Every month we receive 20–50 thousand calls with questions about ATM servicing. Most often, engineers call to find out the status of an application, gain access, check software versions, etc. Or cash collectors - to understand whether there are any malfunctions on the ATM they are dissecting. The questions are the same in 90% of cases.
We took engines for voice automation and speech technologies, combined them and got a robot that helps people, connected it and put it on the line.
The functionality was the same as that of the operator, but the engineers fundamentally did not want to communicate with the robot. Even if it was a typical question “is everything okay with the ATM?” Then we changed the voice to a pleasant female one, tested it in AB with a male one - and the number of switches to an operator from a female robot fell: 24% of processing with Denis and 65% with Julia.
When we were thinking about how to automate the support service and process thousands of standard requests, we considered the option of creating a mobile application for these purposes. But they quickly dismissed this story for two reasons. The first is that the Internet of the mobile operator that is used by the employee servicing the ATM - an engineer, cash collector or other representative of the enterprise as part of automating access lists - is not available everywhere. Some ATMs are located in places where the only communication method is voice. For example, the Zima station. There is an ATM there, but no internet access. And in some factories the Internet may be jammed due to the specifics of production and in the name of security. Accordingly, even if we had at least a dozen special mobile applications, they would not save us in such conditions.
The second reason is that not all heroes wear capes, and not all service employees use smartphones. We cannot (and should not) force them to purchase devices that support the mobile application.
In turn, we ourselves do not see the advisability of creating an application for outdated versions of different operating systems. What if a person has an old push-button telephone? The same applies to options with different chatbots and support via instant messengers.
Accordingly, unlike a mobile app, a phone call that can be made from anywhere and on any device allows customer support to truly become more accessible.
For example, an ATM must be left in some unusual condition, which is not described in any of the regulations or is rare. For example, the upper cabinet does not close completely. Or there was an error in identifying the collector: the personnel number, the group of ATMs that it services, and some other data did not match for some reason. The share of such calls in the total is small: with unsuccessful identification - 15.93%, with incorrect - 4.82%, with successful - 79.25%, according to data for August 2023. For other months the statistics are approximately the same.
In such a situation, the operator is connected. He figures out what happened. If there is an identification error, we cannot trust the work to be carried out through a voice assistant, because we risk the security of the ATM, and security is the main principle of service. At the same time, we also cannot lose the call. Let this not be a bank client, whose anger with a dropped call would harm the bank’s reputation. But by leaving the collector without feedback, we risk leaving the ATM in an inoperative state, that is, inaccessible to the client. The collector will simply leave because his time is limited.
For example, collectors called to report that they had finished work and to find out if they could leave. They stood with money at ATMs and were nervous, waiting for an answer. It was necessary that all these calls, which require a formal response without consultation, were taken by non-operators.
What other problems did we want to solve by introducing voice assistants:
The robot Julia talks to engineers and cash collectors. She advises them, and this robot can answer the question of whether the cash collector is allowed to leave and whether he leaves the ATM in working condition. It will check the status of all equipment on the servers to see if everything is shown as working properly. If something goes wrong, for example, a jam occurs in the receipt printer, she will offer help and tell you how to solve the problem.
Before the introduction of our voice assistants, situations often arose when cash collectors formally carried out their procedures, changed cassettes and left without making sure that the ATM was working. I had to call the collection service again and pay for this trip. Cash cassettes move back and forth, recalculation takes place at the cash desk - the bank suffers losses. The appearance of Julia helped get rid of this problem, because you can quickly reach her by phone and get all the necessary information. Voice assistants are always in touch and work 24/7.
The work was carried out based on the Agile approach using the Scrum methodology. Depending on the backlogs, we combined from two or three to 12 specialists into a team. We have key deliveries for the year, they are divided into quarterly releases with deliveries every two weeks and feedback. The work was carried out in stages. We needed:
And we always had parallel development teams for our incident management and a processing development team. When it was necessary to improve the product in related systems or obtain information from them, the system analyst described the technical specifications for integration and data, and then submitted requests for improvements to neighboring teams.
Today we have moved further: most of the development has moved to streams that unite teams in a certain direction. For example, I lead the streaming product cluster, which develops all service products. There is a stream that develops processing technologies. There is a stream that develops software for ATMs. Within the stream and between them, we set tasks to create a single product, synchronizing with the supply cycle once every two weeks. Thus, the whole story with voice assistants is integrated within the overall range of the platform, the products that we support.
Tasks get into the backlog in three ways - these can be requests for automation, support, or feedback from our business customers. For example, they can calculate what a feature is missing. Conventionally, ask voice assistants to ask collectors: “How are you feeling today?” For us, this is a meaningless task, but if the customer finances the creation of the product, we can implement such functionality. But often we receive more useful feedback and gradually correct and improve something.
