Conversational Interfaces Will Punch UI Right In The Face
Or, “Current Benefits & Challenges of Conversational Interfaces (CI) over traditional User Interfaces (UI).”
Abstract:
The advent of chatbots introduces the dilemma of conversational interfaces [CI]. How do these compare to traditional UI? What are current benefits and challenges in successfully approaching CI for business chatbots? And why is it key to think beyond technology when developing CI? Let’s uncover some of these topics together. There is a real opportunity for customer engagement via truly thought out CI or Conversational Experiences [CX] and businesses that take this seriously will greatly benefit from it.
A little bit of background:
When thinking of bots (virtual assistants) in 2014~2015, I was more concerned about the integration project as a whole: How would the chat box fit design-wise into the webpage where it would live? Should it be an “avatar” or the company logo?
How much knowledge would the bot have? Should I import the company’s current FAQ and take it from there? What type of formulation of the same customer request should it handle?
Was it able to answer social questions or compliments like “what time is it, what are you doing tonight, which city are you based in or can I have your number?”
Other considerations were what would trigger an escalation to an agent? What channel would it be able to escalate to: live-chat, voice (mainly those 2 back then), what additional channels [today]?
Back then, I was not fully focused on the chatbot as a “ Conversational User Interface”. I wasn’t necessarily thinking of it as a tool that needed to deliver a full fledge conversation. Today it’s a whole different ball game. While thinking about the above points remains valid, one needs to pay close attention to how to create a natural flow between the bot and the user and even making the conversation fun.
Defining User & Conversational Interfaces:
Before we jump into the nitty gritty let’s take a step back and try to define what stands as UI as well as CI. It is important to point out that both share the same goal: which is to provide, retrieve and display the [most] ‘useful’ information that will answer or solve the user’s need, request or problem.
At a high level:
A User Interface a GUI [Graphical UI] in which users find the information they need pointing and clicking, navigating on the company’s application. We are most familiar with GUI. For the more techies also a CLI [Command Line Interface] in which the users retrieve the information they need through ‘programatic’ commands and the system/app responds to it by displaying the relevant results.
A Conversational Interface is by definition a much smoother experience as it leverages the most human characteristic to achieve the same goal: human conversation.
Dear User, “come as you are” and get [the same or] better results by simply writing [or speaking] as you would do in IRL [in real life].
Benefits or why I believe CI offers a better experience:
Today, I believe CI presents several benefits from traditional UI. Indeed, traditional UI tend to present many elements [headlines, side panels, buttons and menus] which can turn the focus away from the user’s initial quest.
Let’s illustrate this by an example or use case:
You now need a mortgage and need to find out and understand somewhat quickly what types of mortgage options would be available for you at your bank. Well, the moment you log into it you’ll likely be distracted by a number of things: Current Balance, New Promo Advertising, or Unread Messages some of which may be generic or from your financial adviser. Does it sound familiar?
Conversational Interface: an experience focused from the get-go
CI offers a great opportunity to qualify and take the customer through the exact journey s-he’s looking for. And because humans are different but at the same time alike, CI gives companies the ability to project a “unique experience” for each and everyone of us. A crucial point in nesting an intimate or relevant experience is leveraging all customer context including previous live chat transcript, purchasing history, CSAT or NPS scores, recommended products or services, you name it.
Typically, the information presented above is usually coming from several disparate systems. Or, luckily it could be living in your CRM. So if your Chatbot-Builder-Software allows you to [easily] leverage [that is integrate] all of these ingredients in a unified configuration-console, then you’re golden and will be able to tailor a custom experience for each user.
There are Cost optimisation benefits for companies using CI: reducing cost by deflecting interactions to channels typically managed by humans is a classic example. And, if you’d like to take this topic beyond: What about [new] companies only developing CI to interact with customers? No more web or mobile apps or traditional engagement channels? Wouldn’t that optimise your budget for customer engagement? #foodForThought
Conversational Interfaces & Audience Maturity:
Another “benefit” or rather a reason as to why CI works well today is due to audience maturity [vs. 2014~2015 for example]: 1.3B users on Facebook Messenger. The audience is used to chat conversations.
Finally CI allows companies to extend their brand and services where their customers are very likely to be: on Facebook Messenger, Slack, Telegram, Skype, WhatsApp and What-Not.
