How a tailored approach can fight fraud at scale
Avoid the quick fix
In a busy corporate world, it’s easy to see the attraction of technological quick fixes. But these ‘point’ solutions tend to only address one particular problem, quickly become unsupported and can often be difficult to integrate with other business processes. Many remain as silo solutions throughout their lifespan.
According to a Forrester study, 80% of retailers are still operating siloed fraud solutions due to integration problems.
It’s far better to think strategically and invest in technology that can be used to solve more than one problem. These are some of the key benefits:
- You can build a repeatable skillset around that technology
- Retaining those skills means you are less likely to build tomorrow’s legacy
- You can remove the silos and share data, processes and interfaces across the business.
- You can start small but grow to enterprise-wide.
- You’ll get a better return on your investment.
So what technology should you consider? Automation can effectively handle the volume and frequency of potentially fraudulent transactions hitting your business (see our blog ‘the fraudsters have stolen a march’ for more on this). The big question is, which type to use?
There’s a lot of focus on and investment in Robotic Process Automation (RPA) right now. It’s a system that’s good for things like production lines, but it’s task-centric and uses the ‘if this, then that’ decision tree approach which can’t cope with the complexity of multi-faceted decision-making. A more human and sophisticated option is a better way to go.
Or human centric?
The human brain solves problems through intuition and inference. It subconsciously uses a decision map that can come from any direction and take you in any direction.
Apply this decision map approach to a bank’s systems, for example, and you have a fraud prevention program that’s able to apply the knowledge that you’re going on holiday to spot any suspicious activity that happens in the branch, over the telephone, online or a through an app. Similarly, it could flag up transactions where elderly customers are withdrawing their life savings to pay a fraudster – stopping a withdrawal that appears above board but is in fact happening as part of a fraud.
TAILOR TO YOUR BUSINESS
The Rainbird modelling process gives you the ability to build enterprise-scale fraud detection and identification systems that are based on your own expert knowledge of your business and the threats you face. Our ‘expert-down’ (vs ‘data-up’) approach makes it easier and quicker to react to newly identified or suspected trends by having your subject matter experts develop the solution themselves.
In a recent case study, we took the logic from over 50 of our client’s fraud cases, interviewed their best performing subject matter experts and tailored a solution to their specific requirements.
CREATE YOUR DECISION MAP
When building decision-making models in Rainbird, the author does not need to chain together linear pieces of logic to form a tree, but instead creates a map. This is a holistic model of knowledge, and will include numerous small nuggets of logic pinned to an ontology. Where it is necessary and appropriate, Rainbird will consult with humans along the way to get more data. The final judgement made by Rainbird based on the ontology, the logic and the data will always include a certainty score and, to meet compliance obligations, a full audit trail to show how that decision was reached.
Conclusion – invest in cognitive
It pays to invest in cognitive platforms like Rainbird for fraud prevention because:
- They’re a strategic, long-term solution
- They can be used to automate decision-making in a scalable way
- Decision-making is more nuanced sophisticated
- They can scale your own in-depth knowledge of your business and fraud threats
Ultimately, tools like Rainbird make it simple for your business to create a fraud prevention program that gives you the control and capabilities to stay ahead of the fraudsters.