Why your customer support needs to adapt
In today’s business world, customers demand more than ever before. They want 24/7, personalized, and immediate service – regardless of their purchase value or period of customer loyalty. Consider this:
- American Express found that 64% of customers say access to customer support is very important to them when considering doing business with a company.
- A study by NewVoiceMedia found that 35% of customers would never use a company again after just one bad support experience and that 62% of customers have switched to a competitor following a bad experience.
And with the surging popularity of digital messaging channels and social media for all forms of commerce, customers now expect businesses to engage via their phones – much the same way they would chat with their friends. However, this can get costly for businesses with multilingual customer bases spread across time zones.
Proto’s regulator and operator clients – from the financial services to healthcare industries – are often running contact centres in at least three languages. As these organizations scale, the cost of support agent turnover & salaries, and customer wait times start to really bite.
Today, the traditional methods of customer support simply aren’t sustainable.
The question is how can businesses improve their support operations to meet these customer expectations – without breaking the bank?
The answer is customer support automation and chatbots. But not all chatbots are created equally – and most lack capability for the local and mixed languages that customers across the emerging world use every day.
How multilingual chatbots work
Customer support automation is the use of technology to resolve customer queries without the need for human agent intervention.
Chatbots provide 24/7 support across unlimited customers simultaneously. They never get tired, quit their job, or make customers wait.
There are many chatbot solutions in the market, generally within two categories:
- Rule-based chatbots
- Artificial intelligence (AI) chatbots
These are offered by vendors that specialize in chatbots, live chat, or natural language processing (NLP) engines.
Rule-based chatbots are the simplest type of chatbot, using a fixed decision tree to determine how to respond to a customer query. You can easily spot rule-based chatbots as their responses tend to be short and lack the conversational flow of a human conversation.
At their most basic, these chatbots ask customers to choose from a list of options. At their most advanced, these chatbots search for keywords, such as ‘credit card’ from a customer query such as: "How can I cancel my credit card". However, if a rule-based chatbot encounters any text outside of its pre-defined decision tree, such as local & mixed languages, slang, or misspellings, it will show confusion and ask customers to repeat themselves at the risk of causing annoyance
Rule-based chatbots are good for very simple tasks, such as FAQs but they are ineffective, at best – and harmful to your brand, at worst – when handling free-form customer conversations such as payments, complaints, and sales.
In contrast, artificial intelligence chatbots use a core technology called natural language processing to improve the understanding of customer queries over time. AI chatbots that are trained on large NLP datasets can also engage in small talk and conduct multi-step conversations with the appearance of human intelligence.
Similar to rule-based chatbots, AI chatbots also use a decision tree to determine how to respond to customer queries. However, these chatbots are continuously retrained as they collect conversational data to feed into their core NLP engine. Over time, this improves the accuracy of the chatbots’ response to complex customer queries.
How to get started with chatbots
Businesses looking to get started with chatbots can either build their own solution or use a platform from chatbot, live chat, or NLP vendors.
Chatbot platforms are almost always a more cost-effective solution than building a chatbot from scratch. Additionally, most chatbot platforms offer a wide range of features and integrations.
Building a chatbot from scratch
Are there scenarios where this makes sense for your business?
Building a chatbot from scratch gives businesses complete control over their chatbot solution, but it requires significant technical expertise and ongoing resource investment.
A business looking to build a chatbot from scratch will require a team of specialists to help design the conversation flow, develop the chatbot and NLP engine for the required languages, and integrate it within the business's systems. Therefore, before deciding to do so, it's worth seriously investigating the costs.
Scenarios, where this may make sense, are when a business’s security policies require 100% internalization of technology systems, or chatbot platforms do not offer must-have features – most often in the emerging worlds, this can be certain local and mixed languages spoken by the majority of customers.
Using a chatbot platform
Chatbot platforms provide a cost-effective alternative as they typically have a low startup cost and offer a software-as-a-service (SaaS) pricing model. The chatbots and NLP models in these platforms were already developed and tested. In the most ideal situations, a chatbot platform will specialize in your industry, with large datasets to constantly retrain and improve the intelligence of your chatbots.
The implementation time will also be much shorter as most chatbot platforms have a scrutinized design and deploy process and a simple user interface for your staff to understand and edit the chatbots.
Choosing a chatbot platform
There are many chatbot platforms available on the market and businesses should consider their specific needs before choosing a provider.
