<img height="1" width="1" alt="" style="display:none;" src="//www.bizographics.com/collect/?pid=5286&amp;fmt=gif">

Digital Engagement Blog

3 Things the Self-Driving Car Teaches Us About Intelligent Assistance

Picture this.

You want to go somewhere.

  • Your office.
  • A restaurant to meet some friends.
  • Aunt Alice’s house for pick any holiday.

So, what do you do?

  • You jump in your car.
  • You say where you want to go.
  • You arrive safely, on-time.

This is the promise of self-driving cars.


Lessons from Autonomus Cars

And, it’s about to happen...at least that’s what we’ve been led to believe.

But is it?

It may seem obvious once you truly think about it, but that future of self-driving cars is likely much further away than we’ve been led to believe.


Mostly because it’s a cool story. The idea of jumping into your car, telling it where you want to go, and then just arriving safely is the thing of science fiction. It’s exciting. And so, we want it to happen.

Sexy Terms

According to Steven Shladaver, one of the pioneers of the Intelligent Transportation Systems program in the United States, this is precisely why the popular media and automotive industry have confused the issue. Self-driving...autonomous...driverless are sexy terms.

(Sounds like customer service in 2016, right?  Chatbots...Messaging...Artificial Intelligence… no more agents!…anyone confused?).

At the end of the day, it’s all about automation and intelligent assistance.

And as it turns out, automation of discreetly human actions isn’t so easy.

In other words, what humans are able to quickly learn and perfect—for example, treacherous weather and the changing road surfaces it causes, or avoiding that crossing guard who happens to be standing in the middle of the road—is not so easy for technology to overcome.

All of this is indeed interesting...and incredibly cool...but you may be thinking, what does any of this have to do with customer engagement? What could driverless cars really tell me about the customer experience and customer service?

It’s all about automation.

3 Lessons

Here are three lessons to take from the journey of driverless cars, from an idea born in science fiction all the way to that neo-nostalgic drive you’ll take someday without a second thought...you know, over the river and through the woods to Grandma's house in your horseless and now driverless carriage.Self-Driving Cars and Customer Service

Lesson 1: There are Different Levels of Automation

All automation is not created equal.  

To get the automotive world, popular media, and consumers all on the same page in regards to self-driving cars, SAE International—the largest automotive and aerospace standards-setting body in the world—created standard J3016: Taxonomy and Definitions for Terms Related to On-Road Motor Vehicle Automated Driving Systems.

The standard breaks the automation used for autonomous vehicles into six levels, 0 (representing no automation) through 5 (full automation):

SAE table.jpgSource:  SAE Automated Driving Levels.  Download a full-size copy here.

For the Customer Experience world, and especially in Digital, this has significance. There are multiple levels of automation, and all of them have a place in a comprehensive digital engagement strategy.

 Cars and Customer Service:  More Overlap Than You’d Think

So let’s take a look at the automation levels of self-driving cars and translate them for customer service. The automotive industry has deeply considered automation modes and failure modes.  In customer service, we can put our SAE hat on and think of actors, automation types and failure modes.

  Cars Customer Service
Actors Driver, Automation, Car Customer, Automation, Agent
Automation Types 6 levels  ~ 6 levels
Failure Mode Revert to driver Revert to agent

Let's look at what six levels of automation in customer service would look like:

Level 0: No automation

Customer can talk to an agent at anytime. Some organizations use this level as a specific customer engagement strategy.

Level 1: Customer Assistance or Basic Automation

Level 1 sounds like this: “We use an IVR to automate basic tasks like authentication and account balance.” Easy and limited risk. The fallback is allowing an opt out from the IVR to speak with a live agent.

Level 2: Partial Automation

Increased automation allowing for more transactions to be completed than the basic tasks of Level 1. This could happen in an IVR or on a website. Most businesses are at level 2 or 3.  Fallback is to a live agent when needed.

Level 3: Conditional Automation

Increase automation levels using chatbots or other intelligent assistance. Here, machine learning helps understand customer intent and guide them accordingly. Fallback is to a live agent if needed.

