How to Get Started With Conversational AI
Let's take a look at how to get started with conversational AI and explore what can be achieved with conversational AI software.
Join the DZone community and get the full member experience.Join For Free
An ever-expanding list of benefits and a growing demand for voice interfaces has placed Conversational AI high on the list as a key component for any digital transformation strategy. With everyone from industry analysts and the Board recommending investment, the next question is how and where do you start with conversational AI?
There is no doubt that conversational AI will be the defining technology of the next decade. From improving customer service and driving online sales revenue, to new ways of differentiation using voice interfaces, the speed with which conversational AI is being embraced is breathtaking. According to Gartner by 2020, 25% of customer service and support operations will integrate virtual customer assistant technology across engagement channels, up from less than 2% in 2017.
As the demand for conversational AI intensifies, the discussion is no longer why should I implement? But how do I start?
That's a complex question to answer. The conversational AI technology available is staggering, but only a few of the platforms on the market today meet the stringent demands of enterprise use. There is also the issue of limitation. For example, if your experience of conversational AI has so far been constrained by the technology used, then it's likely your vision of what it can do for your company is also stilted.
Rather than curb your vision, map out what you'd like to achieve then look for the technology that will enable it - the end result is far more likely to deliver greater returns.
What Can Be Achieved With Conversational AI Software?
To help you get started, we're introducing a series that looks at the basics of conversational AI from the business benefits to the nuts and bolts of starting your first project. You'll get to understand what's achievable in realistic timescales and within budget.
At Artificial Solutions, we've been building multi-lingual conversational systems for global enterprises for many years. In this series we've combined the knowledge of our research teams and engineers, with the experience of our advisors that are already helping our customers realize their second and third generation conversational AI projects.
We're going to take you behind the scenes of conversational AI and delve into the challenges and pitfalls, discuss the many different ways data can be used, look at how enterprises are scaling up their applications to include additional language and device support, as well as taking a look at what the future holds and how enterprises can prepare for it today.
We'll answer some of the most frequently asked questions on conversational AI, such as do I need a chatbot? (The answer may surprise you!) And consider other issues such as integration, data privacy and security.
Delivering the Key Must-Haves of a Frictionless Experience
To give you inspiration, we'll take a peek at some of the projects we're working on at the moment and a deeper look at conversational AI applications already in use, from the initial vision to the benefits they've delivered.
Gartner states that by 2019, half of major commerce companies and retailers with online stores will have redesigned their commerce sites to accommodate voice searches and voice navigation. According to Google, 20 percent of searches in their search engine already occur through the voice assistant. In a world where speed and convenience are fast overtaking price and brand loyalty, a conversational AI interface delivers the key must-haves of the frictionless experience — effortless interaction over any channel, personalized content, and intelligent, humanlike conversation.
As a consumer, ask yourself, would you go to the retailer where their interface only understands "I'd like a red dress"? Or to the one that not only interprets your utterance of "I'm looking for a red dress like the one that Jennifer Lawrence wore to the Oscars" but knows your size and that you always buy shoes to go with an outfit hence recommends those too plus lets you know delivery options, all in the same virtual breath?
In a nutshell that's the difference between conversational AI and the voice recognition that most vendors offer. Conversational AI has the capabilities to understand the user, deliver personalized content on the fly using information gleaned from the conversation, known preferences, integration with back-end and third-party databases and more, to deliver the response the user expects.
Integration with back end systems and third-party databases is an essential part of delivering value in conversational AI. There's nothing more frustrating than being told you need a service engineer to fix a problem but you need to switch to a different channel to schedule a visit. Particularly when as the customer, you know that telephoning to book an appointment involves an interminable call-center menu. If the virtual assistant is good enough to diagnose a problem, it should be intelligent enough to arrange the service appointment too.
It perhaps also highlights the need to map out what you want to achieve before limiting your vision by the constraints of the technology you chose to build with.
Extending Business Benefits With Conversational AI
But the benefits to the enterprise go deeper than for example, just selling an extra pair of shoes. Every conversation can be analyzed, even when complying with data privacy regulations. This information provides a wealth of knowledge that can be used to personalize the conversation even further while improving your bottom line through greater understanding of trends, weak points, and new growth opportunities.
