Clearing the flow of calls without magic and SMS

Companies invest in attracting calls to potential customers, while other companies want to make money on it. And not always honestly. For example, creating a large flow of calls that will never lead to a deal. Our Virtual PBX is struggling with unscrupulous call providers.

Not all calls are equally useful.

Unwanted calls to the company clog the channel and take away the time and effort of employees. We distinguish four types of such harmful messages:

1. Telephone hooliganism
2. Fraud
3. Spam
4. Frode

The first type - the least dangerous for the company. Firstly, it is not widely used, and secondly, it is elementarily suppressed with the help of the Virtual PBX. Usually hooligans choose a victim and order calls to her number through callback services. It is enough to add the IP address of such a joker to the blacklist and the problem is solved.

The second kind, fraud, can do much more harm. By breaking into the company's SIP lines, violators can dramatically deplete the balance by calling premium numbers, often international ones.

In most cases, this is due to the carelessness of the SIP-telephony subscribers themselves. Almost any cloud PBX on the market (and ours, of course, not an exception) has the ability to ban calls to long-distance, international and premium destinations. And so we move on to the categories of calls that are much more difficult to fight.

Call blocking by voice (and not only)

With voice spam cope much more difficult. Sometimes it is associated with fraud (trying to force the listener with a provocative message to call back to the premium number), but this phenomenon, fortunately, is not of a mass nature.

And yet there is a way to fight. If auto-dialing is done using the complex, the auto-dialing system + call-center operator, when a company makes cold calls and offers its products, not all of our customers equate such activity with spam. Some even find something useful in such commercial offers and make purchases.

If an auto-call is made by a robot, then it’s difficult to cope with its consequences: calls are clogged with channels, seriously distracting employees from work, and distorting call-tracking.

Clearing incoming calls from robot spam is an interesting task that big data helps us to do. The system simply notes calls that in its opinion are spam.

What is characteristic of a robot? He makes a lot of calls at the same time, he can go through the numbers when he calls, and the clients send the numbers to the black list. He dictates the same message, and even if it is personalized (depending on who the robot is addressing), then with the same voice.

Vote! This is what among other things makes it possible to unambiguously identify the robot. In addition, he does not listen to the other side and interrupts her (have you ever heard a conversation between a spam robot and an IVR? - this is very funny).

After checking the call for spam, you can do whatever you like with it. For example, create a special rule according to which all calls with such a label should be sent to the secretary, and not to the sales managers. Or completely reject applications that meet the conditions:

• the subscriber number in the list of numbers that generate a suspiciously large number of calls;
• the subscriber number in the list of numbers that generate many simultaneous calls (the larger the network, the better such citizens are visible);
• the subscriber number is in the network of the operator, which generates a disproportionately large number of calls in our direction;
• the first few (10-15) seconds of the connection record are the same (the robot reads the message);
• and - the cherry on the cake - the voice is recognized when compared with the voices of robots from the updated database.

Behavior blocking

Frode is a real headache for clients (and ours too) who are willing to pay dearly for attracting a targeted call. First of all, these are real estate and car dealers.

We call Froder a person who is ready to portray a target client for ~ 50 rubles. Such ads regularly appear on sites where they offer "high earnings on the Internet without investment."

The problem is that these “clients”, having completed their questions, dissolve without a trace, and the company loses money twice: to attract a call and to work a manager with such a dummy.

Nevertheless, there is something to cling to here. Froder is characterized by the fact that he never turns into a deal. Order the work of Froder advertising platforms or agencies that want to fulfill the plan or show that they are better than others.

Mixing 2-3% of garbage into the stream of honest calls makes little economic sense. Therefore ... therefore, the rogue pretends to be a markedly lower conversion rate.

Froder has his own pattern of behavior: he is unlikely to “shine” his personal phone to make a paid call, rather take advantage of the SIM card he purchased specially for this purpose, which is then removed from the phone before the next “job”: to the managers of the company whose product he was so keenly "interested", they did not bother with calls.

Many froders do not pass the check for frequent disconnections. In addition, you will not call one victim from the same number twice - you need new SIM cards, but the voice is the same. And this is the second point that we check.

Here, only the main steps were listed, in general, the verification algorithm is much more complicated and takes into account many behavioral factors. We will not reveal the remaining stages, in order not to facilitate the work of the froders.

In general, this is a much more difficult task compared to the robot's anti-spam protection. But, fortunately, absolute efficiency is not required. It is enough to be able to identify sources of calls that generate suspicious activity more often than others - this information is quite enough for our customers to optimize their work.