Tough Talk – with Appriss Retail
Why did you change your name to Appriss Retail when Sysrepublic was so well known?
Sysrepublic was acquired by a US-based company named Appriss. Appriss is a diversified data and analytics company that has been solving a variety of problems in public safety, healthcare, and retail businesses, all under a core mission of utilising knowledge for good. Appriss Retail is the unit that combines the capabilities of Sysrepublic and a US-based retail analytics company called The Retail Equation.
Appriss saw that the application of advanced analytical techniques including artificial intelligence (AI) can help retailers unlock valuable insights in their own data to improve their financial and operational performance, enhance customer service, and reduce loss.
Since the acquisition of the two companies, Appriss has invested significantly in R&D and launched Secure 5™, the most advanced retail intelligence and exception based reporting platform that incorporates advanced, AI-based analytical tools.
Why would a retail company that does not use EBR analytics choose to implement a system? and
Why would a retail company with EBR already wish to upgrade to an EBR with artificial intelligence (AI) capabilities?
EBR is all about identifying and analysing anomalies within a retailer’s data, processes, or people’s behaviours. Retailers with a handful of stores and very limited operations will often rely on store personnel or spreadsheets to look for a simplistic set of anomaly-indicators. They can fix the obvious problems and write off all the other losses –fraudulent returns, theft, short shipments, etc. –as a cost of doing business. However, this manual process does not scale as the retailer grows in store count, in geographical dispersion, and in the complexity of product mix and supplier mix. Retailers of this size usually add in exception based reporting (EBR) analytics. It works to monitor for exceptions, and an EBR system might be what a mid-size retailer needs to begin with.
Traditional EBR deployments resemble reporting systems – systems that may come pre-packaged with certain reports and let the retailer develop others based on their experience. While they are a substantial improvement over spreadsheets and manual oversight, they are still quite limited. They rely almost entirely on users’ experience and their ability to specify exactly what they are looking for. As a business gets more complex, AI-driven models are far more effective. The models complement human experience and are able to analyse massive quantities of data and identify complex patterns. Large retailers often have more situations to be solved. By sheer number of storefronts and employees, they are more vulnerable to loss. When you add to that the level of technology that fraudsters can easily employ—online bar code generation and receipt forgery tools are just two of them—and the amount of work they willingly devote to their actions, the need for the ‘best’ solution becomes clear. That is when AI outshines over basic analytics. It helps solve complex problems with a reasonably sized LP department.
When AI is added to analytics, it streamlines the initial phases of analysis and investigation, and it makes the outcomes of in-person investigation or store audits more productive. The process is iterative, and, thanks to machine learning and feedback loops, the more it is used the more accurate it becomes in identifying anomalies within your company.
Is there a material difference in the results between EBR analytics with and without AI?
Yes, in three ways. First, AI can uncover more fraud or procedural problems in a fraction of the time that EBR or conventional analytics require. Second, it can raise all users’ performance. Consider that an analyst or investigator with years of experience using conventional methods will far outperform a neophyte because the pros know all the ins and outs, the revealing techniques. AI systems can be built to include those techniques along with their massive computational power. The result is that the system can elevate the performance of all investigators (new and seasoned) to maximise their productivity. Third, through machine learning, the system uses the results of each investigation to improve analytics that trigger future investigations.
AI is complex. Will I have to hire IT people to write queries or interpret results?
If you are creating a home-grown system, then the answer is probably ‘yes’ because in this case AI would be a tool kit that your developers would use. However, if you are buying a solution designed to address the unique nature of retail loss, then the opposite is true. A retail-focused, commercial AI solution for loss prevention will remove labour hours from query writing, analysis, and reporting. Given the promise of AI, many vendors have begun to claim that their tools incorporate AI. Retailers would be prudent to perform their due diligence to determine whether vendor claims are just traditional EBR capabilities being marketed as AI versus true AI with its advanced capabilities for high efficiency.
I have situations beyond my store ePOS, but I’m too busy to address them?
AI models can be used to address issues beyond the ePOS, like inventory and ecommerce situations. It can find problems and point investigators toward the most productive outcomes. It can save up to 50 percent of the investigative time you might otherwise spend and produce more accurate results.
Where is EBR headed? Is AI here to stay?
Today, most Loss Prevention departments use traditional exception based reporting (EBR) tools to monitor their employees. While the concept of EBR is going to remain because its focus is on detecting the unusual, outlier behaviour using data, the technology behind leading-edge EBR systems will, over time, be supplemented or replaced completely by AI. Appriss Retail sees AI as the future.
EBR tools used by retailers are similar to the tools used by the credit card industry in the 1980s to detect payment fraud. In the early 1990s, the credit card industry mostly abandoned the exception based reporting approach in favour of methods of detection based on predictive modelling and machine learning. A similar trend was seen with healthcare fraud about 15 years ago, when the industry’s tools converted from rules and queries to predictive modelling. Our estimate is that retail loss prevention is primed to rapidly evolve in a manner comparable to credit card fraud detection.
Can AI really improve analytics–or is it just another buzz word.
Without a doubt, AI can improve analytics, especially loss prevention analytics.
AI applies a complex mathematical model to transaction or other data and learns, using thousands of variables, the most useful elements in an employee’s transactions for identifying cases that are most likely to result in an action. With an AI model, the top ranked exceptions will have a much higher success rate than those that are generated from traditional EBR. Over time, an approach with a higher success rate will make loss prevention departments far more efficient and will decrease employee fraud rates.
As a software user, though, you need to be sure that your vendor is really selling artificial intelligence, not just calling old technology by a new name.
More sales are shifting to online. Can fraudsters hide there?
Fraudsters know how to exploit system weaknesses, but fortunately EBR can be applied to online transactions. You do not want to lose revenue and offend good customers by rejecting odd-looking but legitimate purchases. AI models can review much more data and offer results faster than conventional EBR. For example, AI can help your stores handle omnichannel shipping and ecommerce returns quickly and accurately in real time. Through machine learning, the system can adapt to new fraud techniques, making it more difficult to conduct fraud via any channel. It even moves beyond fraud-related losses to help you identify procedural issues, such as problems in the fulfilment process.