Automated self-service is Silicon Valley’s trick for automating the service economy. By tapping end users as a new workforce, it also fuels machine learning and the knowledge economy.
Automating the Service Economy
Automation has dramatically reshaped the contours of our agricultural and manufacturing sectors. Now, with a boost from machine learning, it is remaking the service economy. Its footprint, recognizable in all the service markets it touches, shows up as a disruptive new ability for end users to serve themselves. Intelligent, automated self-service is what automation looks like when it hits the service economy. It will revolutionize the way we understand work and is paving the way for the knowledge economy.
Just like agriculture and manufacturing companies, service firms automate by replacing employees with machines. The difference is that most of the work of service firm employees ultimately entails directly serving—and interacting with—end users. When service companies automate work, that contact between end users and employees is reduced or eliminated, so end users interact with a company through its various technological interfaces. Examples include bank ATMs, automated gas pumps, self-checkout counters at supermarkets, e-commerce websites, and even publishing tools like the one I am using to share these words with you.
When we learn to see it, most of what we think of as the “high-tech sector” is actually just the service economy becoming automated. Firms like Google, Amazon, and Uber use self-service technologies to disrupt one market after another. The underlying premise behind most Silicon Valley startups, too, usually boils down to “Use technology X to disrupt service market Y before anyone else.”
The New Work of Serving Ourselves
Many of us have mixed feelings about self-service. As Craig Lambert points out in his book Shadow Work (Counterpoint, 2015), companies frequently use self-service technologies to reduce labor costs by dumping tasks such as grocery checkout onto end users. With that said, nearly two-thirds of consumers respond that, when it comes to interacting with companies, they prefer to answer a question or solve a problem on their own. We like the flexibility and speed of performing our own searches rather than working through a human service provider, such as a librarian or travel agent. We usually find it faster, more convenient, and less expensive.
As machines enable us to do more things for ourselves, they eliminate service jobs without necessarily eliminating service work. Automated self-service technologies enable us to work directly with a company without interacting with its employees. The travel agent no longer books my flight; I do. That job disappears, but not necessarily the work. These interactions may not feel like work, since we engage in it voluntarily and aren’t compensated for doing it. Shopping on Amazon or booking flights on Expedia feels more like consumption than actual work, but that is only because our notion of work has yet to catch up with an automated service economy.
All service businesses—whether automated or not—generate value through a co-creative partnership with their end users. Whether it’s an airline, a streaming music service, or a restaurant, the value of a service emerges as we engage with it. Self-service technologies just replace employees as the interface through which end users and the company co-create that economic value. Self-service technologies act as a medium or membrane through which companies automate engagement with end users. They are the new engines of value creation, employing the workers of the new economy: end users.
Intelligent Automated Self-Service
Tools such as speech recognition, speech synthesis, machine vision, and emotion recognition lessen the burden these new “workers” face in doing self-service work. In the process, machine learning makes it possible to automate more and more services that once relied on the subtlety of the human touch. Machine learning thus acts as a catalyst for an automated service economy.
This relationship between service automation and machine learning is reciprocal, however, because self-service technologies also play a critical role in harvesting the end user data that fuels much of today’s machine learning. One of the fundamental design principles of such technologies is that our interactions with them automatically generate a stream of behavior data, and so, while high-tech service disruptors empower billions of people around the globe to serve themselves, they simultaneously generate vast quantities of data that allow the likes of Google, Facebook, and Amazon to dominate the field of machine learning.
As automated self-service technology gets smarter through machine learning, it becomes intelligent automated self-service. The automation drives down costs while the intelligence improves the usefulness and attractiveness of the service. Better service generates more usage, more data, and ongoing rounds of increasing intelligence, all of which feeds back on itself. This virtuous cycle is the engine that high-tech disruptors use to devour one service market after another. The result is an upward spiral of domain-specific, artificially intelligent, automated self-service.
Growing the Knowledge Economy
From the perspective of economic productivity, these increases in automation and intelligence aim to increase the efficiency of end users, not of employees. The question is, to what end?
Google does not pursue its mission—“To organize the world’s information and make it universally accessible and useful”—out of the goodness of its corporate heart. Instead, Google (and Facebook) have convincingly demonstrated the huge profit potential of extracting knowledge from end users. Amazon and Uber operate with business models that differ from either of these two advertising giants, but what these companies—and in fact, all companies now leading the automation of the service economy—share in common is the ability to transform the information gleaned through end user engagement into useful, and extremely valuable, knowledge. Indeed, their profits ultimately hinge on this ability.
Machine learning transforms raw data into information that, when combined with automation, is in turn transformed into applied knowledge. The automation of the service economy is thus the precursor of the rise of a new, knowledge-based economy, and automated self-service technologies are becoming the primary interface through which we embed human knowledge into machine intelligence. In this new knowledge economy, our myriad contributions of data are transformed into a kind of coding—a synthetic intelligence, part human and part machine—that directs the automated operations of machines. Our interaction with automated self-service technologies is the new work of humans, and through these contributions we embed human volition and meaning-making into a new, synthetic intelligence—and into the automated knowledge economy in which it thrives.
Automated Self-Service is Taking Over. Here’s How it Happened.
