One trend I’m optimistic and pessimistic about, at the same time, is automation. This entails AI (artificial intelligence), machine learning (which has significantly improved artificial intelligence), and the improvements of optical and speech recognition software (also through machine learning). Advances over the last few years have begun to find practical applications and are slowly being introduced in several sectors.
These advances are going to significantly improve our standard of living in many ways. The first, a focus for many businesses, is customer service. Through natural language processing (NLP), chatbots, and AI the channels we communicate with businesses, response times, accuracy and quality of information, and total interaction time are going to decrease and consumer satisfaction will increase. Here is a use case to compare and contrast current versus a potential future scenario.
Today when you call your bank, insurance provider, or phone company you walk through their automated phone menu, typing or speaking your unique identifiers, routing your call to the right department, and finally you get to ask your (usually) simple questions. If your question isn’t simple, you may be sent to a different department or escalated to a manager.
In the future, you can have a fluid conversation and while you’re asking your questions, the system validates your identity using voice recognition software and can answer any question, from any department. You’ll never wait on hold, you’ll always speak to the right person, and always get the right answer the first time. If you require escalation, the “let me speak to your manager”, you’ll be routed to a human manager in the appropriate department who has your account validated and pulled up, and can view a transcript of what has already been said on the call. If the conversation was heated or you previously mentioned keywords indicating escalation would be required, maybe “refund” or “lawsuit”, the AI might’ve already alerted the manager to follow the conversation closely. All of these are going to make your experience, as a consumer, infinitely better.
This scenario is optimistic, with several different technologies involved. Most of these technologies already exist but haven’t been combined together. Citi Europe is using voice recognition to verify caller’s identities. Watson, IBM’s Jeopardy-winning AI supercomputer, already distinguishes nuances in the English language and can return relevant results and answers to difficult questions. Chatbots, the text versions of customer service are being introduced through several channels consumers already use like Facebook Messenger, WhatsApp, and WeChat. So this scenario isn’t far in the future.
Automation will make our jobs easier and more rewarding, with impacts for consumers as well. Gone are the days of manually entering data, mindless repetitive tasks, and double checking information sources (reconciliations). Machines can do all of these, faster and more accurately. A few examples are automatically processing simple insurance claims, scanning paper copies and automatically extracting relevant information, and providing quotation information when receiving email or fax requests (RFQs). These are happening now, using software from companies like BluePrism. There hasn’t been widespread adoption yet, but it’s coming with a Japanese insurance company replacing 34 people with computers and ING, a Dutch bank, planning to replace 5,800 workers over the next five years. As technology improves to perform these simple tasks, workers will transition to perform value-added tasks requiring creative thinking. Consumer’s benefit as interactions become faster and less expensive. For instance Lemonade, an insurance startup in the US advertises it can pay a claim in less than three seconds, an incredible improvement against insurance giants that still require faxes, mailed papers, and maybe calls/follow-ups. This will carry over into filing taxes, applying for mortgages or other financial products, and almost any situation you complete forms or paperwork. It’ll represent a significant improvement for everyone.
As we identify applications for robotics and automation a new workforce will be recruited to high-paying jobs that train, improve, and design better robots. These jobs will continuously improve speed and accuracy while reducing cost. As the monotony of a job is automated, these workers will find a more stimulating environment. Expect these roles to go first to engineers, mechanical and software, so the jobs will pay above average to significantly above average salaries.
As automation replaces monotonous tasks, improves quality, and decreases cycle/production times businesses will see costs decrease in labor, materials, and other unexpected areas. Over the last two decades, millions of US jobs have been offshored to lower cost labor markets like India and the Philippines. Again these costs will be reduced as these roles are easily automatable and the robotics license costs 70% less than an offshore worker. After the easier roles are automated, expect businesses and technology begin to automate higher value roles in the US. As businesses reduce their costs, some of the savings will be transferred to consumers in the form of lower-priced goods and services. Imagine an insurance company with lower premiums and higher quality of service for the consumer and is more profitable for the business (by leveraging emerging technologies instead of employees and legacy technologies). It’s a win-win for both parties.
These emerging technologies are slowly finding their niches in the workplace and society. The impacts are greater than the few examples I’ve listed and will continue to grow as new applications are identified. Great advances are being made in the arts and design, where the outcomes remain to be seen but can be expected to be great.
As mentioned, I’m also pessimistic about these changes and the larger impact they’ll have on society. Can we continuously automate jobs without negative impacts? I don’t think so. Eventually, I think we’ll automate so far, we risk collapsing society inwards, through decreased demand.