We made a simple IVR, where, by listening for a long time and moving through a complex menu, calls were categorized according to basic directions. The operators then supplemented this information with algorithms for solving basic queries. We were primarily interested in the points of intersection between the volume of calls and the ability to delegate them to robots as much as possible.
The center's technologies recognize Russian speech quite well, but not all callers have perfect diction. Some nuances of speech can “break” the voice assistant, and it will not respond. These are speech defects, unclear diction, and a strong accent. We solved this problem by using keyboard input. For example, if the robot was unable to recognize the ATM number because the caller has trouble pronouncing numbers, it will ask you to enter their number manually on the phone. We didn't invent this. This is a classic telephony feature. We simply integrated it into our service.
We retained the ability to connect the operator, and the specialists waited for his response, ignoring the robot. The solution was found experimentally. At the very beginning, the voice assistant spoke in a male voice. It was the robot Denis. We did not attach any importance to what gender our robot would be. And the engineers were very indignant when the female voice contact center was replaced by Denis. I admit, we thought for a long time what to do. And then, as an experiment, they replaced the male voice with a female one. This is exactly how our Julia appeared. By the way, her name is associated with our business analyst and key leader Yulia.
As a result, our number of calls that were transferred further to the operator dropped sharply. According to the latest data, for September 2023, 65% of calls are resolved through a voice assistant and 35% of calls are accepted by operators. A year ago, in September 2022, robots handled only 24% of calls. And in 2021, the figure for September was several times lower - only 2%.
In general, all our experience with the implementation of voice assistants is the integration of one working technology into another. If you believe history, the first voice assistants appeared in the last century, in the late 30s. For our part, we suggest how to expand their functionality.
We took engines for voice automation and speech technologies, combined them and got a robot that helps people, connected it and put it on the line.
The functionality was the same as that of the operator, but the engineers fundamentally did not want to communicate with the robot. Even if it was a typical question “is everything okay with the ATM?” Then we changed the voice to a pleasant female one, tested it in AB with a male one - and the number of switches to an operator from a female robot fell: 24% of processing with Denis and 65% with Julia.
When we were thinking about how to automate the support service and process thousands of standard requests, we considered the option of creating a mobile application for these purposes. But they quickly dismissed this story for two reasons. The first is that the Internet of the mobile operator that is used by the employee servicing the ATM - an engineer, cash collector or other representative of the enterprise as part of automating access lists - is not available everywhere. Some ATMs are located in places where the only communication method is voice. For example, the Zima station. There is an ATM there, but no internet access. And in some factories the Internet may be jammed due to the specifics of production and in the name of security. Accordingly, even if we had at least a dozen special mobile applications, they would not save us in such conditions.
The second reason is that not all heroes wear capes, and not all service employees use smartphones. We cannot (and should not) force them to purchase devices that support the mobile application.
In turn, we ourselves do not see the advisability of creating an application for outdated versions of different operating systems. What if a person has an old push-button telephone? The same applies to options with different chatbots and support via instant messengers.
Accordingly, unlike a mobile app, a phone call that can be made from anywhere and on any device allows customer support to truly become more accessible.
When is an operator needed?
But there are times when you cannot do without the help of an operator. These are moments related to organizational issues, and we have not automated them yet.For example, an ATM must be left in some unusual condition, which is not described in any of the regulations or is rare. For example, the upper cabinet does not close completely. Or there was an error in identifying the collector: the personnel number, the group of ATMs that it services, and some other data did not match for some reason. The share of such calls in the total is small: with unsuccessful identification - 15.93%, with incorrect - 4.82%, with successful - 79.25%, according to data for August 2023. For other months the statistics are approximately the same.
In such a situation, the operator is connected. He figures out what happened. If there is an identification error, we cannot trust the work to be carried out through a voice assistant, because we risk the security of the ATM, and security is the main principle of service. At the same time, we also cannot lose the call. Let this not be a bank client, whose anger with a dropped call would harm the bank’s reputation. But by leaving the collector without feedback, we risk leaving the ATM in an inoperative state, that is, inaccessible to the client. The collector will simply leave because his time is limited.
The era before voice assistants
Previously, before voice assistants, you could wait up to half an hour for a response from support services, since we are, after all, not a contact center, but a company involved in centralized management of the service. During peak times, the number of parallel calls could be several times higher than the number of calls during quiet hours. At the same time, appeals could be of a formal nature.For example, collectors called to report that they had finished work and to find out if they could leave. They stood with money at ATMs and were nervous, waiting for an answer. It was necessary that all these calls, which require a formal response without consultation, were taken by non-operators.
What other problems did we want to solve by introducing voice assistants:
- increase network availability, reduce response waiting time;
- save money that can be used to develop and improve the service.