The seeds of our primary interface with machine intelligence rest in the humble origins of the vending machine. Here is the brief story of this evolution of self-service.
My recent experience with this very thing: I booked an Avis rent car online. As my plane landed and was taxiing, I received a text, telling me what slot my car was in and what type of car. We found our way through the rent car garage and located our car, keys inside. Not a soul was anywhere in sight. As we drove out of the exit, a human being asked to see my drivers license. A few days later, as I entered the airport grounds, I received another text, telling me where to drop the car, and checked me out, emailing a receipt. We parked in the return lane, again no person in sight, and off we went.
The amusing thing is that soon thereafter, I received the typical survey of my experience, and the survey asked about my dealings with various people, like during the rental and return. There was no “N/A” option on that question. I think they need to update their survey!
Thanks for sharing that, Bill. Yep, that is exactly what I’m talking about. The effect can be downright eerie at times — like we’re the last person on Earth.
The other night, we had dinner with my mother-in-law at a Chinese restaurant here in Seattle. There were no regular menus; just an iPad through which we were to enter our choices. My son quickly figured out that each of us could access our own view of the menu via our phones by scanning a barcode that was plastered onto the table top. The concept was cool, but we ended up spending ten minutes or so, just figuring out how to use the system and place our order. Less work for the restaurant. More for us. They still had plenty of wait staff around but one could imagine that changing once this tech gets ironed out. I think that’s going to really change what it feels like to go out to a restaurant. High-end places will still have staff, of course. But for mid and lower-tier places, this is probably where we are headed.
Sigh.
Self-As-A-Service: the final frontier.
You first. 😉
I want to share three aspects.
1 Automation is not just making us serve ourselves. It is also unlocking value from a level which was not accessible to us through intermediated service. If I can order food from any restaurant even if far, and pick and choose from ites from different restaurants, it is value not available earlier.
2 It is supposed to enable overcoming conceptual complexity and also detail complexity of interaction. In some situations we have better ability to interact on conceptual complexity. In a travel arrangement, I may have many different points of views in selecting the alternatives. I can do them better myself than explain all of them to a intermediary. In some other cases detail complexity such as filling some data again and again can be a hassle.
3 With the digital service system, intermediated service has not totally vanished. In India, I see many ‘service providers’ helping customers to reap the benefits of this digital world by performing the ‘self service’ on behalf of their customers!
Thanks for your comment, Sowmyan. I always appreciate your weighing in with your thoughts.
On your first two points, I think that these automated self-service technologies open up all kinds of economic value that were simply not feasible before. There is great power in enabling end users to directly manipulate the ‘dials’ that shape the flow of services. We can take as much or as little time as we want in getting things just the way we want. To be sure, there are situations where these systems are implemented in ways that are clunky and simply transfer work to end users. This is the case with many of the older-generation checkout systems in grocery stores. But Amazon’s Amazon Go stores have a kind of self-checkout that is much improved and I think will be very popular for certain types of retail environments because it has focused on not dumping more work onto the end user.
On the third point, I agree. There will be plenty of services that will blend automation with employee support. Ideally, these systems will free up people so that they no longer have to act like robots.
Gideon, Really interesting article.
An implication you have not touched on is human isolation. We are entering a world in which many of our needs are met without human interaction. Yet it is human interaction that is crucial to our mental health.
There was once a time I booked my travel through a travel agent. I am glad to be able to book my own travel — much more efficient for me and I can better tailor my travel to my preferences. But I have lost a human contact — a live person to confer with. I don’t miss that loss — or do I?
If replicated in many, many contexts, I can live my life with fewer and fewer human contacts. Some observers have suggested that our societal problems with drug addition, suicide, and depression are related to isolation. This service automation seems to be driving us toward a world that diminishes our human contacts — and may drive us also to mental despair.
An excellent point, Fran.
We are social animals. When we interact with fully automated systems, it can feel isolating and even alienating (see Bill’s comment above).
What I’m not sure about is where this all heads. There is one version where find ourselves ushered through more and more automated commercial systems that eventually crowd out what it means to feel fully human.
I am more optimistic, however, and doubt that people will put up with this direction for long. The answer that leads us to another path lies in fixing our underlying economics. The reason I say this centers on how we handle the freed up time from automation. Will people have the resources to enjoy the resulting freedom? Perhaps we will find that these systems will most focus on those routine interactions with others where most of what we encounter is scripted, corporate behavior. Perhaps it will free up space for humans in the loop to concentrate on building real connection with us. Or perhaps we become free to spend more time in experiences that are non-commercial in nature.
There are potentially positive paths here, but only if we design for these outcomes as opposed to turning ourselves into hamsters on automated hamster wheels.
Thanks for weighing in.
Gideon,
I agree there is a potentially positive path. It’s like when so much food is fast food from franchises, people start seeking out slow food – – unique food, cozy environments. But often acting on that countervailing tendency is dependent on having enough money or time or both to pursue those healthy impulses. If people are sufficiently desperate to simply survive, it’s very hard to do that.
Agreed, Fran. Questions about these technological systems are so very tied into the underlying economic systems in which they are embedded. It’s impossible to separate them.
Pingback: I heard the news today …
Pingback: meshedsociety weekly #223 | meshedsociety.com