Adding intelligence to voice data

Technology now makes it possible for devices such as domestic appliances, cars, video games and consoles, TVs and wearables to include conversational capabilities which combine with and enhance any existing “intelligence” to exciting new levels of interaction and personality – without the device having to be connected.

One of the companies leading this new field of customer engagement is Artificial Solutions, which has just introduced Teneo 5, which is described as a true hybrid model that seamlessly integrates linguistic and machine learning. Machine learning techniques deliver the closed-loop of automated learning, optimizing and improving; while Teneo 5’s linguistic elements ensure human control for optimum performance.

Automotive Industries (AI) asked Andy Peart, Chief Marketing and Strategy Officer, Artificial Solutions, what are the key features/improvements in Teneo 5.

Peart: We’ve placed a great deal of emphasis on how conversational data is generated and used. Our automotive customers enjoy a level of visibility, flexibility and control of their conversational AI projects that is unmatched in the market-place. Conversational data generated by the application is integrated seamlessly across all modules of Teneo 5 enabling real-time access to business data insights and data-driven development that automatically enhances the interaction and maintains the system. In addition, key stakeholders can measure the performance of the conversational application with KPI reporting and management information.

Another key area is a hybrid approach that provides a native interface which ensures seamless integration between linguistic and machine learning. Finally, Teneo 5’s enhanced technology stack makes installation simpler, and allows for faster roll-out of new environments, particularly in applications where the auto-maker choose to self-host.

AI: What is the response from the OEMs?

Peart: Extremely positive, both in Europe and the US. We’re working with a number of the leading manufacturers and suppliers to deliver conversational AI solutions across a range of use-cases ranging from infotainment systems, through customer service to driving customer engagement on websites.

I think one of the main reasons Conversational AI is set to become such an important component of the wider AI story is that the benefits impact both the customer and automotive OEM.

Customers value conversational interfaces because they are fast, intuitive and convenient. For automotive OEMs, conversational AI offers a way to build a more personalized and engaging customer experience. We’re seeing automakers typically focusing on four main drivers for their investment in Conversational AI:

Drive revenue: Intelligent chatbots guide customers on a buying journey including gentle upselling. Anonymized conversational data can be used to understand trends and better interpret customer sentiment, providing invaluable insight that informs product and service development.

Peart: By automating a proportion of the calls, emails, SMS, social media messages and live chat sessions conversational AI frees up time to allow employees to focus on higher-value customer engagements. Automated customer services bots operate around the clock, eliminating the need for out-of-hours service and can reduce calls to service centers by up to 40%.

Offer new methods of customer engagement. Automating customer service using a virtual assistant can scale across millions of users, and allow customer issues to be answered in an efficient yet human-like manner, 24 hours a day.

Build differentiation. A voice interface provides an intuitive replacement to a complex menu system for home automation and security.

AI: What are some of the challenges?

Peart: When we converse as humans, we are unpredictable. We don’t say what is expected. We use our own terminology. We branch off into tangents. We circle back. We miss out crucial facts and figures, and we ask for clarification. It becomes extremely challenging when you add multiple languages across different channels and industry sectors, as well as supporting multiple input modes including voice, text and touch.

Automakers therefore need toolkits – specialist developer platforms – that include all the necessary components to build, deploy and maintain conversational AI systems. We like to think Teneo is that platform – it’s been designed specifically for enterprises to build, deploy and analyze the world’s most intelligent, humanlike and capable conversational solutions.

AI: How can Teneo help in rapid deployment of NLI applications in cars and keep costs down?

Peart: Teneo is already known for reducing development timescales from potentially years to months -or even weeks – when compared with other conversational AI platforms. The seamless integration allows data to be shared in real-time across all modules of Teneo – delivering a wide range of benefits from AI-powered improvement suggestions to automatically prioritizing the best training data to use for optimum machine learning algorithms.

At the same time, Teneo’s front-end graphical interface provides immediate access to data-driven insights, enabling the conversational AI designer to easily understand what’s working with the dialog flow, and more importantly what’s not. Manual adjustments can be made where necessary.

The interface also allows key stakeholders to measure the performance of the conversational application and understand customer satisfaction levels at a glance. In addition, detailed reporting of the conversational AI data delivers actionable insight to other areas of the business from product management and marketing to distribution. Further, open architecture has made Teneo one of the most flexible conversational AI platforms available, with automotive OEMs able to plug-in and seamlessly and integrate their own AI assets.

AI: How many languages does Teneo cater for?

Peart: Teneo allows business users and developers to collaborate on creating sophisticated, highly intelligent applications that run across 35 languages, multiple platforms and channels in record time. There are three primary use cases for conversational AI within the automotive sector: customers, employees and devices. In some instances, it maybe that an application may support more than one use case type. For example, a customer service app aimed at helping customers select an appropriate make and model, to drive them to a car-configurator and book a test drive.

Conversational AI is typically used:

Between enterprises and customers: Automakers are able to create frictionless journeys for their customers as they interact over a wide variety of digital channels and devices, and to understand the true voice of the customer, thereby opening new revenue opportunities.

Between enterprises and employees: Conversational AI is being increasingly used to extend the robotic process automation (RPA) proposition, allowing for RPA and other AI assets to be integrated into conversational applications to deliver “zero intervention” solutions for high-volume processes.

The ‘in-car’ experience: Conversational AI is gaining strong traction in the automotive markets where reliance on clunky menu systems to operate various devices are seen as a barrier to engagement.

AI: are you addressing the privacy issue?

Peart: Our platform provides businesses with the flexibility to self-host on site or use cloud hosting options that can be tailored to the most exacting security conditions, across multiple geographies and legal requirements. Additionally, the knowledge Teneo gathers is stored in one place which streamlines the querying and interpreting of conversational data and allows for the easy identification of any personal identifiable information (PII) and delete it if needed. It is also possible to pseudonymize PII data to enable conversations to still be used for statistical analysis and data insight even when the PII has been removed.

AI: How does company meet the General Data Protection Regulation (GDPR) requirements?

Peart: Teneo’s data analysis tools, integrated within the platform, already enable this by an easy-to-use graphical interface that allows enterprises to analyze free format, conversational data in real-time. Originally it was developed to enable information to be reused instantly back into the conversation, for maintaining the systems, and for insight into business trends. Modifications now enable it to be used to identify personal information and, depending on the organization’s GDPR policy, to pseudonymize or redact the data, but still gain value from the remainder of the information. Storing all the conversational data in one area makes it easy to search and retrieve, and to process the data in accordance with an organization’s GDPR policy before the information is stored.

AI: What are some of the challenges facing connected and conversational cars and how does Artificial Solutions hope to meet these?

Peart: The menu-driven command style of most automotive apps does little to engage the customer or make their lives more convenient. Conversational AI changes this by allowing customers to ask questions and give orders using their own words. Pre-built apps only offer a limited functionality. This restricts your ability to differentiate the brand and vehicles within the range. Teneo allows seamless integration into other systems, creating a conversation that goes beyond the boundaries of the vehicle to interact with services such as home automation, charging stations, or road-side assistance.