Enterprise Bot Adoption: 5 Expert Tips that Make Enterprise Bots Excel
Enterprise bots are alive and kicking! There are a plethora of terms that are being used in this space but one thing we recognize is that enterprise bot adoption is soaring.
Bots are being planned or implemented by many different industries and across a broad range of use cases, both B2E and B2C.
The Gartner Hype Cycle: Stages of Adoption and Maturity
The Hype Cycle, made famous and branded by Gartner, is a way of looking at the different stages many technologies go through as they emerge and mature.
Gartner identified five overlapping stages in eloquent terms as follows:
- Technology Trigger: This is the early emergent stage when technology is conceptualized. There may be prototypes but there are often no functional products or market studies. The potential spurs media interest and sometimes proof-of-concept demonstrations.
- Peak of Inflated Expectations: The technology is implemented, especially by early adopters. There is a lot of publicity about both successful and unsuccessful implementations.
- Trough of Disillusionment: Flaws and failures lead to some disappointment in the technology. Some producers are unsuccessful or drop their products. Continued investments in other producers are contingent upon addressing problems successfully.
- Slope of Enlightenment: The technology’s potential for further applications becomes more broadly understood and an increasing number of companies implement or test it in their environments. Some producers create further generations of products.
- Plateau of Productivity: The technology becomes widely implemented; its place in the market and its applications are well-understood. Standards arise for evaluating technology providers.
The pace at which different technologies (or businesses) cycle through these phases differs widely but it’s still an interesting way to view and understand the dynamics behind mass-market adoption or even technology extinction. While AI bots are still nowhere near maturity in many enterprises, I believe that the hype cycle tool from Gartner sheds some interesting light on where businesses currently are in their adoption of conversational AI and natural language solutions. It may even be an indicator of where we may be heading in terms of wide-scale and more sophisticated business implementations of chatbot platforms.
The Emergence of Chatbots
While many articles date the advent of the chatbot back 50 years, the real rise of chatbots has become a business reality more recently. From WeChat launching their chatbot ten years ago to Facebook, Google, and Microsoft opening their messaging platforms to chatbot developers in 2016, the frenzy has accelerated in the last 3 years as the technology trigger was ignited. Startup bot development and conversational AI companies emerged in many different forms. Some targeted the booming bot developer community while others honed in on specific point solutions or single-purpose customer service bots, like the virtual assistants that act as basic FAQ ambassadors.
With companies like Google, Facebook, Amazon, and IBM launching conversational products and services, the media hype quickly exploded and expectations soared. Chatbots were set to conquer the world of customer service and more, resolving all our queries day and night with ease and personality. Ultimately bots were on a fast path to displacing human agents in contact centers as well as the need for human interactions across many other operational departments. So while the shiny new future of a chatbot world was highly publicized so too was the fear factor around their future.
But despite the naysayers and the doomsday headlines warning that AI bots would replace the human workforce, the market for bots began to take off. During 2016 and 2017 many businesses plunged headlong into the promises of natural language and bots. By 2017, the chatbot market had begun to attract large investments with an estimated 180 bot startup ventures attracting $24 bn in funding and more AI bot companies emerging on their heels.
Enterprise Bots: From Over-Hyped to Realistic and Achievable Expectations
Despite the flurry of venture capital activity around bots, 2017 didn’t live up to the expected rise of the enterprise chatbot. Some embarrassing and epic chatbot failures were reported causing some household big brands to pull them immediately. The visibility of these failures often outweighed the success stories around AI bots, subsequently leading to frustration and disappointment. The trough of disillusionment was in full swing for enterprise chatbots.
However, all was not lost. Despite bots being over-hyped, the general feeling amongst businesses remained optimistic around their potential for increasing operational efficiency and driving new engagement models. Media headlines in late 2017 predicted that 2018 was going to be the year that chatbots join the enterprise. A Forrester Research report, “Predictions 2018: The Honeymoon For AI Is Over,” predicted that “in 2018 we’ll finally move beyond the hype to recognize that AI requires hard work to plan, deploy, and govern it correctly”.
So what changed? Like many over-hyped highly visible technologies, early bot implementations suffered from overpromise, underdeliver syndrome. Businesses expected too much from a game-changing technology that was still being tried, tested and refined. Many thought that AI bots would be able to replace humans overnight and handle any type of customer query without failing. As Forrester rightly stated, “AI requires hard work to plan, deploy, and govern correctly”. For enterprises that are familiar with application development, this hardly comes as a surprise but add to it the complexity of natural language and we can understand why early chatbot implementations failed on many fronts.
