Chatbots are saving businesses valuable time and money. By 2023, chatbots will handle 40% of all customer interactions. Combine the power of chatbots with the popularity of WhatsApp, and a whole new set of opportunities emerge. So, how can you create a WhatsApp chatbot?
Why create a WhatsApp chatbot?
In 2021, 85% of businesses will compete on customer service alone. As such, improving the customer experience is a top priority. Businesses are learning how important it is to meet customers how and when they prefer, and it's no surprise that WhatsApp is at the center of the conversation.
WhatsApp is the most popular messaging app in the world. With over 2 billion monthly active users and 100 billion messages sent every day, there's no denying the messaging giant's power.
Naturally, businesses are eager to use WhatsApp to communicate with customers. While small businesses can use WhatsApp without automation, larger companies often require more. They need solutions that can help them reach their customers at scale.
With that in mind, WhatsApp released the WhatsApp Business API. This allows larger businesses to build on the standard messaging experience that WhatsApp offers. Now, you can create a WhatsApp chatbot and the tools you need to deliver great customer service.
Who can create a WhatsApp bot?
Not everyone has access to the WhatsApp API, which means that not everyone can create a WhatsApp chatbot. WhatsApp only works with a handful of WhatsApp Business Solution Providers (BSPs) worldwide. In fact, working with an unauthorized provider is a violation of WhatsApp’s terms of service.
At Hubtype, we’re proud to be one of the few BSPs that WhatsApp trusts. We've helped brands like Levis, Caixa Bank, Volkswagen, Michael Kors, EPSON, Guess, and other household names develop WhatsApp chatbots.
Steps to create a WhatsApp chatbot:
Step 1: Request access to the API
The first step is to request access to WhatsApp's API. Your application must go through an official BSP (like Hubtype). Fill out this form, and we will reach out to you if you are eligible to use the API.
To date, we have seen the majority of companies accepted by WhatsApp are multi-national or enterprise-level companies.
If you’re a small or medium-sized business, it’s still premature to jump on board, but you can apply through Hubtype now, and if you aren’t approved immediately we will notify you when the API becomes more broadly available.
Given the popularity of WhatsApp, we receive many applications from all over the world. This means it might take a while for your application to get approved, so it’s best to apply early.
In the meantime, we encourage you to talk with us about how you can use chatbots on WebChat and other messaging apps.
Step 2: Set up a business account
After you've been accepted through a BSP, the first step is to create a WhatsApp business account and set up a line of credit. This account will allow people to easily identify your business and find important information about your company.
Step 3: Define the use case
How exactly will your WhatsApp chatbot help your customers? Across the entire path to purchase, a significant majority of people engage in messaging with businesses. You'll need to consider your audience's unique needs in the awareness, consideration, purchase, and care phases.
Step 4: Strategically introduce automation
After you've defined your use case, get started with Minimum Viable Automation (MVA). This is just enough automation to make things easier, not more complicated. The two most common types of chatbots are rule-based and AI chatbots.
In general, decision tree chatbots are a good fit for the majority of our clients. Decision trees help you define the pathway of conversation to guide potential customers to a specific goal.
If you're not sure which type of automation is right for you, read: Rule-Based Chatbots vs. AI Chatbots: Key Differences.
Decision trees are like flowcharts; they begin with a single question and then serve the user follow-up questions based on the previous choices. They are excellent for solving specific and well-defined use cases. It’s best to never have more than 5 layers of decisions, nor more than 3 choices at any level.
Keywords can also be used to trigger a guided conversational flow. Start with specific keywords that are relevant to your use case. You can always add more later.
Other things to consider when building a WhatsApp chatbot
Manage customer expectations
Your WhatsApp chatbot should build trust and loyalty. Don't try to fool users. Make it clear when a customer is talking to a bot and when they're talking to a human.
If the customer experience is excellent, it won't matter whether or not they are being helped by a human or a bot. Most customers are happy to solve their query faster, or eliminate extra steps, even if it means their interaction is with a bot.
Being transparent and authentic about this builds trust and shows that you respect your customers.
Replace text with graphics where possible
The human brain processes images faster than text. So, choose visual elements to replace text where appropriate. Take advantage of buttons, quick replies, carousel selectors, and list selections to require less work from your customers.
It is easier to select a date on a calendar view than to write out a date in text. If there are two choices in a decision branch, two buttons work better than requiring the words to be written out.
Remember that experiences have evolved beyond text-only conversations. A good WhatsApp chatbot uses calendars, buttons, carousels, time selectors, and other rich elements.
Using these features significantly speeds up a massive number of processes for a customer, and increases the chances of a customer completing a goal.
Plan for the human handover
A clear and direct path to human escalation is necessary when using automation. Common handoffs include human agent handoff (seamless from bot to live-chat) or a phone number to reach a live agent.
When a human agent can be reached for complex cases, and without too many failures of the bot prior to reaching an agent, your customer satisfaction ratings will continue to trend positively. Focus on a seamless transition from bot to human agent and back again. Prioritize immediate responses and low friction.
Note that specifically for WhatsApp chatbots, direct human escalation is a requirement in all user-initiated messages when a business chooses to use automation. Businesses should have a direct and prompt human escalation path available for customers in chat threads, whether it is a live agent, an email address, or a phone number.
Have an out of office hours strategy
A WhatsApp chatbot helps provide service to customers 24/7. If your human agents won't be available for a handoff, make sure you have a backup plan, and at minimum clearly communicate the operating hours of the channel to the customer.
Make sure to collect opt-ins
In order to send proactive pushes, customers must opt-in to messaging on each specific platform such as WhatsApp and Messenger. Ask users to opt-in on your website during a transaction, or to update their preferences in their account profiles.
Start fast, scale later
Analyze how customers use the channel, then adjust accordingly. Learn from the limitations of your WhatsApp chatbot, and expand based on where you're seeing the most opportunity.
For example, if 70% of people ask for tracking information, this is a good area to integrate with your shipping platform. Just don't lose sight of your original use case, and make sure that the potential actions your customer can take are clear.
Use the right agents to handle WhatsApp communication
Dedicate full-time agents that specialize in written communication. Agents that talk to customers via messenger must possess specific skills. Depending on your brand's tone of voice, you may want your agents to use emojis or GIFs.
A good practice is to develop a conversational style guide or brand guidelines for your agents. Developing tools like this will help you onboard new hires and empower current employees.
Combine messaging with APIs to deliver exceptional customer experiences. Open-source frameworks like Hubtype’s allow for robust integrations between messaging platforms and AI, NLP and NLU providers, CRMs, APIs, analytics, and other business tools.
The result of linking a WhatsApp chatbot to other tools is hyper-personalized experiences. Customer details and purchase history can be saved to deliver comprehensive conversational experiences that beat even some of the best customer service representatives.