Without personalised attention, customer loyalty is hard to come by today. 63%1 of customers expect personalisation as a basic standard of service. Anytime, anywhere access to an unlimited array of choices coupled with low switching costs, makes it easy for customers to switch to brands who provide a more delightful, personalised experience. One-third2 of shoppers, of all ages, are quick to abandon brands that don’t meet their expectations. Yet, the customer experience capabilities and mindset of many companies have failed to keep pace. They continue to use legacy technologies that are unable to scale and provide personalised attention to customers, threatening their competitiveness.
WHY use chatbots: Understanding the drivers
Today, instant gratification has become a customer expectation – the norm rather than a luxury. Customers expect rapid responses to their queries and concerns, and chatbots allow businesses to meet this expectation.
60% of millennials, for instance, regularly use chatbots and over 70% of those users report a positive experience. A cohort of over 83 million individuals, millennials account for a quarter of the US population and spend over $600 billion annually. Independent and tech-savvy, they will account for 75% of the workforce by 2025. Brands must proactively cater to the preferences of this increasingly important, discerning customer segment with significant spending power or risk losing market share.
Well-designed chatbots are a win-win for customers as well as businesses. Customers enjoy benefits such as 24-hour service and instant, personalized responses while companies stand to realise the following outcomes:
- Cost savings from chatbot powered self-service, reducing the number of agents required for live support.
- Better call handle times with chatbot augmented agent support.
- Expansion of self-service options across channels.
- Increased revenue through improved online conversion.
WHERE to use chatbots: Assessing use cases
The State of Chatbots Report asked users, “What do you predict you would use a chatbot for?” The top use cases listed below reflect the top frustrations reported by customers.
- Getting a quick answer in an emergency – 37%
- Resolving a complaint – 35%
- Getting detailed explanations – 35%
- Finding a human customer service assistant – 34%
- Making a reservation for a service – 33%
A chatbot is a computer program capable of simulating human conversations. The interaction takes the form of text messages, voice, or both – all in natural language. The Chatbot Use Case Framework in Table1 highlights some simple use cases to help you get started.
|Use Case||Overarching Intent||Sample Customer Query||Typical Entities (i.e. data types involved)||Chatbot Actions|
|Answer frequently asked questions (FAQs)||Find answers quickly||How can I upgrade my current mobile phone plan?||Customer ID, current plan, upgrade options||Lookup the current plan for the customer and display upgrade options.|
|Capture and qualify sales leads||Improve sales conversion||I’m looking to buy health insurance.||Age, gender, occupation, medical history, i.e. the dataset needed for health insurance underwriting||Collect the needed data, send the data to the underwriting process, and display health insurance options in real time.|
|Create and manage customer service requests||Resolve a high percentage of service requests efficiently – without live agent support||I need to schedule an appointment to resolve an issue with my utilities.||Customer ID, preferred appointment time||Look up the customer ID, for plan, location, payment status and display appointment time choices.|
Table 1: Indicative chatbot use case framework
The use cases identified in Table 1 are only indicative. Chatbots can be deployed to achieve a variety of customer service goals – from getting a quick answer in an emergency, resolving customer problems and getting detailed answers to finding a human customer service assistant, making reservations, upgrading plans, and providing suggestions for next best actions.
HOW to use chatbots: Starting with a pilot
Making the most of your investments in chatbots requires a strategic well-planned approach to their deployment. It is recommended that businesses start small, with a small pilot to build business confidence, and progressively identify more mature and sophisticated use cases over time for larger impact. Here are three things to do when kick-starting your pilot:
1. Ask the right questions
Constitute a multi-disciplinary team and answer the following questions – contextually and in detail.
- Which business metric are you trying to impact?
- What are your primary use cases and user personas?
- Which conversational interface will suit these user personas?
- What knowledge sources will the chatbot access. Are they adequate?
- Do you have conversation maps for the chatbot to use?
- What legal, privacy norms must you adhere to?
Mail your responses to Firstsource experts and receive a complimentary session designed to help you develop a custom-built, actionable roadmap based on your chatbot priorities.
2. Match specialists to the building blocks
The top-level architectural components of a chatbot typically include User Interface, Dialogue Management, various APIs, Data Sciences, Knowledge Base, Analytics, Infrastructure, Security and Privacy. It’s important to find the right-fit talent for each of these tracks to ensure integrated development of all components – right from the get-go.
3. Develop key metrics
Not only is it crucial to identify relevant metrics across key areas (see Table 2) but also monitor them contextually. For instance, a long chatbot-customer interaction could either mean an engaging customer conversation or that the chatbot struggled to provide the right answer quickly – depending on the context.
|User Metrics||Chatbot metrics||Business goals|
|Total||Number of conversation starter messages where a user starts an interaction with the bot.||Uplift Net Promoter|
|Users||CNumber of bot messages per chat session.||Score (NPS)|
|Active||Number of messages the chatbot is unable to process.||Lower repeat contacts with live agents.|
|Users||Total conversations per day.||Improve cost avoidance|
|New||Goal Completion Rate as a percentage of total conversations.||Increase sales|
|Users||Fall Back Rate – percentage of times the chatbot failed or experienced a near-failure situation.||Conversions|
Table 2: Identifying relevant metrics
Once you have successfully implemented a pilot, the next step is to leverage the right tools to evaluate your progress. Check out our other blog on chatbot strategy framework and maturity model – ‘3 tools to evaluate your progress on chatbots‘ to learn more about how to assess your chatbot journey success.