As well as being the topic of the moment, Artificial Intelligence (AI) is a subject close to my heart — my honors thesis and masters studies at university in the late 80s were focused on deep learning and artificial neural networks (ANN). The O’Reilly Artificial Intelligence Conference held recently in San Francisco offered a lot of valuable insights into how AI is increasingly being applied in all sorts of industries, and increasingly so in the customer service space (AI in customer experience). As part of the Cyara investment in our AI future, Cyara’s Chief Engineer Thomas Fejes and myself attended the conference.
What was immediately interesting was that while the field of AI has grown exponentially in terms of the scope and complexity of problems being solved, in many ways the core of the architectures and algorithms around ANNs has largely remained unchanged over the last 20 years+. The key factor enabling this change has been the availability of highly parallel computational hardware at a reasonable price point.
When I first started working with ANNs, the tool of choice at my university was our new Convex “super-computer” (a laughable description when compared with modern PCs). These super computers were barely capable of training even the most rudimentary of ANNs in realistic timeframes, while still at a cost of thousands of dollars per minute of compute time. In those days CPUs ruled. In the past 5+ years GPUs have been at the forefront of massively parallel compute power, and today we’re seeing Google’s TPU (Tensor processing unit), Intel’s custom FPGA ASICs, and will in the future see computers based on quantum mechanical phenomena. Slowly but surely, we’re moving towards the commoditization of equipment capable of mimicking at least some basic brain functions.
Wider Applications of AI
But what are companies doing with this extra processing capacity? There’s not a moment that goes by these days without some media outlet talking about autonomous vehicles. The core of that technology is high-speed computer vision and image processing. Deep learning ANNs are at the center of this burgeoning technology space — video cameras linked to AI engines capable of identifying people, other vehicles, road markings, signs and traffic signals all in real time.
Almost as innumerable as the different architectures of ANNs is the number of applications for this technology. Applications covered at the conference were many and varied; a small snapshot included natural language processing, emotion sensing, medical analytics such as early cancer detection, Pinterest’s use of automated image characterization, Microsoft using AI to keep a sailplane airborne by drawing on how birds fly, and autonomous driver policy development through massive data collection of global dash-cams. The applications are endless.
AI in the CX Space
Closer to home in the CX space, vendors like Nuance with Nina, Genesys with Kate, Microsoft with Cortana, Apple with Siri, Amazon with Alexa, and Google are applying artificial intelligence to attempt to provide improved CX on a massive scale. Customer service bots are enabling conversational automated assistance with more human-like personalized interactions. Natural language understanding (NLU), intent mapping, and emotion sensing are key AI applications in this space.
AI at Cyara
Cyara is investing in building a strong in-house capability in AI and machine learning. Firstly, we’re working on understanding the new AI technologies being applied in the market, allowing us to better build frameworks targeted at monitoring and validating the operation of these AI-driven customer service platforms. Within Cyara, we’re also building solutions that utilize machine learning algorithms for natural language recognition and computer vision with the aim of allowing us to provide more realistic synthetic customer and agent activity to improve our CX and AX assurance capabilities.
The next big technical obstacle within the deep learning space is the voracious appetite these algorithms have for training data. As the leading provider of CX monitoring solutions worldwide, Cyara is in a unique position to utilize the massive data we analyze on a daily basis around the globe to develop CX assurance tools trained on the most diverse set of real CX data.
The O’Reilly Conference provided a brief insight into much of what is going on in the AI world today, and a considerable part of that work is on building better customer service solutions. Cyara is using its experience in CX assurance, and combining it with AI technologies to ensure these AI-driven customer service solutions continue to drive improved CX. As AI bots become more prevalent in all areas of our industry, Cyara remains focused on that drive, and in the future, may help prevent your organization’s “Skynet moment” before it ever happens!