Robot Julia and robot Sigma
We created voice assistants and divided responsibilities between them. We have assistant Sigma. It automates access lists (remember the story about an engineer who needs to get to a factory where there is an ATM). Anyone who needs to check the access list communicates with her.The robot Julia talks to engineers and cash collectors. She advises them, and this robot can answer the question of whether the cash collector is allowed to leave and whether he leaves the ATM in working condition. It will check the status of all equipment on the servers to see if everything is shown as working properly. If something goes wrong, for example, a jam occurs in the receipt printer, she will offer help and tell you how to solve the problem.
Before the introduction of our voice assistants, situations often arose when cash collectors formally carried out their procedures, changed cassettes and left without making sure that the ATM was working. I had to call the collection service again and pay for this trip. Cash cassettes move back and forth, recalculation takes place at the cash desk - the bank suffers losses. The appearance of Julia helped get rid of this problem, because you can quickly reach her by phone and get all the necessary information. Voice assistants are always in touch and work 24/7.
Entering the service department
Integration of voice assistants took about six months. Initially, the team included two developers and two business analysts. Now the composition of the core team looks like this: lead, IT lead, two business analysts, two business process automators, four developers and testers.The work was carried out based on the Agile approach using the Scrum methodology. Depending on the backlogs, we combined from two or three to 12 specialists into a team. We have key deliveries for the year, they are divided into quarterly releases with deliveries every two weeks and feedback. The work was carried out in stages. We needed:
- transform business idea into technological requirements;
- make automation at the contact center and integration with the speech technology center;
- provide initial scripting of the trees along which the voice assistant walks, depending on what it hears from the person.
And we always had parallel development teams for our incident management and a processing development team. When it was necessary to improve the product in related systems or obtain information from them, the system analyst described the technical specifications for integration and data, and then submitted requests for improvements to neighboring teams.
Today we have moved further: most of the development has moved to streams that unite teams in a certain direction. For example, I lead the streaming product cluster, which develops all service products. There is a stream that develops processing technologies. There is a stream that develops software for ATMs. Within the stream and between them, we set tasks to create a single product, synchronizing with the supply cycle once every two weeks. Thus, the whole story with voice assistants is integrated within the overall range of the platform, the products that we support.
Tasks get into the backlog in three ways - these can be requests for automation, support, or feedback from our business customers. For example, they can calculate what a feature is missing. Conventionally, ask voice assistants to ask collectors: “How are you feeling today?” For us, this is a meaningless task, but if the customer finances the creation of the product, we can implement such functionality. But often we receive more useful feedback and gradually correct and improve something.
The speech is lively and categorical
We dealt with all problems related to parsing through the speech technology center. For example, dividing calls into categories. Who called, why they called, how they fixed it - all this had to be tracked in order to determine the segments where they call the most and where it is possible to resolve issues without the participation of people.We made a simple IVR, where, by listening for a long time and moving through a complex menu, calls were categorized according to basic directions. The operators then supplemented this information with algorithms for solving basic queries. We were primarily interested in the points of intersection between the volume of calls and the ability to delegate them to robots as much as possible.
The center's technologies recognize Russian speech quite well, but not all callers have perfect diction. Some nuances of speech can “break” the voice assistant, and it will not respond. These are speech defects, unclear diction, and a strong accent. We solved this problem by using keyboard input. For example, if the robot was unable to recognize the ATM number because the caller has trouble pronouncing numbers, it will ask you to enter their number manually on the phone. We didn't invent this. This is a classic telephony feature. We simply integrated it into our service.
Engineers don't want to talk to a robot
It would seem that there are no problems. We set up the work of our voice assistants, provided for everything and did everything technically so that the issue could be resolved without the participation of an operator. But they did not take into account the human factor. Our collectors and engineers admitted that they like to communicate with the girl from the contact center - lively, sweet with a pleasant voice.We retained the ability to connect the operator, and the specialists waited for his response, ignoring the robot. The solution was found experimentally. At the very beginning, the voice assistant spoke in a male voice. It was the robot Denis. We did not attach any importance to what gender our robot would be. And the engineers were very indignant when the female voice contact center was replaced by Denis. I admit, we thought for a long time what to do. And then, as an experiment, they replaced the male voice with a female one. This is exactly how our Julia appeared. By the way, her name is associated with our business analyst and key leader Yulia.
As a result, our number of calls that were transferred further to the operator dropped sharply. According to the latest data, for September 2023, 65% of calls are resolved through a voice assistant and 35% of calls are accepted by operators. A year ago, in September 2022, robots handled only 24% of calls. And in 2021, the figure for September was several times lower - only 2%.
In general, all our experience with the implementation of voice assistants is the integration of one working technology into another. If you believe history, the first voice assistants appeared in the last century, in the late 30s. For our part, we suggest how to expand their functionality.
