Preparing Ourselves for AI and Job Automation

How industries are being automated and how we can survive and thrive


Raj Shroff

2 years ago | 16 min read

This article first appeared in The Startup

In the United States, a businessman named Andrew Yang ran for President in 2020 on the promise to give every American adult a universal basic income of $1,000 a month. That’s $12,000 a year, in cash, no questions asked.

Yang is not a socialist that just wants to redistribute wealth. He recognized that the most common jobs in America are being automated away and that people and communities that depend on these industries will struggle. He realized that Artificial Intelligence will transform entire industries and automate both blue-collar and white-collar work. He talked about this in his book The War on Normal People.

Yang planned to implement a value-added tax on certain goods and services such as Amazon sales, Facebook ads and robot truck miles to help pay for his $1,000 ‘Freedom Dividend.’

I am not qualified to comment on how good his policies were but Yang correctly realized that AI and automation will drastically impact many popular jobs. It is already happening in manufacturing and will soon spread to most blue-collar and white-collar industries.

While AI & automation will undoubtedly increase productivity, they will also target many common jobs — routine jobs that require high school education (warehouse or call center work), white-collar data processing jobs (analysts), and even some medical and manager-level roles. Any job that is mostly logical, repetitive, and routine is a great candidate for automation.

Now consider that the U.S. Federal Reserve classifies approximately 62 million jobs as routine — or 44% of total jobs. That is a lot of potentially automatable jobs in just one country.

McKinsey predicted that one-third of U.S. workers could be automated away by 2030 and that between 400 million and 800 million people globally could lose jobs to automation.

Job types at greatest risk of automation. Source: The Economist
Job types at greatest risk of automation. Source: The Economist

This is a global issue. Surprisingly, very few world leaders and educational institutions are taking steps to lessen the impact on the workforce or prepare us for a radically different future of work.

The good news is that we can, as individuals, prepare ourselves for the AI & automation wave. We don’t have to become coders. However, we can all strengthen certain natural skills and take actions that make us less replaceable and more valuable.

The first step is understanding how automation can affect our jobs and career paths. Jobs and careers that we used to think of as ‘safe’ and ‘lucrative’ in finance, law, and even medicine will decline in the near future.

AI & Automation Will Affect Many Common Jobs

AI and automation are already replacing many blue and white-collar jobs. That includes both office work and non-office work, low skilled and high skilled work. Automation will increase as businesses embrace the cost and productivity benefits, as they well should.

Routine jobs are at most risk from AI and automation. Doctors, lawyers, accountants, analysts, traders and journalists who perform routine activities are potentially automatable. Jobs requiring advanced degrees and specialization are at a surprisingly high risk of automation.

These jobs won’t disappear overnight. However, we may soon see far less demand for these jobs in their current form, especially in developed countries.

Importantly, automation will also affect manager-level jobs. When there are fewer workers to manage, companies need fewer managers and may pay managers less since they have fewer direct reports.

Manufacturing & Warehouse Work

Jobs affected: Factory, assembly line and warehouse work

During a Democratic debate in Ohio in October 2019, Yang mentioned that most of the decline in American manufacturing jobs was caused by factory automation. AP fact-checkers agreed, citing research showing that 88% of job losses between 2000 and 2010 was due to automation.

Warehouse work is also ripe for automation. Amazon is the most prominent example of large scale warehouse automation but large warehouses and logistics centers around the world are also following suit.

Boston Dynamics has these cool AI robots that do warehouse work. These warehouse bots use computer vision tech to see where they are going and identify boxes to handle. They also don’t waste time sleeping, getting sick, or forming unions. They certainly don’t need wages or health insurance.

Source: YouTube
Source: YouTube

Call Centers & Customer Service

Jobs affected: call center agents, customer service reps, telemarketers, some sales agents

Call center work is ripe for automation and is already underway. AI-enabled chatbots and conversational interfaces are being used to handle a wider range of queries.

According to Statista, 66% of global call centers are located in the Americas. India and the Philippines are also leading players due to outsourcing. Banks, telecom and consumer service companies tend to have the largest call centers.

At the same time, Deloitte found that 56% of call centers plan to invest in AI in the next two years and 33% plan to invest in robotics and automation. Cost-cutting is only one reason— companies lose $75 billion annually to poor customer service. Customer dissatisfaction is expensive.

Chatbots and conversational AI answer routine questions instantly, reducing call times and increasing customer satisfaction. The tech is getting better at handling more complex queries (AI bots learn over time) and understanding nuances in human conversation.

Amazon Lex provides tools for call centers to build conversational chatbots using voice and text. Kasisto is another chatbot vendor targeting financial services. Check out their demo for Mastercard here. On the CRM side, Rocketbots helps companies automate customer support, outreach, and lead generation.

In the short term, we might actually see call center employment increase. Chatbots will start handling way more call volume, but complex issues will still need human intervention. This could create short-term demand for call center workers.

As the tech gets better, however, we will start to see many call center jobs get automated away. This is a concern in places such as Asia, where call center work is currently plentiful and can pay reasonably well.

As AI gets better at holding human conversations, it could take away certain jobs in sales and telemarketing, for example.