Challenges today in building Conversational Interfaces:
This was all very nice but as experience shows: nothing comes easy. The Chatbot arena is no different and there are a number of challenges yet to tackle to deliver the smoothest experience.
Users expect a lot from Chatbots (& AI in general). We expect Chatbots to be like a “digital stewardess” for products or services: to somewhat know exactly what we need, how to solve our problem, and even bring us snacks. Ok, not the last part.
And while NLP + NLU (Natural Language Processing & Understanding) are totally here: these technologies have yet to become “fully” human-proof . But this, is happening by the hour! The main NLP & NLU challenge is that these projects remain expensive. From my experience in previous Natural Language Voice endeavours: one needed to consider 50k$~100k$ to build, on top of training the system for 3~6months with human agents, which incurs additional costs.
ITR (Interactive Text Response) or Flow-Oriented / Scripted chatbot conversations can do the trick. BUT, if not conceived thoughtfully [from the beginning], these projects can become quickly very tedious. In particular, when NOT understanding or having the answer for the user.
Finally, another main challenge is integration with the companies’ customer/user data.
Solutions or how to approach Conversational Interfaces:
Luckily there’s always a solution to each of these challenges and the near future holds smarter problem-solving capabilities by the day. Let’s list a handful of key points I believe necessary in building a Conversational Experience for your Chatbot today:
- A bot is a bot. Clearly state that your company’s bot is what it is: not a human.
- You’re bot may not have all the answers and if that’s the case it will/should happily escalate to a human.
- Leverage both a “flow based” approach spiced up with “language intent recognition” technology if you can.
- Consider that CI is a project on its own and that you need to treat it as such. Define its scope, plan it, plot it, test it, launch it, assess its ongoing results constantly and re-iterate it over and over again.
- Writing is for writers and Coding for developers: what this means is that while you want you’re chatbot to be operational in multiple platforms and you’ll get the right Dev Team or Solution for this, the same applies for the most important part of your CI: the conversational aspect.
- When thinking about your CI: narrow it down to the simplest way possible which essentially means taking your user/customer from point A (Mortgage info) to point B (getting the mortgage from your very service).
- Once this is clear imagine and write all plausible scenarios that could take a user from point A to point B. All potential dialogues should be created.
All the above introduces an important ingredient to the recipe:
The importance of a Chatbot “screenplay” writer or Conversational Experience Developer. This is a must! because building a killer CI is a team effort combining: tech, business, and conversation/data-driven experts to jam together!
Text conversations can quickly become boring:
Spiced it up and mix it with an extremely focused selection of photos, videos, forms, carousels. Wait, did you just say this comes from the traditional UI world that we so desperately want to go away from? You’re right! That said, bringing the best of that world won’t harm you’re CI, on the contrary!
Integration with 3rd party systems is always a pain but there are solutions out there to programatically or “wysiwyg-ally” feed your chatbot with your Customers’ Context/Data from 3rd party system. This is crucial. So Make it happen.
Finally, do leverage your Analytics, a Test & Learn routine including AB-Testing, and combining methods that allow you to compare with other channels’ performance.
A few Conversational Interfaces before the end:
Retail: use a chatbot to change the address of your order delivery.
Car Leasing: leverage a chatbot to generate qualified leads for your business.
Education: offer a chatbot experience to cleverly engage students in learning.
Journalism: how a conversational article experience can blow your mind.
Finance: what if your bank offered the smartest CX?
[IMO &] In A Nutshell:
CI offer a better experience than traditional UI as it allows companies to focus the user’s experience exactly on his/her concern/request/problem. Because users are (a) ‘messenger-matured’ and (b) able to “naturally” get the information they need in a conversational fashion: CI may offer higher conversions than traditional UI from lead generation, to higher satisfaction-rates for customer service as well.
Although CI presents a number of challenges today in particular regarding NLU and 3rd party integration, there are great solutions and methods out there that can get your company in the chabots for business major leagues.
It is a great moment for companies to embrace CI but a caveat in its success is thinking about it as team project that must combine tech, business, and conversation/data-driven experts. The future only holds better and better CI and companies are going to benefit.
Thank you. If you found this article interesting, please 👏 below, I’d really appreciate it. You can also follow me here on Medium, and/or on Twitter @lecharleslozano
