Adventures in Artificial Intelligence Part 4: Human Service Tasks AI Will Replace
With companies like Forrester predicting artificial intelligence (AI) investments will jump 300% in a single year, it’s clear that AI is not just a passing fad. Adventures in Artificial Intelligence is a series for savvy field service professionals looking to stay on top of AI trends, and their impact on service.
In our previous post, we explored machine learning and its potential applications for field service organizations.
This week, we’ll be getting more specific about which human tasks artificial intelligence will replace in both the short, and long-term. But first, some news. The AI landscape is shifting rapidly, and key players continue to shake things up on a daily basis.
Our first story features Google, who revealed at their recent annual developer I/O conference that their neural network and artificial intelligence software is designing new AI systems more effectively than Google engineers. That’s right, the AI they built is now autonomously building better AI.
This project, dubbed AutoML, involves deep learning techniques and neural networks that effectively mimic processes found in the human brain. Google is using this software to, “design networks for image and speech recognition tasks. In the former, the system matched Google’s experts. In the latter, it exceeded them, designing better architectures than the humans were able to create.”
How will this impact service? As AI gets better at recognizing images, speech patterns, and facial expressions, it will likewise be closer to evaluating whether equipment requires service based on simple images.
Here’s the full scoop on the future of AI at Google, if you prefer to watch:
Halfway across the world, a Swedish startup named Gavagai AB announced this month they’ll be using artificial intelligence language analysis software to monitor and decipher the language of bottlenose dolphins.
Does that mean future middle-school children will choose between French, German, Spanish, or Dolphin? Unlikely.
While a dolphin language analysis appears rogue at first blush, Gavagai AB has assured many that this challenge will directly improve their software’s ability to interpret words, and noises alike. They report this investigation will improve their analysis of human language and emotional triggers, as well as help them interpret sea creatures’ behavior, or even someday in a galaxy far far away – interpret space alien languages.
In a solar system a bit closer to field service, OpenAI recently announced they have created an artificially intelligent robot capable of “one-shot imitation learning.” The robot effectively watches a single simulation of a task, learns it, and performs it without flaw.
In the screenshot above, a human simulates the stacking of blocks using virtual reality goggles (right) and a real-world robot precisely mimics it in one-shot, without flaw (left). We can not overstate what a huge leap forward in robotic ability this represents.
This type of technology will be a real game-changer for service in the coming years. Just imagine a few of the potential scenarios in which we could use simulations to teach robots to:
- Perform service tasks in dangerous scenarios
- Troubleshoot equipment in inclement weather
- Fix space equipment while in-flight
- Fix pipes, networks, and dams submerged underwater
- Manually troubleshoot equipment that previously required having humans present
Without further ado, here are some human tasks in field service that AI (and robots) will replace in the near, and long-term.
Human Tasks AI will Replace
While autonomous vehicles, or even field service drone taxis, are not quite ready for market adoption in most countries, there’s no denying that at some point in the future our cars will drive themselves. Artificial intelligence software that enables autonomous driving is rapidly reaching its market potential. In fact, Nvidia and Audi recently reported they’ll have a driverless car ready for market by 2020.
Many service engineers spend half their time behind the wheel each week. If freed up from driving, they would have more time to perform strategic tasks while in transit, including:
- Getting familiar with their next customer challenge
- Communicating with previous, or future customers via text, or voice
- Fielding photos or videos of broken equipment from their next service site
- Connecting with dispatch to discuss scheduling, conflicts or customer queries
These are obviously just a few of the potential values of freeing technicians up from the wheel.
2. Dispatch & Scheduling
While many scheduling softwares already leverage some automated features, the majority of future dispatch and scheduling tasks will be completely supplanted by software.
These AI dispatchers will be capable of interpreting, and mimicking human scheduling behavior, and will be far superior to humans at dealing with crisis scenarios like storms, major power outages, or even earthquakes.
Naturally, there will still be a need for humans to manage the dispatch software itself. Likewise, service engineers will still need to interface with their human counterparts in troubleshooting schedules, optimizing routes, and more. This isn’t doom and gloom. We expect a hybrid human-software approach will be necessary.
In all, we expect this process to happen slowly, but you should start seeing AI dispatch and scheduling automation broadly adopted within the next three to five years.
Here are a few changes you can expect:
- Voice-activated technology will be introduced that allows dispatch managers to distribute tasks without any typing. Likewise, field engineers will be able to accept or decline schedule changes using this same voice-activated technology.
- Machine learning algorithms will be built based on the most efficient dispatch professionals. Dispatch software will continue to be optimized based on this information.
- Comprehensive programs which can autonomously manage dispatch and scheduling, and evolve while doing so, will enter the market. This may take years, but will completely change dispatch management as we know it.
3. Self-service & Customer Communication
The first question any IT service professional will likely ask when you call is, “Have you tried restarting it?”
Likewise in field service, millions of customer calls result in millions of field engineers getting in millions of vehicles driving to millions of service locations only to find out the path to resolution is so dirt simple, the customer could have gained resolution all on their own.
THAT is a big fat waste of time, talent, and money.
In the coming years, artificial intelligence software will alleviate the need for needless service visits by interfacing with customers to resolve many of their requests through self-service, on-demand equipment assessment, and more. PwC recently reported that this type of AI could eliminate as many as 80% of current service requests. Joe Atkinson of PwC remarks,
“Instead of sending every service ticket to a dispatcher, we can route it first through an algorithm and determine the best avenue to solve the problem. Is there an online FAQ that provides the customer a quick path to resolution? Would a simple customer-performed troubleshooting step, like restarting equipment or replacing a simple component, solve the issue?”
As the field service industry faces a rapidly aging workforce, who struggle to keep up with customer requests, AI will be a key component in bridging the talent gap.
Here are a few ways AI will help customers resolve simple service challenges:
- Chatbot technology will be available to customers who can ask specific questions about equipment, parts, common service scenarios and more. These chatbots will instruct customers on the simple steps they can take without service, and when a scenario warrants involving their service provider, or technician.
- By embedding artificial intelligence software and sensors in complex field equipment, devices that breakdown will be designed to automatically send information to dispatch, or the service team about exactly what isn’t working. Service would effectively find out something is broken before the customer does.
- Years in the future, many customers will have voice-activated devices connected to IoT sensors embedded on all the mechanical equipment they own (or in enterprise settings, the office manager deals with). Customers will simply ask their voice-activated “hub” questions about malfunctions, mechanical device status, and more.
While AI will certainly replace many human tasks, we feel the most strategic roles in service can never be supplanted by machines. Whether robots can learn service tasks, software that will re-organize a dispatch schedule in seconds, or drones that can survey hard-to-reach equipment, we will need even smarter service technicians and managers who can humanize the technology for customers.
In the end, good service is about good relationships with customers. We believe artificial intelligence can help us all achieve that end.
For more on artificial intelligence and technology trends, subscribe to the Field Service Matters blog.