HOW SPEECH TECHNOLOGY HELPS MANAGE EMPLOYEES AND EARN MORE
The larger the company, the more difficult it is to control and improve the quality of service for each client. But imagine that it is possible to analyze every dialogue between an employee and a client, while not creating a whole quality control department.
Or call 1000 subscribers in just a few seconds, without having a huge call center. All this can now be done with the help of speech technology.
Fonemica Executive Director Igor Basko talks about what speech technologies are and how they help Russian business effectively:
Fonemica Executive Director Igor Basko talks about what speech technologies are and how they help Russian business effectively:
Recognize customer needs
Automate and standardize business processes
Save time and earn more
How speech technology helps business - with case studies and case studies
Speech technology is an artificial intelligence that recognizes speech, analyzes and synthesizes a person’s voice.
Speech recognition and speech analytics help you quickly and objectively analyze conversation recordings. Sales managers can find out:
Does employee follow script
How well does he know the assortment
Does it talk about stocks?
How often does the client refuse
Is the manager friendly enough to communicate with the client
Whether the speech is well delivered or parasitic words slip through the conversation
Does the customer base divert «to the side»
What is the reason customers call most often
How often are customers dissatisfied with something
Here is an example of how this works:
To track whether the script manager is working, it’s enough to create semantic blocks in the system that the employee should say. For example, a greeting — a presentation of the main product — working out objections — a story about a promotion — closing a deal — farewell.
In each semantic block — a set of phrases that employees usually use
After processing the record, the manager receives a ready-made report for all employees in one table, where it is clear which staff is moving away from the script and communicating incorrectly with the client.
The system automatically generates a rating and statistics of conversations for each employee
If necessary, each conversation can be analyzed in detail. For example, listen to the context in which the operator used stop words or how he offered a particular product. This is useful if you can see that he often offers customers certain products, and there are few closed deals..
The center displays the number of mentions of words from each stage of the script. By clicking on a word, you can go to the recordings right to the moment of mentioning and listen to the context
Manage staff more effectively and build a training system
All employees receive the same introductory information about products, services and work standards — be it a training, seminar or acquaintance with a knowledge base. But to determine «by eye», who and how learned the knowledge, where the competencies are weak and how well each employee knows the assortment is practically impossible until a precedent in the form of a dissatisfied client arises.
The Fonemica platform has detailed analytics for each employee. As a result, it is clear who thinks for a long time about the answer, who incorrectly explains the advantages of the product, and who moves away from the script.
Based on the report, it is possible to identify «gaps» in the skills and knowledge of each and decide what to do with it: conduct a conversation, send for training or say goodbye.
The results of the introduction of speech analytics in retail
The retail electronics chain wanted to find out what questions customers most often call the contact center and how operators process calls.
For this, the records of existing conversations were uploaded to the speech recognition server and transferred to text.
The platform automatically processed the received text and identified key topics of conversation.
Speech analytics helped determine that:
39,4% of requests — are questions about the range, ordering, delivery and installation
in 9,6% of conversations, customers report a problematic situation with the product and express dissatisfaction
in 8,7% of cases, the operator is incompetent in the matter and cannot help, which means that almost a tenth of the calls do not help the client to close their need.
As a result, operator gaps in knowledge, competencies and communication skills were identified. Due to the fact that the system also generates an individual report for each employee, management was able to quickly formulate an effective staff training plan.
By the way. Employees cannot look at themselves from the outside and understand what they are doing wrong in a conversation with a client. In the process of dialogue, you do not pay attention to where you were good and where you could work out differently.
We decided that it would be cool if every day employees receive an objective feedback about their work and analyze what they say and what they forget. Therefore, they made a section for employees in the service where they can track their statistics and independently «pump over» skills.
An analysis of the dialogs makes it clear whether the manager made a sale or an additional sale, and track how transactions are closed.
But speech analytics is used not only to control the work of employees. We made it possible to analyze and speech customers. For them, you can also create semantic blocks of dialogs and add them to scripts. This helps to evaluate customer loyalty, move to the formation of an idea of customer experience and improve sales techniques or revise the range.
If you configure client speech analysis, you can:
«Hear» every review or wish. For example, the system can track the phrases «not enough», «I would like to», «but do you have any» and the like. At the end of each day, the manager sees who and what was dissatisfied with what else can be offered to customers and how best to do it.
Fonemica has developed a report that contains the most commonly used phrases. Their analysis will help identify customer needs before competitors think about it, or find weaknesses in service.
Find out what reduces sales. The system breaks down the dialogue into roles and understands where the employee is talking and where the client is.
The platform highlights the dialogs where there are phrases «it doesn’t suit me», «no, I don’t need it», «thank you, another time», and helps to understand why the sale breaks down and what part of the script needs to be redone.
In a conflict situation that has reached the leadership, chaos and confusion often begin. And with the help of scripts for clients, you can quickly receive information about conflicts, tracking where the client began to show a negative attitude. The leader listens to the context and only then draws conclusions, will understand, apologize and correct the situation.
Assess the quality of work of each employee at all points of sale. And no mystery shoppers
You can analyze the communication of employees with customers not only by phone. Voice recorder badge will help you get the same analytics and reporting for line staff: sellers, consultants, couriers, sales representatives, drivers or promoters.
