Periodic Reporting for period 2 - CLARA (The first superbot to audit calls)
Berichtszeitraum: 2020-10-01 bis 2021-09-30
In every Call Center it is necessary to monitor calls and verify that voice interactions with clients or leads conform to the guidelines and policies of quality established by the company. Calls have plenty of useful data, but only a small percentage of the recorded calls is being randomly evaluated and large amount of relevant data is lost, since monitoring is mainly done manually and it’s impossible to review all the calls. This results in poor customer service and ultimately business and profit losses.
The objective of CLARA project is to deploy a unique Super-Bot Speech Analytics system, to offer an effective way to leverage customer call data in an automatic way. It offers a variety of must-have features to improve customer service experience. The Speech-to-Text feature can automatically convert full audio recordings into text to easily identify root causes in customer interactions. The tool is capable of entirety transcribing all recorded calls, and then analysing the interactions using advanced speech detection technology. The audio signal analysis is able to evaluate customer – agent sentiment and provide a more robust and precise approach to customer support analysis.
Through machine learning, the system allows to discover actionable insights for effective analysis of calls and interaction with customers, leading to a more efficient business management and eventually to improve brands performance and benefits.
The objective of CLARA project is to deploy a unique Super-Bot Speech Analytics system, to offer an effective way to leverage customer call data in an automatic way. It offers a variety of must-have features to improve customer service experience. The Speech-to-Text feature can automatically convert full audio recordings into text to easily identify root causes in customer interactions. The tool is capable of entirety transcribing all recorded calls, and then analysing the interactions using advanced speech detection technology. The audio signal analysis is able to evaluate customer – agent sentiment and provide a more robust and precise approach to customer support analysis.
Through machine learning, the system allows to discover actionable insights for effective analysis of calls and interaction with customers, leading to a more efficient business management and eventually to improve brands performance and benefits.
During the project, the technical maturation (WP1) and integration of the technology (WP2) have been implemented according to the established project plan. Those tasks set the basis for the subsequent validations (WP3) and pre-commercialisation (WP4) activities, successfully implemented during the second period of the project. In this regard, successful integrations and proof of concept have been developed, and several agreements with relevant stakeholders have been achieved. Likewise, the communication, dissemination (WP5) and management (WP6) tasks have been effectively implemented, allowing to achieve the expected results of the project. The product name was rebranded to UPBE for commercialisation purposes, and multiple online dissemination actions have been deployed with good reception and participation of the audience, raising awareness of our solution and facilitating the next stages towards commercial roll-out.
CLARA project offers a unique and effective way to leverage customer call data, in an easy and automatic way.
First, through a well-developed User Interface business user are now able to set up and implement machine learning models in an easy and extremely fast way. This way we overcome two concerns that the speech analytics traditionally have. One is the need of IT support for setting up projects, second, large budgets reducing the amount of companies that can use this technology, third, long periods of time to implement changes to campaigns. Call Centers are environments very dynamic requiring constant changes very fast, sometimes from one day to other. Giving these tools to the business user we have overcome these concerns. In fact, one of our commercial claims is “Artificial intelligence applied to phone calls driven by your operations team”
From a technology perspective, we have been able to achieve the following:
1. Become the most accurate speech to text engine for Call Center domain in Spanish language.
2. Develop a quality and compliance insights template “branded as UPBE Score” which can identify automatically over 50 categories in phone calls and provide a score for each interaction.
3. Develop a set of dashboards allowing the operations team to understand what is happening in their call center. Through multiple visualizations this can be assessed from a campaign level, agent level or even call level. If there is any critical error or alert in a call the system will trigger this information to the user.
4. Perform both text and audio signal analysis and combine both approaches to provide the most accurate assessment of sentiment analysis in phone calls.
5. Develop more sophisticated approaches to provide the most important metrics to Call Centers as a pre-set-up module on our product, thanks to all the training and learnings achieved. This reduces the personalization and set-up of customers and will accelerate our go to market strategy, providing the only tool technology able to do this with specific models per industry.
Therefore, the technology developed in CLARA project has a clear potential to improve companies’ competitiveness at EU and global level, by improving their customer services, performance, and revenues, as well as by creating qualified employment. This will also have a positive impact on society since companies will be able to offer an optimal customer experience with satisfactory results.
First, through a well-developed User Interface business user are now able to set up and implement machine learning models in an easy and extremely fast way. This way we overcome two concerns that the speech analytics traditionally have. One is the need of IT support for setting up projects, second, large budgets reducing the amount of companies that can use this technology, third, long periods of time to implement changes to campaigns. Call Centers are environments very dynamic requiring constant changes very fast, sometimes from one day to other. Giving these tools to the business user we have overcome these concerns. In fact, one of our commercial claims is “Artificial intelligence applied to phone calls driven by your operations team”
From a technology perspective, we have been able to achieve the following:
1. Become the most accurate speech to text engine for Call Center domain in Spanish language.
2. Develop a quality and compliance insights template “branded as UPBE Score” which can identify automatically over 50 categories in phone calls and provide a score for each interaction.
3. Develop a set of dashboards allowing the operations team to understand what is happening in their call center. Through multiple visualizations this can be assessed from a campaign level, agent level or even call level. If there is any critical error or alert in a call the system will trigger this information to the user.
4. Perform both text and audio signal analysis and combine both approaches to provide the most accurate assessment of sentiment analysis in phone calls.
5. Develop more sophisticated approaches to provide the most important metrics to Call Centers as a pre-set-up module on our product, thanks to all the training and learnings achieved. This reduces the personalization and set-up of customers and will accelerate our go to market strategy, providing the only tool technology able to do this with specific models per industry.
Therefore, the technology developed in CLARA project has a clear potential to improve companies’ competitiveness at EU and global level, by improving their customer services, performance, and revenues, as well as by creating qualified employment. This will also have a positive impact on society since companies will be able to offer an optimal customer experience with satisfactory results.