For one large telecommunications company, the single action of consistently delivering the right contact information across all of its customer touchpoints reduced its call transfer rate of 80% to almost zero.
We'll talk more about the benefits and how they are realized in the next post, but customer engagement is high on the list. Why? Because fully engaged customers represent a 23% premium in terms of share of wallet, profitability, revenue, and relationship growth.
Increasing Customer Engagement With Conversational AI
Mention conversational AI to most people and chatbots immediately spring to mind. But chatbots are only the beginning of your conversational journey. Indeed, many enterprises are now skipping this step and using conversational AI to reach a wider remit across multiple touchpoints within the brand experience. Think of conversational AI as the evolution of chatbots.
In this way, they can improve customer engagement from the very first initial interaction right through to nurturing the ongoing relationship, even providing the connection in the final product such as home automation or in-car entertainment. It's an essential step in increasing customer loyalty in a landscape that is becoming highly competitive as customers start to choose speed and convenience over price.
By looking beyond a point solution to solve a particular problem such as resetting a password, and implementing a digital assistant that can actually interact with the user through the entire customer journey from the initial contact to after sales care, these organizations are moving closer to delivering a frictionless experience to their customers.
Kindred Futures, for example, are passionate about creating real value for its customers. It's new Teneo based chatbot bypasses click-button or key-word driven clinical responses, which struggle to cope with the breadth and depth of a regular conversation, to not only understand the complexities and nuances of gambling language but also learn and adapt in order to get better answers faster.
Launched on the Unibet Facebook page, the conversation-based interface enables customers to do more; from finding odds and placing bets, to asking about fixtures and getting help. Users are also able to connect to a Customer Service agent to escalate issues and questions.
Where to Start Your AI journey
If you're a global company, it might be tempting to start your first implementation in your most productive market. But choosing a region or aspect that isn't performing to expectations can yield better results. These areas offer low risk but potentially give higher returns. It's easier to iron out the kinks with a smaller subset of users and then scale up ready to roll out globally once you've proven your application performs to yours and your customers' expectations.
Planning a series of sprints towards your end goal is also a good idea. It allows you to garner customer feedback at each stage, to broaden the scope of the project if required and provides much-needed proof points to ensure internal stakeholders are as enthusiastic of the development as your customers.
Choosing the right development technology is critical from several different standpoints. First on the list is ease of use and capability. All too frequently one normally excludes the other, leaving an enterprise either with a chatbot that does very little or still trying to achieve their vision of the conversational AI application with no end in sight several months on. They will, however, have spent a vast amount of resources on specialist developers, training data and duplication of efforts with each new channel or language supported.
Data Ownership Is Key in a Conversational AI Platform
Which brings us to the thorny issue of data ownership. With the large majority of technology vendors, there is no need to worry because the chances are that you won't even have had access to most of your conversations to know what you were missing in the first place. In addition, unless you have the right tools in place, data privacy laws may prevent you from gathering and using the information anyway, which can negate the whole point of having a conversational AI interface in the first place.
It's also important to consider how you might want to develop your conversational AI systems in the future. As pointed out recently by Will Mace, Head of Kindred Futures, there are plenty of Manchester United fans in China that will never get to Old Trafford, but a virtual reality experience has possibilities. You can discover more about Kindred's journey with AI and conversational systems in this video. Widiba, Italy's premier online bank, has already launched its virtual reality branch where customers can walk around, talk to an advisor or access information via screens.
Both of these organizations are prime examples of how building on a platform allows for conversational AI systems to be expanded easily to new languages, devices, and services. Many projects we're involved in are going beyond their original remit as organizations start to understand the exciting potential of conversational AI. And not just as the next phase of development, but into new business areas, such as the IT Helpdesk or HR. Again, this is where a flexible platform approach to development comes into its own by enabling enterprises to integrate with other AI assets already developed.
As consumer demand and expectations grow, the impact of conversational AI will be felt in every aspect of an organization's business. At the same time enterprises are challenged by a rapidly maturing technology, a lack of resources, and a wealth of misinformation.
Published at DZone with permission of Andy Peart, DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.
Mastering Go-Templates in Ansible With Jinja2
Using Render Log Streams to Log to Papertrail
AI and Cybersecurity Protecting Against Emerging Threats
Revolutionizing Algorithmic Trading: The Power of Reinforcement Learning