Companies should create a chatbot readiness checklist to help define a bot design strategy.
Chatbots have taken the world by storm with their ability to take insanely complex technology and make tasks look incredibly simple. Think of the ease of asking Amazon Echo for your bank balance or requesting a ride via Facebook Messenger. No wonder many businesses are eager to roll out chatbots.
In the excitement to move forward and achieve business value quickly, however, the organizational readiness that bots require is often not fully appreciated. Unfortunately, this can lead to implementation surprises.
So what do organizations need to prepare for when considering a chatbot implementation? The readiness runway includes the following key areas.
Identification and prioritization of business value
Organizations should determine the value of deploying a bot, to both the intended end-users as well as the organization itself. This should be done at the transaction (or “use case”) level, as this will help prioritize necessary bot features. Delivering real value will play a key part in keeping users engaged with bots on an ongoing basis.
This includes analysis of whether certain interactions should be rolled out in the bot channel at all. For example, a wealth management client wanted to build a financial data access bot for their advisors on Amazon Alexa. One interaction that was ultimately discovered to not fit very well was the delivery of detailed client transaction data. The user experience of reading a typical statement would have been much better as opposed to having a bot read them out one by one.
Given the restrictions on UI design imposed by the available platforms and devices, the user experience factor becomes critical in bot design.
The platforms have begun realizing this too. For example, FB Messenger’s recent update is an implicit admission that user interactions driven purely by the current state of NLP technology might be leading to less than optimal user experience. Another critical area is in the design considerations for the use of bots by people with disabilities.
Finally, perfecting the user experience during user authentication is a critical design consideration for voice-based chatbots like Alexa. This is an area often taken for granted.
Ease of discovery and hand holding
For all the hype and attention around bots, they are still a new channel. It will take some time for users to become more comfortable with devices/platforms and their capabilities, what bots are available and the many functions that bots can perform.
It is critical that organizations make this initial user journey as smooth as possible. Helping users discover bots easily is likely to play a large part in their rapid adoption. The bots that are hard to find and navigate don’t get used! Organizations can take a page from the communication strategies that they used to generate awareness about their mobile apps a few years earlier. The availability of high quality user-help/demo content can also help smooth out those critical initial interactions.
Chatbots do not work in “data isolation” – for most user interactions, they will have to access (typically via APIs) some other application, internal or external, in order to get or update data.
However, organizational systems might not be in shape to fulfill this requirement. Many legacy platforms are not API-enabled. For those that do make services available, organizations should make sure their services strategy is mature enough to handle these new requests from bots.
For a financial services client, we built a bot that integrated with their newly purchased BI platform. This platform provides users a first class user experience through its UI. However, the bot development team was quickly disappointed to discover that its API offering significantly lagged its UI capabilities. This led to a course correction to integrate their bot with another, more mature reporting platform that had a better services architecture. Now they can roll out multiple interactions through their bots, quickly refine these interactions and put them into production.
Bot platform capabilities
It’s important to track the stable and beta capabilities of bot platforms & devices. Since the platforms are so new, there is a significant amount of flux in their features and user experience, which is expected to continue in the near term.
For example, a media platform company wanted to make their videos available on the recently released Amazon Echo Show device. We had recently built an audio playback feature via Alexa Echo/Dot devices for them, and their assumption was that video playback would just be an extension of this. We then discovered that the playback controls that are available to users for these videos are not as refined as those that the platform has made available for audio playback.
Companies should create an “Organizational Bot Readiness” checklist to help take a multi-dimensional look at their chatbot requirements and build a bot design strategy that is likely to work best. Of course, as bot platforms and capabilities evolve, this checklist will evolve too.
Think of it as a way to learn the language that the bots speak in order to better chat with them.
Rajesh Kamath heads Incubation and Solutions for the Financial Services business unit at Incedo Inc. a California-based technology services firm specializing in data management, product engineering and emerging technologies. With nearly 17 years of experience in technology, mainly in financial services, Rajesh and his team help clients innovate in emerging technologies like big data, data automation, text analytics, machine learning, chatbots and predictive modeling.
The InformationWeek community brings together IT practitioners and industry experts with IT advice, education, and opinions. We strive to highlight technology executives and subject matter experts and use their knowledge and experiences to help our audience of IT ... View Full Bio
We welcome your comments on this topic on our social media channels, or [contact us directly] with questions about the site.