In the previous blog, we discussed the trends which are changing the way consumers prefer to communicate and interact with businesses and services and the rapid evolution of customer engagement. These trends are briefly summarized below:-
- Mobile consumers are spending most of their time on messaging platforms. Messaging platforms have seen tremendous growth in active number of users.
- Bi-directional and synchronous communication is becoming mainstream.
- Messaging and voice-based interfaces offer greatly streamlined and frictionless interaction.
It is a no-brainer really, to see that brands which want to engage customers in conversations and push customer acquisition, will need tools and technology leaning heavily on Machine Learning enabled, natural language processing capable agents. You guessed it right, we are talking about platforms which can enable Chatbots & services and integrated in messaging platforms, voice assistance platforms, apps, websites, etc. It can ultimately turn out to be a hybrid approach of delivering some use cases (Repetitive and Robotic) solely by bots, while in other areas bots do what they are good at doing (prediction), and leave the judgment to humans.
In this blog, we tackle the questions of use cases and how we should go about building those.
Chatbots and Voice Assistants are a Medium
The idea which we should keep in our mind is that our end goal is not just to have a medium. The goal should be to have a medium, which delivers customer experience to strengthen the relationship. In this regard, KPIs which treat customer relationships as an asset like Customer Lifetime Value or Net Promoter Score will appropriately measure the impact of these touchpoints. Chatbots are particularly suitable to capture and calculate Net Promotor Score (shares or mentions, invitations to a conversation). Having this goal makes it easier to wade through the myriad possibilities and select low hanging fruits to deliver high impact, revitalizing customer engagement.
Chatbots and voice assistants have usability in customer acquisition, different stages of customer purchase journey, usage and customer service.
- Chatbot living in any messaging platform has rich access to customer’s profile. Compared to a user browsing anonymously for the first time, you already know them slightly better. This provides opportunity to influence the first contact and convert into a purchase by using personalized recommendation of products and offers which make sense for them.
- Success of messaging platforms lies in making it useful and effortless for the user, aiding customer acquisition. For example, mention of “Taxi” leads to popup bubble for Uber. Particular keywords can be used as data detectors for suggesting a bot living on platform.
- Natural invite to a conversation is probably the most important discussion in acquisition, as it relies on your brand ambassadors who are trusted by potential customer.
Purchase is a decision making process for the consumer, which exists along a continuum from very easy to complex. We can equate the complexity of decision to cognitive effort required on customer's end. In this regard, looking at stages of customer purchase process still remains a good way to analyse the cognitive effort and then design the experience accordingly. Following picture depicts elements involved in determining evaluation intensity for a customer:-
Cognitive effort required for a particular purchase will have bearing on:-
- What parts of decision making can happen on voice or chat?
- Will customer need to switch context (i.e. go from chat to app or website)?
- Can you detect and provide stimuli?
|Stages of Purchase Journey||Chat Bot||Voice Assistant|
|Seek||Suitable to provide personalized external stimuli||Initiating unwarranted conversation on a home device or phone is against all good customer experience principles|
|Discover||Suitable for short-tail catalogue, low involvement, positive attitude buying aided by visual-rich cards, AI-assisted discovery||Limited applicability as verbal communication requires too much information to be kept in working memory. Stepping through the options can be a tiring experience|
|Evaluate||Cumbersome, but can be solved creatively using same semantics as asking a friend or expert for opinion. Ask bot to get an expert join the conversation. Great applicability in augmenting physical retail experience||Evaluation of alternatives on voice only medium without a superior general purpose AI will impair customer experience|
|Purchase||Highly suitable to go through with decision and complete the transaction||Suitable to go through with decision and complete the transaction. It can enable you to push the shopping cart which consumer has prepared but not completed, or order something from the Wish-list|
|Engage||Feedback, share with your contacts/friends. Possible to engage on the platform of their choice with 1:1 personalized communications||Possible to get feedback for recently placed orders|
To summarize the above discussion in broad rules of thumb:-
- Chatbots have a clear edge over voice medium when it comes to purchase.
- With current set of technology, chatbots seem to be less suitable for high cognitive effort and purchases with high number of alternatives.
- Chatbots are highly suitable to covert personalized offers and recommendations into purchases, provided accuracy of personalization and prediction, high amount of trust and brand loyalty.
Chatbots can be used to deliver FAQs, structured troubleshooting and more rich information about the product on fingertips of customers. For subscription and services business, it could mean providing billing & balance related information, usage alerts, etc.
Customer service was almost an afterthought some years ago but is now essential to customer engagement. Chatbots have profound application in providing customer service. Unlike contact centres where expansion comes with significant cost implications, upscaling chatbot-based customer service will come at very less incremental cost. Delivered through cloud-hosted platforms, it can scale quickly in high trading periods and scale down when things are quieter. Consider following use cases which can be delivered through voice assistants or chatbots:-
- Guided visual response in chatbots: Chatbots can provide structured response which is visual, by use of rich cards.
- End of IVRS in voice: Earlier, interactive voice response was necessary to route to appropriate agent or quick informational tasks, as technology to process voice into required intent was not available readily. This can now be eliminated and customer’s intent can be directly connected to fulfilment like “Ask BrandX when is my order arriving”.
- Escalation: A mechanism to resolve exceptional and unusual cases should be built in the bot/voice functionality.
Discovery & Interaction with Physical
Chatbots have use cases in product discovery and integration of physical with online. Chatbots can be used to provide in-store assistance with product showcase, deliver real-time offers, and facilitate consulting with experts about products.
Discovery of bot or voice service: There are open questions around bot’s discoverability on the platform. There can be multiple approaches like should bots be suggested organically by platform, based on data detectors, context, etc. There could be keyword-led promotion like search engines. The best option will be to get introduced by an already connected member. This space is still developing and messaging platforms will play similar role which app stores play now. Voice assistant apps also have similar discovery of skill/service issues as Alexa, Cortana & Google are the interpreters that link user’s intent to the fulfilling service or skill.
Right technology: Technology involved in developing conversational service is still evolving and space is fragmented with no clear market leader. There is obviously a fear of locking in a technology which doesn’t survive as the standard, which needs to be mitigated. Therefore, clear selection and architectural patterns which simplify making changes later, should be the guiding principles. Delivery of Chatbot capability depends on four core technologies which should be present in the toolkit – Natural Language Processing, AI, Micro-services, and PaaS.
Time to market: Toolkits which are available now (Microsoft bot framework, IBM Watson, Api.ai to name a few) allow quick prototyping and get a functional Chatbot up quickly. With usage, toolkits are gaining domain knowledge which becomes instantly available to use. Organizations can adopt a “test & learn” approach of devising a quick pilot with beta users and assess ROI. You can incrementally add the capabilities which your customers are going to like. Option to leverage bot framework living in cloud with NLP, AI abstracted allows you to build bots for the organization with clear separation of concerns.
Budget and skills: The enabling platforms and services are available with cloud hosting, which means administration and maintaining the core intelligence services are taken care of by the provider. Talent which is required will be of digital specialist or experienced product managers to figure out what features and functionalities should be piloted and how it is integrated with rest of the landscape.