Dr John Taylor: A Pioneer in AI

JohnTaylor, CEO, Actionai | Business Magazines | CIOlook
JohnTaylor, CEO, Actionai | Business Magazines | CIOlook

AI is sweeping across industries with the potential to change nearly every aspect of the way we do business. Getting to grips with the speed of change and understanding its consequences is central to business’ plans for the future. John Taylor is the CEO of action.ai, one of the most highly awarded technology companies in the field of AI;

John’s passion is to change the way humans interact with computers – reducing the friction in current interfaces by supplementing them with intuitive and intelligent Conversational Interfaces, not chatbots.

John combines his experience as a CEO with a background in strategy consulting and as a Think Tank Director. He also brings academic experience via his doctorate at Oxford University and as a Fellow of the Royal Society of Arts in the UK.

Providing truly Conversational Interfaces unlocks a hugely compelling opportunity for consumers to engage with services on the go, 24/7. Consumers can express themselves naturally, just as they would to a human, and receive instant and meaningful responses. Experiences can be delivered via a smart speaker, like Alexa, via the web, or via a mobile device. The consumer now has access to services 24/7 via an intuitive interface that responds and engages with human-like intelligence.

The business John leads, action.ai, has developed industry-leading technology to empower companies to build both voice and text-based interfaces that are truly conversational. Central to this has been building a world-class engineering team with business partner Dr Richard Tolcher, a leader in the field of computational linguistics.

The Main Drivers for Adoption of AI

AI is a broad area but what differentiates AI when deployed effectively is its ability to change the way organizations scale via automation. One of the most striking opportunities is the ability to transform the way consumers communicate with businesses using automated conversational interfaces, often mis-labelled as ‘chatbots’.

What Pitfalls Need To Be Avoided? 

Consumer and B2B focused businesses are well placed to reap the benefits from adopting AI, particularly in the area of Conversational Interfaces. However, companies have tended to assume that it will be easy to build Conversational Interfaces using the mainstream tools on the market but these tools cannot handle the tremendous complexity of human language. The result is brittle and broken chatbot experiences. It is beguilingly easy to build a prototype, but practically impossible to use the same tools to move to a commercial strength experience. action.ai’s expertise in computational linguistics and its Language as a Service offer addresses these shortcomings. Often the best solutions to a problem are not the most well-known, or the most obvious.

One important factor is to understand the linguistic complexity of a use case before embarking on a project. Seeking support to assess the likely complexity can save a great deal of time when designing a project. It is also important not to boil the ocean of language – the Conversational Interface should be focused on users’ needs rather than trying to be a general personal assistant.

Current implementations, for example, basic voice chatbots available via Alexa, are generating very poor user satisfaction and engagement (user reviews paint a damning picture of most experiences). Users become frustrated and then disengage because they are not understood when they express themselves. They can’t hold a conversation with these chatbots – the most they can do is issue an instruction, and even this often fails.

Supporting language is perhaps the hardest problem in AI but it is solvable with the right tools and partners. It is essential to support the core elements of language including ambiguity and vernacular, and also to be able to distill what is relevant from within often complex user language. Handling this complexity pays off many times over as users can now hold sophisticated conversations that are automated and compelling. The aim is to delight users whilst also reducing the costs of engagement.

Advice for Organisations Embarking On AI Projects

John has worked across several industries both as a CEO and advisor. In these roles he’s seen a lot of successful projects, and also a number of failures.

In the field of AI and Conversational Interfaces, a common mistake is to assume that a basic prototype chatbot will be a good basis for a commercially useful Conversational Interface; experience shows that basic chatbots give no useful insights into user desirability or how their users want to converse. Persistence is critical for business success but continuing with the wrong tools and approach is folly.

Before setting out to build a Conversational Interface it is essential to define the opportunity clearly. Once the business case is clear then choosing the right technology is the key – it determines whether the outcome is a failing chatbot or a compelling and intuitive Conversational Interface.

A properly developed Conversational Interface promises to transform the way consumers communicate by providing automated services with human-like levels of understanding.

Users will embrace Conversational Interfaces but they will not use basic Chatbots because they do not understand what is being asked, and this creates huge frustration. In contrast, engagement with Conversational Interfaces will be very high and persist because of the great benefits for the user from a truly intuitive experience.

The Future

As action.ai continues to grow internationally, it is essential to remain focused on the core assets and differentiation of the business. It’s all too easy in the world of AI to be a generalist given the huge demand for advice. But AI is an area in which detailed knowledge and experience has particular value and therefore action.ai remains focused on its Language as a Service product and helping companies realise the commercial benefits from Conversational Interfaces.