First, though, automation will provide the benefits mentioned above and businesses will race to identify opportunities for automation. The race will improve AI & ML, decrease costs, and identify new applications. We’ll quickly move through the low hanging fruit, repetitive jobs, and tasks with high error rates, like manual data entry. Then we’ll move to roles previously difficult to automate, like cashiers and drivers. While we won’t get through all of the lower class jobs, we’ll eliminate enough to significantly impact this already burgeoning group of people. Their currently high unemployment rate will continue to grow and the question is, what jobs will we create?
The automation revolution will have several winners. Computer scientists and developers will see increased demand for new applications and features. Consultants and project managers will be requested to identify opportunities and implement solutions. The Lean Six Sigma gurus will calculate effectiveness and help fine-tune processes to extract further value from what has been automated. There’ll also be a group of people who train, monitor, and test automation products. My concern isn’t that jobs won’t be created, I strongly believe they will. The concern is the jobs created will require skills too far advanced for those displaced. As a society, we can retrain and expect citizens to improve skills over time to meet increasingly complex business demands. But is there an end to this? Can we request our truck drivers to learn python and our cashiers to learn java? Some of them will, but eventually, we risk requesting too much.
Does this have to happen?
I’ve seen several thought leadership pieces about the direction automation will go. Some mention we’ve been anticipating this revolution for over a hundred years, and we’re still waiting, which is true. Others have taken a seriously negative stance, decrying it as the end of society as we know it, which remains to be seen, but likely unfounded. Lastly, some point to the historical industrial revolutions where improvements in technology have created entirely new and unforeseen industries, which is true. So there are other thoughts on if this will happen, the impact, and what we can do it about. In the US, a fundamental economic equation drives my concerns. Of the US GDP and economy, two-thirds (+60%) is driven by consumer spending. If we reduce real incomes through increasing labor supply by automating jobs, consumer spending will decrease. I believe the US needs to push companies to provide higher salaries, at all levels, but in particular the lower levels. US companies are more profitable than ever, they’ve never had more money on their balance sheets, and at the end of the day, the wealthy consumers can only spend so much. Injecting capital at the lower end of the economy has several positive impacts (health, education, etc.) and is reinjected into the economy through spending. Injecting capital at the upper end of the economy, through tax cuts usually, leads to increased investments in the stock market, which doesn’t increase GDP growth and isn’t reinjected in the economy.
In the US, a fundamental economic equation drives my concerns. Of the US GDP and economy, two-thirds (+60%) is driven by consumer spending. If we reduce real incomes through increasing labor supply by automating jobs, consumer spending will decrease. I believe the US needs to push companies to provide higher salaries, at all levels, but in particular the lower levels. US companies are more profitable than ever, they’ve never had more money on their balance sheets, and at the end of the day, the wealthy consumers can only spend so much. Injecting capital at the lower end of the economy has several positive impacts (health, education, etc.) and is reinjected into the economy through spending. Injecting capital at the upper end of the economy, through tax cuts usually, leads to increased investments in the stock market, which doesn’t increase GDP growth and isn’t reinjected in the economy. American consumers need to demand effective and meaningful Corporate Social Responsibility (CSR) from the companies they shop at. By voting with their dollars, they can create meaningful change in corporate policies. We’ve seen this in several companies recently, in particular with the Trump and Breitbart brands, which were dropped by several retailers and online advertisers. This will require a commitment and coordination that’s rare, so we’ll have to see if there’s consumer appetite to wage such a battle.
American consumers need to demand effective and meaningful Corporate Social Responsibility (CSR) from the companies they shop at. By voting with their dollars, they can create meaningful change in corporate policies. We’ve seen this in several companies recently, in particular with the Trump and Breitbart brands, which were dropped by several retailers and online advertisers. This will require a commitment and coordination that’s rare, so we’ll have to see if there’s consumer appetite to wage such a battle.Conclusion
I wrote this piece for myself, to think through some ideas I’ve had on automation, and it’s very short in comparison to all of the facts, information available, and all the different ideas I’ve had. It serves as a starting point, to develop discussions further, and encourage others to think about how this will positively or negatively impact our future and what we should do. I don’t think it’s the end of the world and any change in society would be a slow downhill trot, not a rapid fall. We’ll have the opportunity to use this for good and continue to improve our economy in the right direction if we choose too.
These are my thoughts. Let me know yours in the comments.