The badge records all conversations, records are automatically transferred to the platform for further analysis
ЗRecording takes place all day, after — the system analyzes the recording in the same way as calls, and the manager receives the same detailed report. Inconvenient situations do not remain «behind the scenes». The recorder cannot be turned off, recording starts automatically when the device is removed from the platform for recharging.
The employee’s speech does not mix with the client’s speech. The badge records the conversation in «stereo». This allows you to analyze the speech of employees and customers separately.
The results of the introduction of badges-voice recorders in the network of manicure studios N.A.I.L.S Russia
With the beginning of the use of badges, it was possible to identify a whole pool of problems on all fronts:
In communication — in 24% of the dialogs, parasite words were heard
In the service — a five—day guarantee for coverage was reported to the client only in 63% of cases
In sales — employees made the presentation of the main product only in 54% of dialogs, and additional services and goods were offered only in 37% of cases
Neither selective listening to recordings of conversations of administrators, nor mystery shoppers could give such an accurate assessment of the quality of work and service.
It’s impossible to just say in the back the words and phrases «Hello», «We have a special offer», «Order for this product…», do not approach visitors all day and get a good rating in the system. The neural network not only recognizes individual words, but also analyzes the overall content of the dialogue and context. If the conversation is too short, the system does not count it.
Do not inflate staff
Sales managers, call-center operators, dispatchers, even secretaries or bank consultants can be replaced by one voice bot. Depending on the given scenario, he can make calls, and can receive up to 1 million calls per day.
The voice bot is a program with elements of artificial intelligence. It does not play pre-prepared sound tracks, but speaks (synthesizes) dynamic information based on a subscriber’s request.
Voice bot can:
Make mass calls and report on stocks and new arrivals
Remind customers about appointments, payments or delivery times
Gather feedback and conduct polls
Answer the most common questions
Find out the client’s problem and switch it to the right specialist, depending on the topic of contact
Reporting the weather on Mars and doing another million things — it all depends on the scenario invented.
At the same time, the voice bot will not forget to ring several customers from the list, will not mix up the name of the client and will not forget to mention the most important advantage of the product. He will work both on New Year's Day and on the Eighth of March without a salary or vacation, and he will not miss a call yet, because he went for lunch.
Speech of voice bots is indistinguishable from human. We taught the neural network to change the pace of speech and tonality. As a result, we get a variety of lively variations of phrases, rather than monotonous artificial speech.
An example of using a voice bot in a marketing research
For the manufacturer of baby food, we introduced a robot that phoned mothers with children under three years of age. The bot phoned 10,000 numbers in a day, and the next day the client had a ready-made report with all the necessary data on hand:
Total number of calls
How many answered / did not answer the call
How many answered the first question and the number of answers: Yes / No
How many answered the second question and the number of answers: I don’t buy at all (end of conversation) / I buy once a month / I buy every week / I buy every day.
How many answered the third question and the number of answers: In a hypermarket / In a store near the house / In a specialized store / Via the Internet
The client received the most accurate data, because the voice robot can recognize whole phrases and sentences, understands highly specialized terms, determines the gender of a person by voice, and ends the call if the answering machine is activated.
If the calls were made by live operators, they would need 170 times more time to do the same job, and it would cost 10 times more.
How to incorporate speech technology into your business so that it benefits
To get positive results, it’s not enough to simply process records from points of sale and a call center through the platform. Speech technologies are just tools, and by themselves they will not make a business successful.
Assign Integration Responsibility
You need a person who will:
Understand exactly why this is an innovation of the company, and communicate it to the heads of sales and service departments. The latter often ignore the platform, relying on more familiar working methods than on «some kind of artificial intelligence».
Thoroughly know how the platform works, and help others understand.
Formulate new work standards taking into account interaction with the platform, compile and submit reports on achieved results.
This may be a new employee or someone «easy on the rise», who is not afraid of a new one. Such a person is needed in order to configure the service to work specifically for the tasks of the company, and in which case to help to cope with the rejection and misunderstanding of the service in the team.
Present the platform to employees correctly
Especially if previously the company did not record communication with customers.
Often employees think that the company has started all this in order to «keep everyone under the hood». In this case, they are reluctant to interact with the platform, integration is delayed, and the effect is not what it could be. To form the correct attitude to the service:
Tell us how the platform works
Everything new and unknown is terrible and causes rejection. Give a presentation and explain how it works. Please note that in fact only the service «listens» to employees, tell us about the tasks set for the platform and your expectations from the implementation.
When everything is clear, working in a new way is not scary.
Show employee benefits
One of the main features of the service is working on yourself and your performance. Typically, sellers' salaries are tied to their results. Demonstrate how you can use the platform to hone your skills and improve results, and explain that service is one of the tools to increase their earnings.
For contact center operators and support services, motivation can be introduced that is tied to the «customer satisfaction index».
Explain that you are at the same time
Once again, emphasize that no one is going to follow anyone, on the contrary, the assessment of work based on the results of such analytics is most objective. This excludes the subjective opinion of mystery shoppers or a control service or manager. In addition, employees can always protect themselves from unreasonable reproaches from customers.
Make the most of your data
It happens that department heads simply do not want to understand either the service or the data that it gives. Or do not know how. As a result, the platform is useless simply because it was never used.
The service gives a huge amount of information about how in fact everything works in the company. It is important not just to «store» this data, but to constantly analyze it and come up with options for how to improve the results.