Whether it’s the multitude of vendors, poorly defined use cases or the high percentage of failed conversational AI pilots - building and deploying chatbot solutions presents many different challenges for enterprises.
You might be surprised to learn despite many teething problems, conversational AI solutions like chatbots are rapidly capturing hype with businesses around the world.
Chatbots are bots enhanced with a conversational user interface (CUI) lets users interact with applications in a way similar to human-to-human communication. In many cases, interaction is done via text instead of voice; the end-result is up to the particular use-case.
Essentially, chatbots interact with users through natural conversation and recreate the way we interact with customer service to find answers to our problems or queries. There’s endless possibilities here for efficiency and productivity gains, in addition to innovation and competitive advantage.
In this article, we cover 4 best practices to deploy chatbots - so you can avoid the misconceptions, minimise initial adoption challenges and reduce risk of failure.
But first - let’s go into how and why chatbots are the next big thing in enterprise.
The rise of conversational AI and chatbots in 2019
Chatbots are experiencing renewed attention and fast growth is due to major improvements in natural language processing (NLP), defined by Gartner as the ability of machines to map a spoken or written input to an intent.
With increasing maturity and sophistication in natural language services enabled by AI, popular use cases like automating tasks for customer service or providing human resource support response is an increasingly compelling value proposition for businesses undergoing digital transformation - hence the natural rise in investment in chatbot development and deployment.
There is a real opportunity to leverage AI to improve customer service, user support and knowledge management, which have all driven more than a 106% increase in interest in chatbots this year as per Gartner.
Other key takeaways from the latest research on conversational AI this year:
- By the end of 2019, approximately 40% of enterprises will actively use chatbots to facilitate processes using natural-language processing for interactions.
- New research from MIT Sloan says almost 85% of global executives believe AI will enable competitive advantage and play a bigger role in improving workplace productivity.
- 80% of global businesses want chatbots by the year 2020, according to Oracle.
Chatbots today are being developed and deployed across SMBs and enterprises to improve customer service and business workflows with self-service options and seamless automation that reduces the time it takes to perform generally mundane tasks, with the desired result generally being better customer service and more satisfied customers and end-users.
However, implementation and deployment of enterprise chatbots comes with complex challenges. Here are the major considerations to analyse and assess prior to deploying your own chatbot, as recommended by the top global research firms.
#1 - Develop and tailor chatbots around current expectations
The biggest roadblock and challenge enterprises face when approaching conversational AI and chatbot solutions is the age old question, “Where is the killer app?”
It’s a fairly expected question to ask, and is often accompanied by “What’s the best AI platform?”
The problem is in the current digital transformation landscape, there are very few AI platforms out right now with out-of-the-box applications to showcase chatbots in a standardised way. Businesses often must rely on in-house development and models to get investment on their down and drive their own innovation, but the one major thing Gartner recommends is to keep focused on is deploying chatbots that align to today's current expectations of the technology and build up experience.
Current expectations of chatbots are visible, customer-centric tools that act as the representative of a company’s customer service experience - and carry a broad appeal of a tool that can heavily benefit people’s daily tasks via automation and efficient answers for generic queries.
Focus on deploying chatbots that champion these expectations and worry less about being on the cutting edge of the space, especially if you currently lack specialised expertise across data science and machine learning to create your chatbots in-house and train data.
Leverage pre-built chatbots from external, cloud-based third-party providers that specialise in data preparation and training to build and host chatbots initially, and use the development time there as a springboard for future initiatives as consumer expectations naturally evolve and the technology becomes more self-service.
These conversational AI platform vendors and their pre built models and prepackaged training data are very useful to meet current use cases. Leveraging these preconfigured solutions is an alternative, more efficient path to chatbot deployment that requires less work from the enterprise, but can also help keep you focused on what consumers want from them now.
#2 - Define the strategic fundamentals - goals, users, platforms
Chatbots don’t just integrate into the business without proper planning, preparation and strategy.
Your business must take the time necessary, like with all digital transformation projects, to define and organise the team leaders and key players who will be driving the success of both the setup and deployment of chatbots through the business. This includes those handling the technical implementation, the design, and the marketing towards potential customers.
Business goals must align with your bot initiative to ensure everyone is on board with what it can offer, and to avoid becoming another failed pilot. Define the overall objective of deploying your chatbot, what you want to achieve, why you need a chatbot to do it, and an examination of the current conversational AI platforms on the market to help begin your development.
But it’s not just about getting your internal strategy right. You must also tailor your chatbot initiative around your potential users and the devices they will engage with it on; this includes the persona of your users to build accurate responses into your bot, which requires extensive assessment of several factors (age, gender, job role, language, voice, platform, etc).
Ultimately it’s not about how cutting-edge your chatbot is, but a back-to-basics evaluation of how chatbots can help further your business goals, and how you can focus its functions on your user’s needs and wants to stand out from the competition.
#3 - Personalise your chatbot to represent your brand
Chatbots need to be tailored around your brand identity and tone of messaging.
The customer experience when engaging with chatbots must be enhanced by these helpful conversational AI tools, not dampened. From its name to the specific way it answers customer questions, bots are a vital new representative of any company - and should be treated as such.
A big part of crafting a relatable, effective chatbot personality is the voice; if it’s too robotic, it may not accurately represent your brand image and turn potential prospects away, rather than entice them to engage with the technology and walk away impressed with your implementation.
For example, research from 10Pearls on virtual assistant use studied participants who were asked to interact with and rate four popular voice-heavy chatbots - Amazon Alexa, Hound, Google Assistant and Siri - based on accuracy of answers and personal preference.
The study found that while Alexa was seen as the least accurate of the group, it was the most preferred (35%) because “she seemed nicer” and they “liked” Alexa better.
Customers anthropomorphising our chatbots and virtual assistants with human-like qualities shouldn’t be seen as strange, but rather a natural evolution of what we seek from our business queries and customer service experiences. Studies such as 10Pearls’ only further highlight why it’s key to imbue chatbots with emotion, personality and tone - particularly if voice-based.
It’s key to think of chatbots as your brand’s front-facing avatar or mascot, but it’s also important to remember they need to represent your brand beyond their look or design. When deployed for customer-facing use cases such as customer support on a company’s external website, ensure your conversational AI persona embodies your organisation’s values and represents your business competence with clear, effective communication with difficult to predict customers.
#4 - Prepare and leverage voice-heavy chatbots
Entry-level chatbot solutions in their current market iteration are mostly based on written input due to the inherent complexities and challenges that using voice presents - background noise, poor-quality audio signals, and so on.
However, firms like Gartner are seeing an increasing demand for voice-based chatbots, and recommend enterprises focus on this next wave rather than remain conversative with the text-based iterations.
Amazon Alexa, Google Home and Apple’s Siri are three significant examples of voice-based chatbot interactions across customer service experiences and home-assisted AI today, despite their fairly limited capabilities. But the popularity of natural speech to interact with our devices and businesses we need support from is proving to be key drivers for voice-heavy chatbots.
Gartner notes that while also being more appealing, voice chatbots generally take a lot less effort for customers and end-users to get what they want than typing, lending to its appeal and only highlighting the importance of preparing for next-wave voice-heavy chatbots now rather than later.
How to prepare for conversational AI and chatbots: Next steps
Deploy chatbots solutions based on compelling use cases for today's customer, namely to deliver exceptional customer experiences.
Adopting conversational AI in any business is an evolving journey best guided by sufficient planning and expertise. Chatbots and natural language processing (NLP) especially requires a certain level of business maturity to get right. Our team can help evaluate your organisational readiness for adopting AI, across people, process, technology and data.