January 11, 2017

Conversational interfaces and Next Generation of Design Thinking

In the beginning of Internet there was this idea that if we share the information via your website, customers will be enchanted. This served as means to build a basic online presence and brought customers to the web from various other channels. The next small steps involved imitating offline processes (or workflows) on the web — i.e. product marketing, knowledge sharing, customer support, etc. Companies got a huge bump in ROI when they started engaging customers through their digital channel. This later evolved into a concept known as Omni-Channel where customer can be in touch with the company agnostic to the channels but be helped with common service level, voice & tone.

Then problem started surfacing, helping people through the conventional Brick & Mortar or via Call Center was very different when they came from Website. They expected quick and precise response. The psychology of these users was very different — they needed answers “right”, they cared less about the “brand”.

We all remember the “call-transfers” and not getting the answers after a long wait time. It seemed that going back to store was better experience.

Companies could not figure out — the result was disaster — most of the companies who got Omni Channel investments did not materialize the same ROI because end users were more frustrated. In most of my consulting engagements (especially with Call Center Research with Sprint & SallieMae) I found that people did not want to talk to machine or the person on the other side. They did not trust the answers via call center.

This led to the strategy — which now has become a de-facto — “Divert The Call” [Sprint 2008). The whole idea about serving customers was to engage them with self service and ask each others for help. P2P, or social recommendation became the new mantra. Most of the companies now provide a very good self-service, community and chat to help customers get what they want. The call-center transformed into handling the questions from the community sites and moderate them.

Today if you go to any BigCo’s website, you will find the feature of “community”. There are plenty of tools that can do the job fairly easily. But the problem is, if you are one of those sites which existed since few years, chances are that you will find lots of competing answers. Or, no answers matching your problem. This is same as having a website will lots of content, irrelevant content.

Key Technology & User Trends

SEARCH: There has also been tremendous progress on search engines. Most of the answres to common consumers’ problem is found on Google and other search engines. For most of the users that is the starting point. Search engines know more about user’s life style and needs than any individual company. They also process natural languages much more elegantly than if you have to build the solution out all by yourself.

MOBILE: Most of the searches also emirates from mobile devices now.  That means that users does not have time for research and is seeking quick response that is also succinct — Micro-content as against the whole big articles.

MACHINE LEARNING: One of the most important development in information science is application of Artificial Intelligence and Machine Learning. Search bots and machine algorithms can learn quickly then humans. This has revolutionized the ideas of recommendation engines.

VOICE BASED INPUT DEVICES: Mobile devices has brought one important trends — voice enabled interfaces. Voice enabled interfaces are important part of life because it provides us a lean-back experience with technology (Amazon’s Alexa, et al.) and undistracted (often efficient) means to communicate to the machine (Car Talk, Siri on MacBook Pro).

Welcome to Chat-bots

The evolution of web, technology and consumer trends are propelling Chat-bots to feed into Virtual Assistant role. So that you appreciate Chat-bots, consider following thoughts —

  • Consumers do not want to make a call to call center. Trusting the human voice/recommendations is the issue, also most of time is wasted with qualifying questions that are too basic and annoying. Customer believe that they should use Call Center as a last resort.
  • The access of information is changing rapidly. And most of the users are accessing information far away from their laptop, and work stations. Often in transit, or while driving. In not so far away in future, people may not be reading, searching and browsing website like they do today.
  • If you have been evolving your business on the web, last few years of experience will tell that you have  lots of content. Your editors can’t keep it fresh and your consumers can’t figure out which one is what. And if you are in a regulated industry, this problem is even more damning.
  • Improving on search and IA on your site has diminishing returns — that is, it will improve the engagement metrics only fractionally (unless your site has not been designed since last century). Rather making the content fungible, open to newer devices, search agents, crawlers are going to pay off more.

If you can comprehend the above points, applying Chat-bot idea to your context may be wise. There are several very good examples from variety of use cases. (P2P Insurance), (Customer Engagement).
iPhone App: Pana (Travel), Operator, Merzi (Shopping)

Chat-bots implementation is a design problem to begin with. Machine learning, AI, NLP are technology parts that makes the magic happen and are very important. But there are many solutions and frameworks that do that if one has solved the design problem.

In order to solve the design problem, one needs to comprehend the context of your customer engagement. Develop a good Taxonomy and apply modularity into content strategy. Build the publishing organization to be agile and nimble to configure the unanswered scenarios. Chat-bots should help envision the omni-channel dreams way beyond the existing norms. And finally, chat-bots should be considered a way towards the future — towards voice enabled, or gesture based interfaces.

Chat-bots will give your editorial/operational staff a very quick learning. For example, the old norm of having the articles rated is good to have. But to understand which part in the user journey is disjointed will be quickly noticed with well designed chat-box implementation.

For sometime I considered Chat-bots as a fad. But no more. I am taking this ideas as an amazing opportunity to take the online engagement to another level.