Woman bending down to receive a package from an futuristic robotic delivery service.
PHOTO: M.Dörr & M.Frommherz

For a medium that is as unsparing and spartan in its interactional capabilities as the voicebot, the difference between delight — yes, voicebots can delight you, even those of the telephony IVR stripe — and dismay almost always hinges on personalization. Did the voicebot take into consideration who the human being it is interacting with is, or did it treat the human being as a stereotype — a cardboard, a two-dimensional (at best) cartoon character?

Surprisingly, it doesn't take much to delight voicebot users. Equally surprising is that very seldom do voicebot designers put in the effort to take advantage of even the lowest hanging fruit at their disposal to deliver user delight.

We share here four such low hanging fruit:

1. First Time vs. Repeat Users

For voicebots that are used repeatedly — for instance, managing your 401k contribution once a month, listening to the latest transactions on your bank account, checking in when arriving at a location — a very important piece of information that can be used by the voicebot is whether the human is a first-time user, a recent user or a long-time frequent user.

Knowing this enables the voicebot to determine how wordy and instructive it needs to be. With the first-time user, the voicebot should be expansive in their language and should hand-hold the user, one step at a time, lest it loses the user. With frequent users, the voicebot should be succinct and hands off, lest they annoy the user.

Related Article: Why Voicebots Continue to Disappoint Us

2. The Situational Context

In cases where the reason why the user is engaging the voicebot can be easily inferred from readily available information, the voicebot should leverage that information to streamline their engagement with the human. Making choices is taxing on humans, especially when they don’t see them but have to remember them.

Instead, humans prefer direct "Yes/No" questions. For example, if I have opened an issue ticket recently about a problem I encountered using some software, and I engaged the voicebot. The voicebot should leverage the fact that I had opened the issue ticket and ask me at the very outset of the engagement if I am engaging about that issue. Chances are very high that the answer to that question is a "Yes."

So, instead of:

  • Human: Hey Google, launch Hoot-Spot Support.
  • Google: Hoot-Spot. Welcome. Here’s how I can help you. You can say…
  • Human: Information on an existing ticket.

This interaction is less taxing:

  • Human: Hey Google, launch Hoot-Spot Support.
  • Google: Hoot-Spot. Welcome. The ticket you opened this morning is currently being worked on. Do you want more details?
  • Human: No. Thanks! Bye!

Related Article: How Should a Voicebot React to Verbal Abuse from a Customer?

3. Patterns of Usage for User

Other information that is readily available to the voicebot that could help the voicebot infer the main reason why the human is engaging with it are the patterns of usage by the human. Modern voicebots sit on top of powerful Artificial Intelligence platforms that can, given a history of actions, detect patterns. It may be that the user engages the voicebot every morning to check on their stocks, or every Saturday afternoon to listen to three movie recommendations, or every Sunday morning to get an idea for a brunch recipe.

Take this interaction:

  • Human: Hey Google, ask Stock Picker for my top three performing stocks.
  • Google: Here you go...

The user here has to focus and formulate their question. This formulation takes a lot more effort than this:

  • Human: Hey Google, Launch Stock Picker
  • Google: Hello. Do you want your performing stocks?
  • Human: Yes.
  • Google: Here you go...

4. Patterns of Usage for User Base

The fourth low hanging fruit has to do with patterns of usage not of a particular user, but of the whole user base, or of a cohort of the user base.

The AI platform that modern voicebots live on can also examine the data collected from interactions with sets of users over spans of time and detect patterns that apply across the whole user base or a subset of the user base.

For instance, it may be that every first workday of the month, the voicebot sees a spike in usage, with the spike caused by users checking in to fulfill a compliance requirement. Inferring this, the voicebot can offer users, right out of the bat, the option of engaging about compliance rather than forcing them to listen to a menu of options and only then selecting the compliance option.

To be sure, a percentage of the people will not be engaging about the compliance task, but if 80% of the users are engaging about compliance, it means that 80% of the users will be delighted, while the other 20% will be only minimally inconvenienced, since, if the voicebot is well designed, they would only have to answer "No" to a "Yes/No" question. And if the voicebot wishes to go for all out delight for the whole 100%, they can remember who had said "No" so that next time, the question is not asked of them.

So, instead of this:

  • Human: Hey Google, Launch MacroStrategy.
  • Google: MacroStrategy. Here's how I can help: You can say...
  • User: Compliance check in.

The exchange can be this:

  • Human: Hey Google, Launch MacroStrategy.
  • Google: MacroStrategy. Do you want to start your compliance check in?
  • User: Yes.

Conclusion: Less Talking for Customers

Voice is a time-linear medium that requires the user to listen carefully, speak clearly and in a timely fashion. The less the user has to speak and listen, the more effective the interface is.

Leveraging context and personalizing the behavior of the voicebot to take that context into consideration is an effective way to deliver delight through a medium that is known to frustrate rather than serve. If you have a voicebot that your customers are interacting with, you can easily differentiate yourself with excellence at those sensitive moments when someone wants to buy from you or when a customer or a partner or an employee is reaching out for help.