Customer Engagement

Top 4 foolproof ways to build better chatbots

Share this post:

Key Points:
– Bots need to be truly customer-centric, solve a real problem and work reliably across devices, platforms and channels, 24x
– Focus in one channel before expanding to other environments
– Concentrate on successful engagement
– Continuously improve your bots based on user feedback and know when to start over

Learn to build a chatbot in minutes

 

Chatbots are being created to deliver everything from customer service to event registration and more. But creating a successful chatbot requires more than just implementing basic questions and routine answers. Measuring the real success of your bot requires more than just evaluating your sales figures. Bots need to be truly customer-centric, solve a real problem and work reliably across devices, platforms and channels, 24×7.

Here are four ways to make sure your chatbots are the best they can be:

1. Focus on one channel at first

Chatbots can be used in web interfaces, chat apps, and over voice interactions. But while the intelligence that drives the bot may work across multiple environments, optimizing the user interface for each requires very different analysis, approach and skill sets. And when developers attempt to work across too many channels right from the beginning, they can lose focus. It’s easy to miss things if you’re trying to address technical details on every single channel at the same time rather than tracking the results and optimizing them one at a time.

When creating your own chatbot, it is important to not boil the ocean. Keep your bot’s capabilities limited at first so that you can ensure the functionality works well before expanding to additional capabilities. Being overly ambitious could lead to technicalities, slowly down the development of your bot.

2. Concentrate on successful engagement

Many companies develop chatbots to augment the work done by people and to deliver better customer service. When a chatbot is successful it can result in fewer human interactions and higher rates of satisfaction reported by customers. It can take time to determine customer satisfaction, and measurement can be difficult. But counting the numbers of interactions and comparing them over time is relatively simple and can deliver near real-time analysis of the success rate of a chatbot initiative.

“If we can have humans involved in 500 cases per day instead of 1000 cases per day, that’s a huge cost savings for a customer,” says Joe Beninato, founder and CEO of Banter, Inc. “Another measure of success is how the end user feels. If they’re able to get their questions answered quickly without having to wait on hold or navigate a phone tree, that’s a big win for them.”

Objective measurements like the number of human interactions per day and the length of time users spend waiting for support are relatively easy to come by and provide quick and accurate feedback on the chatbot’s success.

3. Continuously improve based on user feedback

The standard methodology for product improvement is to take feedback from customers and learn what works for them and what they would like to see in future versions of a product. Chatbots have an inherent advantage insofar as collecting feedback because the feedback loops can be built into the bot. This alleviates the need for customers to have separate interactions in which they give their thoughts and opinions.

“Chatbots that are successful provide true value, can problem-solve and simplify tasks and encourage engagement from users, pushing the chatbots to better utilize and learn about the various components of the platform they are hosted on,” explains Anurag Lal, CEO and co-founder of Infinite Convergence Solutions. “This is because as users continue to use chatbots, the chatbots become smarter since increased data collection allows them to improve. A chatbot that grows, learns and advances is a good marker for success, too, as it shows that the platform is being used.”

Developers need to conceptualize the key success factors for the chatbots they develop and build feedback mechanisms directly into them. The data derived from the feedback can then be enhanced with machine learning tools that can assess results and iteratively adapt the chatbot to refine its operation.

4. Know when to take s step back and restart

Many companies are making their first attempts at developing a chatbot, and first tries are sometimes only marginally successful. But marginal success only indicates there is more to do. When a chatbot is promising but not delivering on its initial goals it may be time to rethink and rebuild. Instead of just trying to fix a million little issues, sometimes it makes more sense to restart the process and reimagine the bot you need to build based on the learnings from the first attempt.

One of the keys to creating a successful chatbot is understanding what you don’t know, and making the decision to take a step back and rebuild if necessary. It isn’t always possible to add features to an existing application, and starting fresh can mean incorporating lessons learned from partially successful trials.

Chatbots are growing in popularity but they are still far from being commodity technology. Successful chatbots will be created by learning from early trials and taking advantage of already mature technologies like voice processing systems that can be plugged into development efforts through APIs.

Read our “How to build a chatbot” post to get started with your own bot in minutes.

Learn how you too can build a chatbot in just minutes.

 

Add Comment
No Comments

Leave a Reply

Your email address will not be published.Required fields are marked *

More Cognitive Enterprise Stories
May 8, 2017

How to integrate custom-built annotators into your data pipeline

Watson Discovery Service and Watson Knowledge Studio help businesses extract information and insights from complex, domain-specific content using a simple enrichment structure. Apply document enhancing capabilities to extract information from industry domains. With business logic built into the ground truth, annotations carry meaning pertaining to a business scenario.

Continue reading

May 9, 2016

Watson learns to understand Korean life and language

Today IBM and SK Holdings & CC, a leading Korean IT services company, announced plans to introduce Watson cognitive services to Korea. The significance of this milestone agreement is two-fold. For SK Holdings & CC and Korean organizations, access to market-ready artificial intelligence services comes at a critical time of change and an increasing appetite […]

Continue reading

June 2, 2017

IBM engineer Rama Akkiraju featured on list of top 20+ leading women in AI research

Very few earn the title of “Distinguished Engineer and Master Inventor” at IBM, but Rama Akkiraju’s contributions warrant the distinction. Akkiraju's teams are responsible for many of Watson cognitive services and leverage AI, psychology, sociology, decision theory and consumer behavior.

Continue reading