, transform and load data from complex ERP systems now takes literally 4 -6 days! The behavioral algorithms are pre-mapped and generated upon loading saving hundreds of hours. This allows for significantly more value added time running the analysis and refining results, which in-turn, improve the predictive models that “find-more-like-this.” For those familiar with eDiscovery matters and email reviews, think of eDiscovery tools like Relativity, but for transactional data, not emails.