The Slope of Enlightenment and the Path to Enterprise Bot Success
This year we have witnessed an increased appetite for bots as learnings from highly-publicized bot failures have been better understood. Here are some of the key aspects of enterprise bot solutions that impact better success rates and adoption:
- Craft a Well-considered Enterprise Bot Strategy: Like any new technology (e.g. mobile, IoT, blockchain), the importance of a well-crafted strategy helps guide success and create a consistent approach to bot development across the organization that involves business departments as well as IT. So rather than jumping headlong into building bots, a strategy will help guide decision-making around what use cases are prioritized (these could be B2C and B2E), what the challenges may be, a rollout and deployment plan, a data integration and access plan, and the nuts and bolts of security and scaling. In a recent blog post on the fundamentals of a chatbot strategy we outlined some of these considerations for you to review.
- Create Multiple Bots, but each with Narrower Missions or Scope: Human conversations are tricky and can take many different twists and turns but Natural Language Processing (NLP) technology is advancing and improving. A chatbot needs to be able to keep track of the conversation and maintain context. Asking too much of a chatbot can lead to failure so it is better to start small and build a bot that can handle fewer intents but be able to execute on them. The bot can then take on additional intents or skills over time. Think of chatbots as workers. They can only handle so many tasks, especially when they are being initiated into the organization. By creating mission-specific bots you can build out a whole team of bots that can be used across different customer or employee journeys.
- Use a Virtual Assistant as your Bot Orchestrator: It’s clear that a strategy based around a single chatbot is not sound so if you do start deploying multiple bots that have different missions you need to consider how these will be architected and orchestrated. This is where a central Virtual Assistant (VA) comes in. The VA manages the conversation with the user and routes to the mission bots according to the intent(s). With this type of VA orchestration, an enterprise can centralize things like sentiment analysis, language detection, escalation mechanisms, small talk and more within the virtual assistant. This then becomes accessible to all bot use cases across different departments creating consistency across different user experiences and journeys. It also enables faster and easier scaling of bots. New bots with different missions can be added and recognized by the VA. Here are some deeper insights into our concept of the Virtual Assistant as the bot orchestrator.
- Enhance the Context with Data: Many early bots failed to produce results as they had to hand off to human agents very quickly. This was often due to poor natural language understanding and intent detection or to the fact that the bot hit a wall and couldn’t access the information that the user requested. An enterprise environment has its own systems, processes, and even language. Authenticated customers or employees that interact with a bot present an opportunity for the bot to access their information from different systems and personalize the experience based on their data. Integration with different systems that provide this context and that enable the bot to fill any information gaps (aka. slot filling) rather than requesting it from the user, goes a long way to enhancing the user experience.
- Security is Paramount: There are many security considerations that need to be taken into account when you implement enterprise bots,. Where will conversations be stored? Will sensitive data in conversations be redacted or encrypted? How do you handle disaster recovery and user authentication? These are important aspects that need to be considered to protect the organization from serious security breaches. We are still at the early stages of chatbots in the enterprise but imagine the headlines if some sensitive customer conversations were hacked?
Has the Age of Enterprise Bots Arrived?
We believe so, based on the conversations we are having and the traction we see in the marketplace. We see companies getting very serious about their chatbot strategy. Some of them are recovering from unsuccessful early forays into the space. They understand the drawbacks of some solutions and the need for enterprise-level platforms and solutions that can help them build a good foundation for success.
Recent market forecasts by Juniper Research indicate strong growth in chatbot interactions from 2.6 billion in 2019 to 22 billion by 2023 where “retailers can expect to cut costs by $439 billion a year in 2023, up from $7 million in 2019, as AI-powered chatbots get more sophisticated at responding to customers.” In Gartner terms, I believe that many businesses have moved from Trough of Disillusionment and are on the early Slope of Enlightenment. There are exciting times ahead!
Here are some additional resources that are related to this article that may be of interest to you.
- Chatbots Infographic: 9 Considerations for your chatbot strategy.
- Download our Template: A template to help guide you through issuing a comprehensive RFP document
Definition of the Gartner Hype Cycle https://whatis.techtarget.com/definition/Gartner-hype-cycle