Rudy Dalimunthe ,VP of Customer Excellence, Tokopedia
Digital Transformation, Artificial Intelligence, Chatbot, Robotic Process Automation, and Smart Self-Care have become some of the latest buzzwords in the past 3 years. More people are raising the topics and more companies around the globe have even started to embrace these emerging market trends so as to not get left behind.
However, it is imperative for companies to first understand the bigger picture. Many companies transform digitally merely for cost saving or even worse, cost cutting, as they put customer care as a cost center rather than strategic business unit. In fact, the transformation should actually be grounded upon delivering the best experience, solving problems, and fulfilling the needs of the customers. Unfortunately, many failed to see the blind spot and took the shortcut by creating automation and chatbot to simply shift the interactions from human assisted channels.
The rise of chatbot as a new face of customer service
The ever-changing customer behaviors and trends have spurred the expeditious evolution of customer service. From voice hotline in the 1990s, email channel in the 2000s, web self-service in 2004 to the latest trends of connected omnichannel (bot and live agents), and social media with sales and marketing in 2019.
Driven by growing expectations and high adoption of mobile internet and social media by customers, many companies across industries are inviting their customers into digital and social channels --- and chatbots are gaining popularity and becoming a pivotal part of customer experience nowadays. According to Forrester Research in 2018, 57% of companies globally have already used chatbot, while Global Market Insights states that chatbot market is on the rise in Asia Pacific and keeps getting traction.
Today, chatbot capability has evolved. Some experts classified it into four stages with high involvement of Artificial Intelligence (AI) as it advances:
1. Content-based (scripted) bot: Scripted, keyword-driven, and acts based on manually-crafted rules, difficult to scale since the company has to explicitly update rules as it goes different way.
2. Intent Recognizers: AI & machine learning capabilities are improved by understanding user’s intent, relations between words, and extracting meaning from a request.
3. Virtual Assistant: Able to understand what human is trying to achieve. It goes beyond words; it connects to other systems and leverages user’s profile and insights.
4. Human-like consultant: A strong & human-level AI; it understands contextual factors at human level intelligence.
In Tokopedia, chatbot is positioned in the 2nd stage, which is Intent Recognizers. The chatbot is able to greet customers with small talks, answer frequently asked questions, and take action to solve inquiries. In the near future, we aim for the chatbot to enter the 4th stage as we believe that our agent should only be the advisory hub for the users, while the chatbot will automatically engage and detect the user's issues as they enter. Now, Tokopedia chatbot covers almost 100% of the issues faced by users with more than 8 million interactions per year across channels. Furthermore, our chatbot can deflect 65% of incoming tickets with 3,000% of return on investment (ROI). Non-stop annotation process is done daily by reading fallback tickets and bad CSAT responses, classifying them, and creating new possible solutions if needed.
Recently, Tokopedia has been recognized as the ‘Best Use of Self-Service Technology’, ‘Best Technology Innovation’, and ‘Best in Customer Service’ in 2019 Contact Center World Award by racking up three gold medals for our innovations in chatbots. Prior to that, Forbes listed Tokopedia as one of the 25 companies around the world that demonstrated satisfactory chatbot and omni-channel experience.
As a technology company, we believe that creating an excellent chatbot is fundamental in supporting the evergrowing business with new products being launched every month, resulting in higher interaction on the platform.
No matter what stage you are in now, investing in AI &chatbot requires high commitment from top management considering customer service is rarely considered as a direct road to profitability. Different industries have different velocity in getting their bot to a certain stage. As the magnitude of customer interaction may vary from one another, it takes intensive and high level of interactions with customers of various complexities so that the model can learn in a fast and more accurate manner.
Avoiding the blind spot and misperception in chatbot & digital transformation initiatives
Applying AI & automations could be a little challenging and tricky, indeed. Failed initiatives could hurt company's brand image badly. Hence, my recommendations to have a successful AI, chatbot, and digital transformation initiatives are by avoiding the blind spots such as:
1. Cost reduction are the main and only focus.
We all agree that positive ROI is one of the main company objectives, and in hard times, cost pressure has become a major trigger to start these initiatives. As a result, the company has to launch a half-baked chatbot and push the customers in with no other options. In order to do that better, always start with customer’s problem instead of company’s problem - identify the needs and the key pain points and work backwards to solve the root cause. Sometimes, it only requires a simple rule-based bot, not even an AI, to make things work.
2. Digital = Social Channel?
Many companies try to mix their digital transformations with social engagement and simplify these channels in one single AI and automation strategy. They might see Twitter, Facebook, and Instagram as part of digital channels whereby chatbot is intended to get a major role.
In Tokopedia case, many customers in Indonesia reject bot during their interactions with us in social channels.In fact, they require a more “emotionally connected” human customer service where higher level of engagement can be built and posted publicly.
Through Tokopedia Care, we change the way we do things in social customer relationship management, from reactive to proactive approach. Our Customer Engagement Hub (command center), equipped with social listening tools, allows us to find and respond to relevant conversations on social media, both tagged and untagged posts. These tools enable us to be proactive in solving user's problem before they even know about it. In addition, our engagement with users is also built by producing interactive contents and creatives.
Social channels require different strategy and technology. Social listening tools are highly important in this playing field, as opposed toChatbot & Automations.
3. Digital = Full Online?
We might find some companies have fully digitized all their channels and left no offline channels for customers to interact with customer service representative in person. Nothing is wrong with that. However, for high-context society like many Asian countries, customers can have many implicit messages they do not convey - what is said is not always what they mean to say, and even smart intent recognizer would fail to decipher. This blind spot can be avoided by having faceto-face interaction with the customer regularly. Combining online and offline capabilities in one spot will also work wonders.
There were times when we felt like we have built a good product, predicted the behavior and intent using big data, and ended up finding that some customers come to simply ask, “How to subscribe?”
4. Focus only on CX metrics.
As stated earlier, top management commitment is key, as every customer engagement initiative might not always show a direct road to ROI. However, it is our job as a CX practitioner to stay focused on delivering the initiatives profitably, while focusing on increasing the Net Promoter Score or Customer Satisfaction Score as well. Your Chief Financial Officer (CFO) might need you to convince on how every decrease on these metrics lead to negative ROI, as well as the importance of having stronger commitment to improve the AI and chatbot to further pave a path for revenue generation as a necessary action.
I can always repeat this important message to all CX practitioners: Always start from your customer problems, and not company problems. Be obsessed in providing solution and still meet them in person. Separate social and digital strategy, and make sure your channels deliver the best experience profitably as you’ll need them to convince your Chief Financial Officer to further invest in better AI.