Some factors to keep in mind include:
The chatbot platform should support the languages that are most relevant to your business and customers. Most chatbots are limited to common European languages, with unacceptable performance for languages across the emerging world, such as Cebuano, Hindi, Kinyarwanda, Twi, and Yoruba. This problem is compounded by the mixed-language phenomenon, where two languages are fused within everyday speech, such as the combination of Tagalog and English to form Taglish. Chatbot platforms that build and manage their own NLP engines, such as Proto with its HermesAI™, can deploy chatbots with local and mixed language capability.
2. Local channels ↗
The chatbot platform should offer a wide range of channels so your business can interact with local customers via SMS or their preferred messaging apps.. This is especially important in markets that have skipped the website experience in favor of mobile-based commerce within a handful of messaging apps, such as WhatsApp and Viber.
3. Sample chatbots ↗
Implementation timelines can vary widely across chatbot platforms, so you should consider how long it will take to get your chatbot up and running. Platforms that offer sample chatbots are most likely to offer a rapid deployment capability, as these samples are pre-built with industry dialogue and best practices.
4. Privacy & security
Businesses should make sure that their cloud-based chatbot platform is secure and reliable to ensure data privacy and high availability. Some important items to check are compliance with data privacy regulations such as GDPR, SOC, PCI DSS, business continuity and disaster recovery plan, and penetration tests
5. Ease of use ↗
The chatbot platform should have an intuitive drag-and-drop interface and require minimal training for your team, especially for businesses planning to have non-technical staff manage the chatbot’s dialogue and integrations. Ease of use will also speed up the implementation and maintenance time requirements
6. Contact centre features.
The chatbot platform should offer a wide range of features and integrations that can be used to make the chatbot and your contact centre systems work together. At a minimum, this should include API integrations, customer profile syncing, and case creation.
7. Predictable pricing ↗
The chatbot platform pricing should be transparent with no hidden costs. Make sure to also consider how the platform prices its services - does the price include the number of agents you'll need, how many interactions, how many messages, integrations, etc.
8. Localized customer support.
The chatbot platform should offer a high level of customer support, ideally with staff who are based in your timezone and can speak the same languages as your customers. This is important for quickly reviewing chat history and fixing any problems in the chatbot dialogue.
Chatbot use cases
Now that we have covered the basics, let's explore some use cases for how businesses are using chatbots to automate customer engagement.
Chatbots for customer support
One of the most popular applications for chatbots is customer support. Chatbots can be used to provide immediate answers to common queries and escalate complex problems to human agents.
Most queries received by businesses are repetitive – for our clients, this accounts for 60 to 80% of all customer support requests. These requests should not take time from your human agents – they are best applied to more complex customer cases. Additionally, chatbots can be used to provide 24/7 customer support across timezones – a capability that is difficult and costly to achieve with human agents for even the most well-resourced contact centres.
AI-enabled chatbots can provide a personalized and seamless customer support experience, with the vast majority (80%) of consumers reporting positive experiences with customer support chatbots.
Chatbots for sales & marketing
Beyond reactive customer support, chatbots can proactively engage customers and prospects with personalized messaging at an unlimited scale. These proactive chatbot messages can be targeted according to customer behavior, language, location, and more.
Your business’s marketing team will likely have several ideas for using your chatbot as customer conversion and retention tool, such as:
- Using sneak peek chatbot messages to offer discounts to first-time B2C website visitors, or a free trial to B2B website visitors.
- Sending SMS messages to the phone numbers of crypto or FX website users to alert them to a price drop or buy & sell recommendation.
Chatbots for financial inclusion
Chatbots with local and mixed language capability is a potent tool for including historically-marginalized customers within the financial system. These customers are typically rural residents, women, or members of visible minority groups – and even today, they face social and physical barriers to joining the banked population.
Currently, most financial service providers offer urban-based and in-person processes for account setup, loan applications, payments, and more. Chatbots deployed in local languages and messaging channels are a massive equalizer, especially if they are built with API integrations for completing in-branch tasks.
Chatbots for consumer protection
Chatbots are also effective at processing customer complaints, especially when pertinent data such as complaint category and incident location are buried within other irrelevant or emotional information. This capability is important for emerging market regulators that often face a lack of resources to enforce consumer protection standards across thousands of service providers with constantly evolving product categories.
Currently, most regulators rely on manual processes for complaint category identification, referrals, and follow-ups on overdue complaints. Chatbot platforms with integrated capability for automatic categorization and respondents can free up human agents to focus on complex investigations to the benefit of consumers.
Talk to an expert
Automating your business’s multilingual contact centre is an important but big step to unlocking savings and improving customer experience.
It’s always a good idea to talk with multiple chatbot platforms, gather quotes, and build your own understanding of the capabilities that are essential for your business.
To get started on this journey, book your free demo here with the Proto team.