How is AI used to build contact center solutions at Diabolocom?
Kevin: First, think of a contact center as a place where people who need help connect with agents (human ones) who are there to assist them. Imagine someone calling about a confusing phone bill, a software glitch, or to reserve a table at a popular restaurant for Valentine’s Day. On the other end, there’s an agent trying to help hundreds of people like this every day.
AI, especially machine learning, is transforming how these interactions happen. It makes the process more efficient and satisfying for both parties. Machine learning can handle tasks that couldn’t be automated before. This is because, rather than manually programming every detail, we can now teach the machine to recognize the most important factors for each task using relevant data.
At Diabolocom, we utilize AI, particularly machine learning, to enhance the productivity and satisfaction of both agents and customers in our contact centers. AI enables automation of tasks that were previously manual, like sorting variables for specific tasks based on data.
Can you give an overall definition of what language and speech technologies are, and how they are used in your contact center solutions? How are these technologies implemented at Diabolocom?
Kevin: Language and speech technologies are systems that analyze, understand, and generate human language, largely developed using AI.
They are usually employed to automatically analyze what people say or write, either to gain insights (like diagnosing speech disorders or identifying key topics in thousands of customer messages), or to make interactions with machines more natural through speaking and texting, rather than clicking buttons or pulling levers (such as with chatbots).
Take our Agent Assist solution as an example. It’s a collection of tools designed to help call center agents interact more smoothly with customers by handling the tedious but essential tasks, like fetching documents. This allows agents to concentrate on what they excel at: engaging with the customer in a friendly, efficient, and empathetic manner. We are developing tools that include Automatic Speech Recognition (ASR) systems, which automatically convert spoken words into text in various languages, and tools that automatically gather information relevant to the customer’s needs based on their conversation.
With the sensitivity of data in AI applications, how does Diabolocom handle data security and privacy?
Kevin: We are committed to ethical AI practices. We put a lot of effort in ensuring that data sources are informed about the use of their data, in anonymizing sensitive information, and we adhere to strict security measures. Especially for our European partners, we guarantee that data remains within European servers.
The actual reason why security and trustworthiness regarding data is important to us, is that we strive to build and keep good relationships with our partners, and data safety is a key element to this end. These relationships are what allows us to build features suited specifically for contact center applications.
Given the prevalence of open-source projects, does Diabolocom use open-source resources to build their AI solutions, or are they built from scratch?
Kevin: Yes, absolutely, we do use them! While some of our AI solutions are built from scratch, many are developed using existing open-source projects. Given the high quality of these resources, it would be foolish not to take advantage of them. We not only use these projects but also enhance them. I want to clear up a common misconception: it’s often thought that open-source projects can be used right out of the box. In reality, many of these projects are designed either for general purposes or to address very specific problems. A significant amount of work goes into adapting these tools to our specific needs and ensuring they are user-friendly. At Diabolocom, we have the expertise and resources to leverage these tools to build solutions specifically tailored for contact center applications, thanks to our strong position in the industry and excellent relationships with our partners.
That being said, whether we develop our solutions from scratch or build-on open-source projects, we always maintain complete control over the final product. We do not rely on third-party APIs. This approach ensures greater security for our clients’ data, full transparency in our workflows, and more control over our implementations to better meet our objectives.
Is Diabolocom contributing to the open-source community as well?
Kevin: We strongly believe that open sourcing benefits not only the wider community but also our own company. Creating a broad community of contributors enhances our projects with diverse perspectives, which is crucial as we plan to expand globally. We’re actively working on several projects that will soon be made open, including our recent collaboration with CentralSupelec’s MICS lab, which led to the release of CroissantLLM. We have more exciting projects like this in the pipeline. Additionally, we’ve developed useful resources such as datasets and evaluation tools to address specific challenges we’ve faced, which we think could benefit the community at large.
In line with this, we maintain ongoing collaborations with academia and are always seeking new partnerships. For example we are currently working on participating in research consortia for projects aiming at reducing AI’s carbon footprint or at enhancing AI explainability.
In conclusion, Diabolocom’s strategic embrace of artificial intelligence, as articulated by Kevin El Haddad, underscores a progressive journey towards transforming contact center operations.
Curious to know more about Diabolocom?
