AI is already delivering savings in the fight against fraud

Every business today is facing a host of new threats in the communications world. Telecoms service providers are no different, and the growing challenge that is VoIP and toll fraud is now estimated to cost £25.5bn globally; more than double that of credit card fraud, writes Carl Boraman, the director of strategic alliances and Tollring.

The traditional measures that service providers use to prevent this type of fraud are spend caps and credit limits. However, the methods of attack deployed by these criminals are evolving and changing. More frequent, small, pick-pocketing style attacks have ensured that fraud remains a lucrative endeavour for those responsible. The end result is a communications industry that is perpetually on the defensive, limited to reactive approaches where millions are spent after attacks have already taken place.

These costs are large and unpredictable, and pose significant threat to a service provider’s bottom line.  However, it’s the impact on their brand and reputation which ultimately impacts customer trust and loyalty. Fraud sours relationships, takes time to resolve and damages trust, resulting in poor customer experience and increased customer churn.  For the telecoms industry, where margins are getting thinner and customer retention is increasing in importance, this is particularly significant.

This combination of increasing costs and customer churn seems to be concentrating minds on tackling the problem.

Analytics, combined with machine learning and artificial intelligence (AI), now offers a viable solution that allows service providers to proactively combat fraud before it occurs. This allows a fraud and credit management service to predict trends, learn patterns and make ongoing adjustments, delivering considerable benefits.

For example, AI engines can capture customer interaction data which can then be transformed into valuable insights into customer patterns, trends and behaviours. By combining machine-learned data from AI with traditional communications data, you start to create a real opportunity to offer exceptional customer experience, where you truly know your customers and can meet their needs regardless of method, location or time. This can then be used in the fraud prevention arena by determining whether the activity within a business phone system is genuine without simply relying on caps and limits.

Tollring’s AI-equipped fraud and credit management solution, iCS Protect, has already saved its service provider resellers and customers (across all platforms and partners) millions in fraud damages.  In the past six months alone, Tollring’s partners and customers saved an average of £800,000 per month – this figure is based on the widely published average cost of damage of £10,000 per fraud attack (source: NEC and other companies on the internet).

Tollring’s findings of 96% of fraud incidents across platforms in the past year captured by intelligent analytics suggests that relying on credit capping is vastly insufficient in the fight against fraud. If such a high percentage of fraud attacks continue to be captured by intelligent analytics, we believe this will cause a huge shift in the industry’s approach to fraud protection.

This type of fraud management, driven by intelligent analytics, can drive a culture of shared responsibility, encouraging the whole supply chain to think more carefully about the vulnerabilities in their systems. Operators and service providers will need to co-operate, working together to adopt dynamic, real-time anti-fraud tools so they can detect fraud as it happens and act appropriately.

As an example, BT has introduced a fraud management solution using analytics from Tollring in a bid to prevent criminal activity before it takes place.

Every call that crosses BT’s network now passes through rigorous checks. Machine-learning functionality with intelligent real-time analytics enables BT to monitor behavioural patterns and call trends. The company can implement rules, trigger alerts and automatically block calls to protect against illegal usage and ‘bill shock’. Credit management becomes the final gatekeeper in the process, constraining spend to manage risk.

As a reseller of BT’s services, Daisy Wholesale’s Alex Mawson says, “We can now assign over 100 rule thresholds against a customer site, in addition to adding simple credit locking values. Email alerts are sent when these levels are breached then usage can be restricted. Alerts based on unusual calling patterns allow early analysis to identify fraud and misuse. The system can also automatically restrict services if thresholds are breached, providing considerable peace of mind.”

Customers are now looking to their providers for fraud protection, to ensure their phones and data remain secure and confidential. Reputational damage arising from security breaches far outweighs potential financial losses to both customers and their providers, so as organisations’ objectives align there is a new shift towards shared responsibility for combatting fraud.

We must ensure that every link in the telecoms chain is protected against attack. AI with intelligent analytics is the only way to stop fraud before it happens. By embedding intelligence and analytics tools into the network, providers and their customers can achieve a whole new level of security.