5 reasons customer service chatbots must be part of your omnichannel strategy
I was talking with the contact center directors of a large financial institution lately and they admitted they were having a hard time convincing their various lines of business that chatbot initiatives should be driven by the contact center people. Chatbots (or conversational agents) are rather new to them and they are lacking solid arguments to make their case. In fact, ownership of the whole omnichannel customer experience is moving from the IT people towards the marketing team.
The problem is, chatbots are so hot these days, platform vendors are pushing so hard at the exec level, everyone wants to play with them. PoC initiatives are conducted in isolation in the LoBs or by the marketing department. They can’t wait for the IT team to get budgets approved, projects properly defined and planned, resources allocated, etc. And that’s totally understandable.
For the sake of this post, let’s distinguish the two types of chatbots we find in most enterprises: marketing chatbots and customer service bots. Marketing chatbots are mainly used to create awareness and build the brand. They are a great way to engage conversation with potential customers. They can even drive sales. Customer service bots, on the other hand, offer automated services over a textual channel. They can support customers with their most common tasks (think password reset, subscription renewal, etc.), answer frequently asked questions (opening hours, branch locations, …), and more.
In all cases, customer experience must be a paramount concern. You only have one chance to make a good first impression. At least, for marketing chatbots, the stakes are not high. You may get some funny stories on social media. But at least people will talk about you! You’ll prefer good press to bad one, but at worst it will be seen as a clumsy attempt to build something cute and won’t damage your brand.
But for customer service bots, it’s a different story. We all know that a large percentage of customers are willing to change provider after bad customer service. They want stellar service. But at the same time, we know that only 30 percent of transactions are completed on Facebook Messenger-powered bots. Your chances at failing to deliver great service are pretty high!
The only way you can truly deliver a great customer experience in this context is to make sure you properly integrate your customer service bot in your omnichannel strategy. And here are 5 arguments supporting this.
Visual UX design expertise is not conversational UX expertise
Most organizations have teams of web / visual designers who know how to design great visual user experiences. But conversational UX requires very different skills. The latter has to deal with the intricacies of natural language, turn-taking, how to properly craft messages to get a set of manageable answers, how to handle errors or miscommunications in the dialogue, etc. Even the simplest chatbots need some conversational expertise to do them right (ever stumbled on a bot that repeats “I don’t understand. Please try again.” ad nauseam without changing its dialog strategy?).
These conversational UI (CUI) skills are most often found in the contact center or the firms they engage to design their IVRs and voice self-service applications. That’s the bread and butter of VUI designers. Yes, text channels come with their fair load of challenges that are not present in voice applications, but voice and text messaging have more in common than, say, chatbots and mobile apps.
Expertise should not be scattered across the organization
Letting each and every department build its own customer service chatbots will inevitably spread the experience and know-how across the whole organization. How will you make sure you follow the same CUI guidelines and standards? That the brand image is consistent? And if you add AI in the mix, it may not make sense to have different teams developing NLP-related machine learning expertise. Better have a single team that can apply expertise, experience, and data across all your LoBs.
Customers expect personalization and proactive notifications
Chatbots are a great opportunity to have a personalized conversation with your customers. I, for one, expect chatbots to know who I am, what I did the last times I used them. They have access to my identity, so they can certainly use it to engage in a more personalized way with me (past the simple “Hello Dominique!” greeting). That means integrating the chatbot with CRM systems and the rest of the contact center infrastructure.
For example, if I buy a product from the chatbot, I should be notified of the delivery status of my order from the messaging interface I used, proactively.
Chatbots need human assistance
Customer service chatbots should not be released in the wild alone. Especially if they are using the latest NLP/AI technologies. They need supervision. Conversations need to be monitored carefully to catch problematic conversations in real time. Because your NLP/AI system will make errors. Guaranteed! There will inevitably be requests or answers the system won’t be able to parse properly. The CUI designer can’t anticipate everything. And real people make mistakes, they are unpredictable.
The good news is, using appropriate supervision tools, your customer can still have a great user experience even in the presence of errors from your system. And, more importantly, your system can learn from these mistakes over time!
But supervision means real people behind the curtain. Trained to handle problematic interactions in an efficient way, using specialized tools. Having a centralized team of chatbot assistants can be much more cost-effective than having separate teams spread across the organization. The assistants can be trained to handle interactions from all the different bots the organization deploys.
The chatbot is only the start of the conversation
You wrote your first e-commerce chatbot. Customers start buying products from it. Great! Now how do you integrate it with your customer support team? If customers have inquiries about their orders, where do they go? Can they use the chatbot to do that? Do your customer service agents have access to the whole conversation history? Can they take over conversations started with the chatbot?
Don’t assume you’ll automate 100% of all your customers’ requests. What will happen with those that fail? You need to think about how you will hand-off the conversation to human customer service agents. You want them to use the tools available for all the other channels.
You also want to use the same reporting and real-time statistics tool. Not new tools, forcing you to develop custom aggregated reports. Ideally, you would use those already provided by the contact center platform.
If you want to truly deliver a delightful customer experience, make sure that all your touchpoints are tightly integrated. This is the promise of the omnichannel contact center. So don’t repeat the mistakes of the past and make chatbots another silo!