To Bot or Not to Bot: 3 Tips for Getting It Right

August 4, 2021

To bot or not to bot-04

70% of customer interactions will use emerging technologies like chatbots, machine learning (ML) and mobile messaging by 2022, according to Gartner. As customer journeys continue to move increasingly online, the digital experience a company delivers within and across channels can make or break a brand.

In recent years there’s been an enormous push across industries to integrate tools and technology into the everyday lives of consumers. Some of these tools, like virtual assistants (think Alexa), have been well-received and became mainstream in a fairly short period of time. Others, including automated chatbots, have taken longer to be embraced by the masses. What’s driving the difference? Perhaps it’s because not all automated interactions are alike.

Automation in Customer Support

Customer support is filled with repetitive, administrative and data-related tasks so it makes sense that companies look for ways to offload these activities from human resources. The industry experienced a huge uptick in chatbot implementations across customer journey stages and interaction touchpoints. But not all implementations were designed properly which has led to customer frustrations and some brand backlash.

Companies that dive into automation without first understanding customer expectations for automated interactions are risking poor customer experiences. That’s why many are finding success using a hybrid approach. This method combines chatbot software and live agents, working together to solve customer queries. Using automation and humans to seamlessly deliver customer support is an important strategy for any company. Here are three ways to utilize chatbot applications in your support environment.

When to Bot… or Not

Traditional chatbots were designed using decision trees. These rules-based workflows connect to your tech stack and provide quick resolution for less complicated questions. It’s this type of simple, repetitive, fact-based interaction that is tailormade for automated technology. When data and rules are not complex, pre-defined scripting can be developed that meets the needs of multiple consumers posing a similar question. These bots can be used to prioritize and route customers based on an established decision engine, and since they’re available 24/7 they’re an effective strategy for lowering human-based resource investments around the clock.

More sophisticated chatbot applications, including those that use machine learning, can be launched but then require ongoing tuning. These advanced approaches are flexible and dynamic to meet changing inputs. They also require data-driven insights on an ongoing basis to keep the software constantly tightening rules and messaging. Although the technology is equipped to conduct some self-training, it’s more common to have hierarchical and other rules designed, and updated, by a developer to ensure the software optimizes over time. Often referred to as “independent chatbots” due to their ability to self-train, many business have learned the hard way that failing to apply ongoing human intervention to update taxonomies quickly leads to an outdated chatbot.

Combine People and Automation

On their own merit, chatbots are useful in many ways. From directing customers to self-help to tackling simple queries to understanding requests, they effectively assist customers with basic tasks. However, there are limits to what a chatbot can resolve. While it’s true those that employ machine learning, analytics and language processing tactics are considered more intelligent, with any chatbot there will be a limit to what it’s able to do. As customer issue complexity increases, it’s not uncommon for chat automation to max out its ability to fix a problem. In these instances, the best option is a seamless transfer to a live agent who can further consider both the problem plus ways to resolve it.

Intelligent chatbots can identify when a live agent is needed and implement a seamless handoff. Information captured upstream is transferred to the agent so customers don’t need to re-deliver insights already described earlier. This arguably is the most crucial part of a bot-agent hybrid strategy.

Don’t Hide Your Bot

Bots must introduce themselves as a bot early in an interaction to demonstrate transparency and build confidence. Attempts to disguise a bot as a live person often have disastrous consequences leading to disgruntled customers and occasionally making issues worse. Your customers deserve to know when they’re connected with a bot, not a person.

There are pros and cons to this approach, but ultimately it builds trust with your customers. Customers want to feel like they’re being helped, and recent studies have shown they tend to react negatively when an issue can’t be resolved and later learned they’ve been communicating with a bot. Interestingly, those same customers tend to react positively to the disclosure that they’re talking to a bot – even when the issue couldn’t be resolved. A straightforward message early in the dialogue, like “Hi! Thanks for contacting [Company]. I’m a virtual agent, how can I help you?” will not only ease the customer’s mind, it’ll also allow both them and the bot to stay focused on resolving the question at hand.

In Conclusion

Chatbots are a crucial, cost-effective way to automatically manage customer interactions and deflect the simplest queries away from live-agent resources. Going a step further, a hybrid bot plus agent model is a necessary solution in environments with the high potential for complex inquiries. Blending automation with people can have a large, positive impact on the best ways to service customers.

Subscribe to our blog for more thought leadership