Data Futures: Machine Learning for Profit Recovery
Data Futures: How automation and machine learning are changing the world of profit recovery
Machine learning and automation are taking over the world… not in a Terminator / Skynet sort of way (yet) but their effects are reaching every corner of business and society. Profit recovery is no exception.
Profit recovery is all about process; the process of receiving and validation of data, the process of forensically auditing that data to find those duplicate payments or that unfunded promotional activity and finally the process of recovery. All of these functions follow a set of rules in order to apply labels, for example, “claim” or “no claim”, “duplicate” or “not a duplicate”. This, at the core of it, is what automation and machine learning were built for.
Already automation has had huge benefits for clients; in grocery retail for instance, after the Groceries Code Adjudicator was given further powers to financially penalise retailers in 2015 there was more urgency than ever to deal with profit recovery matters sooner. Automation has provided the means to be able to extract, transform and load data into audit systems for analysis much faster and more effectively than ever before, allowing profit recovery firms to get as close to ‘real-time’ or ‘in-year’ auditing as possible, ensuring payment queries are raised and dealt with well within the regulatory timeframes, and making ‘in-year’ and pre-event auditing achievable.
The next step towards auditing ‘real-time’ is being able to get through this mass of data as efficiently as possible, and machine learning helps APL do just that. By passing through historical data with known outcomes we can analyse new data to derive a predicted result. That may be searching and highlighting documents of particular importance, spotting a duplicate payment to a supplier or matching those documents to promotional activity that it relates to, automatically.
However, the same technology that could help profit recovery firms expand their client portfolios with minimal extra staffing requirements could also spell danger. More and more applications are now coming with machine learning technologies integrated, and it should be expected that any new software purchased by retailers would have smart technologies built in. This could provide huge potential for profit recovery firms as retail clients use machine learning to create more targeted promotional campaigns, driving more volume and therefore more value, with more potential for payment errors.
It’s easy to envisage that with the advancement of the machine learning field continuing at the pace it has been that a traditional profit recovery offering may soon rendered obsolete. Part of Audit Partnership’s approach is to bridge this gap, constantly developing our applications and techniques to provide the necessary intelligence for our customers to prevent loss going forward.
The future has so much potential for profit recovery, but as technology continues to evolve, so must the customer offer provided by recovery audit firms. As we move away from purely a historical offer, we need to harness the power of machine learning to focus on the protection, rather than the recovery of profit.