5 Things To Think About Before Adding a Chatbot

From e-commerce and retail to financial services and everything in between, companies are turning to chatbots to provide agents with added support and to improve customer satisfaction. With new research showing intelligent chatbots are now table stakes, consumer expectations are rising and companies that aren't adopting the latest technologies risk being left behind.

You might have already decided to invest in a chatbot, but choosing the right chatbot for your business is no small task. Of course, you want to find a solution that saves you time or improves your customer experience, but it can be a daunting task to identify where to start.

Here are five questions you can discuss with your team to start the process:

No. 1: What are our goals and use cases?

By determining your business priorities on the front end, you'll be able to better identify the right chatbot solution that will work best in the long term. To home in on your specific business needs, ask the following:

  • Do we need to increase deflection and reduce service tickets?
  • Do our customer service metrics need improvement?
  • Could agent productivity be enhanced? What about average handle time?
  • Do we need to scale service and support operations?
  • Could we reduce costs by implementing cost-effective service automation?
  • Do our customers need 24/7 support options?
  • Does our customer satisfaction score need a boost?
  • How can we receive better online reviews?
  • Do we need additional insight about customer issues and challenges? Could these insights improve our products and processes?

No. 2: What technical resources do we have?

Are you considering buying a chatbot or building your own? The answer depends primarily on your team's technical needs and bandwidth. If you lack the internal expertise or time required to build and integrate a complete chatbot and automation solution from scratch (note: you're not alone), there are commercially available solutions from reputable vendors that are probably your best bet. However, if you have plenty of engineering resources, time, and budget to create a system unique to your company, then you could go the route of building a chatbot.

No. 3: Where do we need to support our customers?

Do you only need a chatbot for your website? Or do you get frequent service requests from social channels like Facebook Messenger or WhatsApp? Perhaps you have a separate mobile app? If you want a consistent experience that works across devices and platforms to answer comprehensive questions before a live agent becomes involved, then you'll need to ensure you go with an omnichannel chatbot solution.

With your priorities determined, it's time to explore three types of customer service automation: bot-building frameworks, platform providers, and chatbot applications.

  • Bot-building Frameworks: A bot-building framework integration takes nine to 12 months, but so too does a business' customizability. These frameworks (such as Microsoft Bot Framework, Dialogflow from Google, Wit.ai, and IBM Watson) provide tools for businesses to build, test, deploy, and manage chatbots. While these frameworks are comprehensive and allow businesses to create fully customizable solutions, they require in-house engineering and AI expertise, application development, and maintenance capabilities. In terms of initial development and ongoing maintenance, this is likely the costliest of chatbot options.
  • Platform Providers: With platform providers you'll get the simplicity and convenience of an embedded solution that's part of a larger platform. Most of the CRM/helpdesk platforms (such as Salesforce, Zendesk, Oracle, etc.) have some built-in chatbot or automation capabilities that can handle straightforward chatbot use cases. Deployment traditionally takes four to six months with some professional service or integration costs. Many of these platforms lack advanced AI and machine-learning capabilities or aren't designed specifically for advanced support use cases, so lots of manual training and maintenance is required. Most CRM platforms integrate seamlessly with support-specific chatbot applications to address these concerns.
  • Chatbot Applications: When businesses want an automation solution that can be customized but doesn't require the cost and resources of building a fully custom, home-grown solution, this is the way to go. These chatbot applications include low- or -no-code automation and can be customized without the high cost and resources of other solutions. Chatbot applications can also be deployed quickly, in one to three months. It's important to note that there are a number of vendors in this category that deliver significant variances when it comes to intelligence (AI, machine learning, and natural language understanding technology), the time it takes to train, deploy and maintain the chatbot, as well as the customizability of the solution.

No. 4: What kind of conversation experience do we want?

There are two types of chatbots: rules-based and intelligent AI. So, what's the difference?

Rules-based chatbots depend on a series of If X, then Y, statements that are manually built and maintained. AI chatbots, by contrast, use machine learning and natural language understanding (NLU) technology to understand the intent behind customer statements or questions and respond appropriately.

Rules-based chatbots provide businesses with a good solution when the customer issues it needs to handle are simple. Since answers need to be built out for every type of question and answer combination, the biggest limitation of these chatbots is the time and effort needed to build and maintain them.

Next-generation, AI-powered chatbots require far less effort to build and train. Why? Because they can understand questions however they're worded and find a common answer to deliver without setup from your team. They also learn and improve with each interaction due to the machine learning technology that's built in, so they get better with minimal maintenance.

No. 5: What are we thinking in terms of total cost of ownership?

This means considering not only the costs of deploying a chatbot but also those associated with scalability and maintenance—we're talking the lifetime costs of owning a chatbot.

With chatbot deployment, you want to keep expectations realistic. The main costs associated with chatbot deployment will include: the creation of content for the chatbots to use, the collection of training data for machine learning, the building and testing of the chatbot rules and responses, and integrating the chatbot with existing business applications and processes. This also includes the development time required to customize and personalize your chatbot.

One of the biggest surprises new chatbot owners encounter is the time and cost it takes to manage and maintain these systems. These costs will vary greatly depending on the solution you've chosen. The good news? If you choose wisely, the investment in a support chatbot is well worth the effort.

There are plenty of factors to consider when choosing the right chatbot for your business, but remember that each question asked and answered gets your business closer to finding the perfect solution for your business and your customers.

Bob Grohs is director of marketing at Solvvy.