This content is part of the Essential Guide: Guide to AI in customer service using chatbots and NLP

Customer support automation for B2B requires human touch

Robert Johnson of TeamSupport and Sandy Crowley of ICANotes offer tips on how B2B companies can use automation to supplement human customer support services.

While AI and chatbots are the shiny new toys customer service departments can roll out to augment human agents, some experts say B2B still isn't quite ready for full-on customer support automation.

Developing technologies, such as AI and machine learning, leave significant room for error, but the solution may lie in streamlining the customer support process and picking the right time for automated customer response tools to hand off to human teams to take advantage of automation in its current state.

There are a few tricks of the trade that B2B companies can keep in mind when using humans to facilitate the customer support automation transformation, said Robert Johnson, CEO of customer support software TeamSupport, a B2B customer support software company.

"Too many employers spend big on automation and bots then try to justify the trend with expensive and unnecessary training for the entire company," Johnson said.

Bots, he said, should never replace humans.

Chatbots not ideal for B2B marketplace

Chatbots aren't ideal technology for B2B, because the business world still values interpersonal relationships, which can be too delicate to trust to bots. A better use of automation in B2B support can be found on the back end to flag service-level agreement issues or violations or to streamline processes, including ticket assignment and updates. These uses improve the customer experience and reduce resolution times.

Just ask Sandy Crowley, CEO of ICANotes, a TeamSupport customer that sells a cloud-based medical records application to behavioral health professionals. The company, bound to Health Insurance Portability and Accountability Act health data privacy rules, uses TeamSupport as a ticketing system and Knowledge Base for customer support, as well as for tracking quality assurance testing and process documentation for support agents.

Customer-facing automation is highly dependent upon the quality of the search term, and users can quickly become frustrated if they are unable to find or interpret the information they are seeking.
Sandy CrowleyCEO, ICANotes

"With the search capability available across all areas of TeamSupport, an agent can quickly search for and retrieve the information needed to address the customer's issue," Crowley said. "Should all of these resources fail to give the agent the answer to their question, they can access the rest of our team via Slack chat channels."

The company, where employees work virtually, hasn't yet implemented chatbots or AI as part of its customer-facing support options. "Customer-facing automation is highly dependent upon the quality of the search term, and users can quickly become frustrated if they are unable to find or interpret the information they are seeking," Crowley said.

That's because the majority of ICANotes users are not especially computer-savvy and prefer having access to and interaction with a trained agent who will walk them through the answer to their question, step by step. ICANotes' highest level of customer satisfaction resulted from providing agents with the resources they need for a speedy and helpful response to customers.

Beware AI fails

There are plenty of examples of bad things happening when a company relies too heavily on automation. For example, using customer support automation to identify the gender of a customer based on first name and automatically assigning a salutation of Mr. or Mrs. could put a company at risk of offending customers. But blunders like these can be easily overridden -- or, better yet, prevented -- by supplementing AI-powered customer service with customer support teams.

Depending on the provider, technology doesn't always leave a lot of room for error, and the acceptance of failure is significantly lower in the B2B world than B2C. Facebook's chatbot, for example, has only a 30% success rate, according to Johnson, and some B2B solutions aren't much higher.

"A service like Facebook is free to consumers, and expectations are low," Johnson said.

B2B companies, on the other hand, might pay thousands or even millions to a service provider, and that payment elevates expectations and makes successful communication experiences even more important.

Making chatbots work

Those who do decide to use bots in B2B need staff that fully understand them and can make up for their shortcomings during customer calls. As with any omnichannel tool, customers need to see a clear and direct path to move past the automated bot and speak to a real person immediately, without starting over.

That doesn't mean customer service personnel should be trained on every trivial aspect of customer support automation. That may overwhelm them. They only need to understand the basic technology, how it's being used by customers and what their role is with the technology.

See a demonstration between a human and chatbot.

"Don't go into the intricate automation nuances of this technology with your entire team, at least not yet, and especially not with front-line workers," Johnson said.

Technology is a tool in their arsenal, and the most effective way to use it is together with other tools, not as a stand-alone solution.

Looking forward

Of course, the relationship between humans and automation is likely to evolve. Customer support automation still has a long way to go, and as for TeamSupport, Johnson has his eye on one budding technology: accurate customer sentiment analysis.

"We're really excited about this technology automatically determining the 'mood' of a written customer issue, based on the tone of the writing from the customer," Johnson said.

But even decades from now, when automated customer support essentially replaces a company's static knowledge base, customers will still need humans to help solve the thorniest problems that bots can't puzzle out.

"It's important as this technology improves that businesses understand its role and do not attempt to 'fake out' a customer by pretending a chatbot is a person," Johnson said.

Dig Deeper on Customer service and contact center

Content Management
Unified Communications
Data Management
Enterprise AI