When you think about chatbots consider they can be used in multiple ways:

  • Customer-chatbot (chatbot services the customer directly)
  • Agent-chatbot (chatbot helps the agent instead of the customer)
  • Customer-chatbot-agent (3-way interaction with the chatbot helping progress the task)

Level 4: High Automation

Intelligent systems (Bots, VA, AI, etc) learns over time and can handle more tasks without agent intervention. (This is what’s happening with cars right now.) Human agents are in the loop in the case a fallback is needed.

Also, consider other ways to use a bot…

I have my personal bot go do something for me. For example, I command my bot to do a task where it interfaces with the automation system in place in an organization's customer service department (perhaps to dispute my parking ticket).  This happens bot-to-bot.  But I’m always there to clean up potential errors.

Or, you might consider Amazon Dash buttons to be at level 4 (even without smart ML algorithms involved). What happens when I hit the dash button? I get more toilet paper. Bot-to-bot-to-toilet paper.

Level 5: Full Automation

A completely automated experience. Since humans are not involved, automated systems provide fallback. One scenario that comes to mind for this level? IoT devices are requesting service.

Autonomous Cars and Customer Service

Lesson 2: Lower Levels of Automation are Still Useful

In the evolving world of Digital, don’t dismiss basic automation. Instead, make it work for today’s customer.

Full automation can be a great thing, but it must come at the preference of the individual. This means all levels of automation—including no automation at all—have relevance.

Customer Service Automation Levels
Level Automation Agent
0 None Always
1 Basic Opt out
2 Partial When needed
3 Conditional If needed
4 High Just in case
5 Full Never


For example, those sensors that help keep a car in lane are considered low-level automation according to SAE. But, are they helpful? You bet. If you’ve ever driven late at night, exhausted, you know.

Likewise, there is a place for even the most basic automation in customer engagement.

IVRs get a bad rap not because of the technology itself, but because of how they’re deployed. Done right, the IVR is proven to lower customer service costs and actually can improve CX scores. And, it’s just about the lowest level of automation available to customer engagement pros.

Next up are chatbots. We saw in 2016 that chatbots created lots of buzz but failed to deliver much of the promise. We recommend using chatbots first by providing level 1-2 automation, and only then moving on to more complex tasks.

Lessons of Self-Driving Cars

Lesson 3: Human Intervention Should Always be an Option at a Moment’s Notice

Automation can be wonderful for all if human intervention is a natural part of it.

The biggest barrier for fully autonomous vehicles? Those discreetly human actions. What if the technology cannot overcome a sudden change in weather? What if there is a software malfunction? What happens to the passengers when a new variable...a new driving situation occurs?

For self-driving cars, this is why human intervention must always be available. Automation, of course, but automation managed by a human.

And, it’s the same for brands. The ability for human intervention and interaction is a must to ensure the best possible outcomes.

“Machines can be programmed to do the next thing right. But only humans can do the next right thing.”  Dov Seidman, via Thomas L. Friedman JAN. 4, 2017

While the idea of a fully automated experience across customer goals and channels and even brands is sexy, for it to reach its potential, there must be the ability for human intervention at the moment the consumer wants it…if not a nano-second before.

That means being able to interact human-to-human when it’s needed.

It means hard rules can be bent to suit the needs of the customer when appropriate.

It means the best possible experience may be delivered—consistently—despite variables that would not compute with automation alone.

What Does This All Mean for 2017?

So will we see mainstream, fully autonomous cars in 2017? (Our friends at NVIDA say 2020)

Perhaps, though probably not.

But, we will continue to see the evolution of digital engagement with automation that grows more sophisticated with each interaction.

And, those brands that embrace all levels of automation without forgetting that human touch? They will be the digital engagement leaders we’re talking about. They will be the brands that thrive.

For consumers, that sounds like a great ride, indeed.

Forrester Chatbots 101.jpg

Chatbots 101

Ready to get moving on automation but need more insight?  This Forrester report lays out Chatbot basics.

Get Your Copy

Topics: Customer Care Performance Optimization User Experience Enterprise Chatbots