The year ahead has a whole lot in store for Invosys as we prepare to launch several new products — along with our new website!

Back in December 2021, we spoke broadly about caller fraud and how fraudsters can gain and use individuals’ or business’ information illicitly.

We also briefly touched on the upcoming arrival of our brand-new caller fraud product, Invosys Call Shield: a one-of-a-kind firewall that detects fraudulent phone activity and prevents criminals from obtaining sensitive personal information or valuable business details. As the official launch date for this all-new solution is on the horizon, we’d like to share a bit more with you.

Whilst we’re not about to spill the beans on exactly how our new fraud solution works, we’re keen to offer more insight on the machine learning aspect of the product to enable our customers to understand how the cogs turn behind the scenes.

Unrivalled data analysis

Each call is made up of ‘metadata’, which are the call’s specific features. And every call’s metadata is unique, acting as its fingerprint or DNA.

Invosys’ Call Shield is able to extract and analyse the metadata of every single phone call, profiling it and establishing whether it has all the typical features of a fraudulent call or not. The features are then profiled to establish whether it meets the criteria for a suspicious call. So, what kind of metadata are so telling?


Are we dealing with an international caller? Is the call coming from somewhere local or overseas? Or, is it being routed via a different country — pretending to be local?

Number presentation

We can see whether the number presented to the call recipient is coming from the same number. If not, the number is being ‘spoofed’, and it’s likely there’s something suspicious going on.

Fraudsters commonly use public telephone boxes to protect their identity when committing caller fraud. However, we’ve got a list of all the public phone-box numbers across the UK, meaning we can immediately spot this kind of unusual activity.

What’s more, did you know there’s an exhaustive list of non-allocated numbers that nobody’s allowed to use? If a call is received from one of these numbers, we’ve got cause to believe it’s not legitimate.

Call source

The metadata allows us to see if a call is coming from a mobile, landline or a virtual source (such as Microsoft Teams). The source presented is compared against the call’s true origin and, if they don’t align, there’s reason to suspect fraud.

Volume of calls

If a single phone number appears to be making simultaneous calls, we know something isn’t right. Without electronic intervention, a standard mobile number can’t have more than one call active at the same time.

Taking advantage of AI

Over the years, Invosys has processed millions of call minutes for our clients. So, we’ve incorporated artificial intelligence and machine learning into the Call Shield solution so that it draws on its knowledge of thousands of past calls to learn and build an algorithm that identifies patterns. Now, the product can immediately identify the features of a fraudulent call and ensure it gets flagged.

We also have an extensive database of known telephone numbers associated with fraud. Thanks to our exclusive partnerships with industry bodies, this database is one of the most comprehensive available and is updated in real-time to ensure we’re working off the most up-to-date information. Invosys Call Shield is linked to this database and will automatically alert you that the call is likely to be fraudulent.

If a call is detected to be fraudulent, we can route the call to an agent, flag it or drop it. We’ll work with you to define a workflow around handling fraudulent calls and ensure you have proper processes in place.

Pushing the boundaries 

We’re constantly feeding our fraud product with as much call information as possible, applying machine learning to billions of data points with the analysis happening in near real-time. The Invosys Call Shield understands the DNA of a non-fraudulent call and knows what constitutes suspicious caller behaviour, and it’s been trained to automatically compare the features of calls it knows to be normal against all inbound calls.

The machine learning aspect of Invosys Call Shield is second to none — particularly as we also incorporate our customers’ feedback in what’s known as ‘supervised learning’. If our customer has received an illicit call, we’ll feed that call’s data to the product so that it can learn based on the patterns it automatically identifies, as well as from real people. The Invosys fraud solution draws on multiple sources, making it an incredibly intelligent and powerful piece of software.

While several caller fraud firewall solutions are available elsewhere, none comes close to Invosys Call Shield.

Built into our best-selling Number Manager platform, the product can be switched on at the click of a button. The software’s ease of integration and set-up speed mean it’s cost-effective and simple to get it up and running.

Keep an eye out for the official launch of the all-new Invosys Call Shield solution. In the meantime, you can get in touch with us at or give us a call at +44 (0)161 444 3333 if you have any questions.