Thought chatbots were trendy? Make way for Interactive Virtual Assistants (IVAs). But can chatbots change the customer experience? Can IVAs like Amelia go beyond the limitations of chatbots? There is plenty to consider.

The customer experience is constantly evolving. It's not just about finding your customers where they are - it's also about anticipating where they are going, what they need, and how your brand can best support them.

Over the past decades, customer experience has become one of the top business priorities. For many industries, customer experience is the most critical differentiator between brands, and can significantly impact customer loyalty and sales.

By approaching all business decisions from a customer-centric perspective, brands can begin to understand the complexity and connectivity of all customer touchpoints (and behind-the-scenes processes). Once a business realizes the value of the customer experience, a customer journey map can assess the current state.

Businesses will find massive pain points through existing processes and technologies such as legacy automation, IVR, and DTMF/touch-tone systems. Businesses are composed of existing processes and technologies, such as legacy automation, IVR, and DTMF / touch-tone systems.

While these systems were once the standard for routing and redirecting calls, they no longer live up to changing consumer expectations. This is where digital transformation with conversational AI comes in. Digital transformation is the process of adapting traditional business processes and adapting them to the digital-first world. In terms of customer experience, this means using conversational AI to enhance the customer experience with (hopefully) effortless and productive automated conversations. But there is more to this technology.

Chatbots were invented in the 1960s, but the recent focus on customer experience, combined with more advanced website integration, has resulted in a boom in popularity in recent years. A chatbot works by being programmed to give responses that match the most similar keywords or phraseology patterns based on what the customer said. This results in a quick response to a customer's questions, but it is not always correct.

These pre-programmed automated interfaces communicate through online channels like a website, and on social media platforms like Facebook Messenger, Whatsapp, Skype, Slack, WeChat, and more. Chatbots are most useful for predictable and straightforward tasks, such as answering frequently asked questions, such as store opening hours.

Organizations often rely on chatbots as the first point of contact for customers. Although chatbots are widespread, they often do not meet customer expectations. Why? Chatbots cannot determine the underlying context, leading to many misunderstandings, repetition, and dead ends for customers.

If a user tries to find an answer to something not in the chatbot's algorithm, they have no choice but to switch to a different channel, which will take more effort for the customer, or they might walk away without it. The role of technology has changed with customer expectations. Once only used to route calls through systems like DTMF signaling and IVRs, the shift to a cohesive customer experience has resulted in technology playing a more prominent role in the customer journey.

With more advanced options like intelligent virtual assistants (IVA), brands can now use the technology as a self-service option with a customer experience that they hope is on par with a human agent. Decision-makers have to weigh the gamut between customer experience and operational costs, and often one is at the expense of the other. However, with more advanced technology and operational efficiencies, an IVA can offer a premium customer experience with faster value and scalable customer service transformation.

What is conversational AI?

Conversational AI is the set of technologies behind automated messaging and voice-enabled applications that enable human-like interactions between computers and humans. Applied conversational AI requires science and art to develop successful applications incorporating context, personalization, and relevance in human-computer interaction.

The fundamental aspect of conversational AI applications is designing processes that sound natural, and the result is indistinguishable from what a human could have delivered. Remember when you last contacted a company and you could have completed the same tasks with the same effort, if not less, than you would with a human? This is the highest quality conversational AI.

Conversational AI uses various technologies, such as automatic speech recognition (ASR), natural language processing (NLP), advanced dialogue management, and machine learning (ML) to understand, respond to, and learn from every interaction.

Human language is full of grammar exceptions, dialects and other eccentricities complicated for AI to understand. Words can have different meanings in different contexts. NLP enables an IVA to understand a customer's language, recognize their intention, and produce a response.

What is conversational AI? Artificial Intelligence IVAs can use multiple AI technologies, including NLP, ASR, and TTS, to understand what a person is saying, process it, and create a formal response. Machine learning is a set of algorithms used to teach a computer to perform specific tasks without being explicitly programmed. IVAs use machine learning and deep neural networks (DNNs) to learn and become more competent as they process more and more transactions. This technology allows customers to speak in their own words rather than follow a particular path. Automated Speech Recognition (ASR), one of the underlying technologies of AI, is the foundational technology that allows computers to understand spoken words. Today's ASR can be trained to understand languages, accents and other variations of the language.

Amelia - forging new ground in conversational AI?

Compare that to Amelia; a conversational AI agent referred to by her developers as "the most humane AI for the business." According to Anil Vijayan, Vice President of Everest Group: Amelia combines the ability to address a wide variety of use cases, such as customer service, human resources support, marketing and IT helpdesk, with the ability to operate across all channels, including voice and offers a virtual agent solution so bright that it is suitable for multiple business needs. (Amelia was included in the Everest Group Smart Virtual Agents Product PEAK 2020 Matrix Assessment).

You can see right away that Amelia is not ready for something as complicated as cancer therapy, which is why Amelia is successful. IPsoft (the company that created Amelia; the company is now renamed Amelia also) falls into the category of cognitive and conversational AI. Amelia, its flagship product, which somehow fascinates me, was recently classified by Everest Group as "Learning and language, automation capabilities, architecture, development and tools, implementation and security, optimization and analysis, vision and roadmap."

Amelia also received an OnPar rating in Everest Group's "Chatbot Readiness and Market Focus" categories. That makes sense. Amelia is pretty far from being a chatbot. Amelia is so powerful that she can train digital assistants for specific tasks without having all the capabilities of a cognitive digital assistant, which sets her apart from RPA and chatbots.

This is very different from RPA and chatbots, which are primarily scripted. To be fair, some chatbots can learn and expand. I will compare RPA, chatbot, cognitive conversational artificial intelligence technologies and solutions in a future article. By then, I will be able to understand Amelia's new effort, Digital Workforce Automation. From what I can tell, it will be more competitive with Amelia on price and time. Amelia will train "Digital Employees" through Amelia.ai, a "one-stop-shop online marketplace where you can search, interview, and onboard digital workers, powered by Amelia, the industry-leading cognitive artificial intelligence solution."

My take

The phrase "conversational cognitive AI" made me cringe a bit. They seem so intelligent, but their "learning" seems limited to the application they are trained for. While AI can get people to believe and do things, it can also be used to power robots that are problematic, when their processes or appearance involve delusions that threaten human dignity.

Putting conversational AI to work in enterprise settings is another matter entirely. I got into that my recent piece, NLP brings interactive analytics forward - but what are the requirements to make augmented AI work on your project?.