Jobs affected: truck, taxi, Uber, bus drivers

Self-driving trucks are here and will soon automate away many trucking jobs in developed countries. Over 3 million Americans are truck drivers and 7 million work in trucking-related activity. Not to mention the communities that are built around trucking. The average American trucker is a 49-year-old male high school graduate. How will they find comparable wages without a college degree at that age?

Self-driving vehicles, in general, are getting more sophisticated and will soon threaten taxi, Uber, and bus drivers. Yes, taxi unions and regulators will fight autonomous vehicles as long as possible, but not in every country.


There are roughly 4 million secretaries and admin assistants in the U.S. according to the Bureau of Labor Statistics (BLS). The BLS also predicts a slow 3% job growth rate for this function between 2014 and 2024.

That is because AI tools such as virtual assistants and financial, email, and calendar management software are already being used. Consider some admin assistant functions and AI software designed for them:

  • Scheduling appointments: creates and manages meetings and integrates with Salesforce, Zoom, and Slack
  • Managing email: Knowmail learns your email habits and filters the most important and actionable emails for you
  • Document management: Filee enables paperless storage and management for things like documents, invoices, and receipts.

If you need proof of how good virtual assistants have become, check out Google’s Duplex virtual assistant making a phone call to a hair salon and talk to an actual human to book an appointment. The cool thing is that the conversation sounds natural, and the human doesn’t suspect it is talking to an AI.


Jobs affected: investment/financial analysts, investment research, traders, junior bankers, financial advisors
UBS trading floor in 2008 vs 2016. Financial crisis plus automation. Source: AtozMarkets
UBS trading floor in 2008 vs 2016. Financial crisis plus automation. Source: AtozMarkets

Financial analysis, investment research and trading activities are candidates for automation because they are logic and process-driven. Consider what financial professionals do on a daily basis:

  • Recognize patterns in data and extract unique insight
  • Create models and decision rules
  • Anticipate future events

AI technologies such as machine learning and natural language processing are already being applied to these tasks. Bloomberg reported that Wall Street reached “peak human” in 2016 and that firms will start shedding jobs in favor of automation tools.

Daniel Nadler CEO of financial analytics firm Kensho had the following to say:

“Analysts, young associates, vice presidents — anyone whose job is moving a column of data from one spreadsheet to another is going to get automated.”

Note that he includes vice presidents — experienced professionals with years of experience. It doesn’t matter what your job title is — if parts of your job contain routine & logical processes, it can be automated.

There are many AI startups providing automation solutions for finance. Apoidea specializes in extracting information from company reports and other documents, summarizes key information, and performs sentiment analysis by monitoring traditional and social media. Banks currently employ armies of very well paid entry-level and mid-level analysts for this work.

AI tools will undoubtedly make banks and investment firms more productive. However, these firms will also require significantly fewer employees in the near future.

Bookkeeping, Accounting, Data Entry/Analysis

Jobs affected: junior accountants, clerks, data analysts, reporting specialists

Regardless of job title, these types of roles deal with routine, logical and repetitive data tasks such as data entry, processing, transformation, and analysis. This includes accounting entries, invoicing activities, payment processing, data validation, and management reporting, among others.

Accounting software such as QuickBooks has been around for a while, meaning that small and medium-size businesses have fewer reasons to hire bookkeepers and accountants. Once these tools become better, we may see less demand for these roles at larger corporations as well.

Software Powered by AI and Robotic Process Automation (RPA) tools are automating data entry and document processing. This software can convert text from scanned documents (e.g. invoices, contracts) into machine-readable text in Word or Excel format. It can auto-populate document templates or databases with info extracted from scanned documents, and even create reports. These tasks currently require lots of manpower, but that will change.

Legal & Compliance

Jobs affected: lawyers, paralegals, record clerks, KYC & compliance professionals

Deloitte predicted in 2016 that 39% of legal sector jobs will be automated, affecting young lawyers, paralegals, and legal secretaries.

A large (and unglamorous) part of working in legal & compliance is document analysis. This includes searching for legal precedents or specific clauses in mountains of documents.

JP Morgan Chase launched a tool called COIN (Contract Intelligence) for intelligent document search. It allows the bank’s legal team to review large volumes of legal documents and extract specific clauses in seconds. COIN uses Natural Language Processing (NLP) and saves the bank’s legal team up to 360,000 hours (15,000 days) annually.

Once these tools become widespread, corporations and law firms will have less need to hire armies of law school graduates or even some experienced professionals.


Jobs affected: radiologists, surgeons, diagnostics professionals

There are numerous examples where AI identifies cancers and tumors better than human radiologists. Turns out that AI can ‘see’ shades of gray on CT scans that are invisible to human eyes.

Robotic surgery is also becoming more refined and may require human surgeons to serve more as supervisors than actual surgeons.

Robot assisting with intestinal surgery. Source: Nature
Robot assisting with intestinal surgery. Source: Nature

While this is great news for healthcare and will enable doctors to devote more time to patients instead of medical processes, demand for some specialized medical professions may fall in the future.

Journalists & Writers

Around one-third of the content published by Bloomberg News uses some form of automated reporting. The tech can analyze a hundred-page company annual reports in seconds and produce an immediate news story with relevant summary analysis. Other news organizations including the Associated Press and the Washington Post are using similar methods.

Journalism students beware. For that matter, authors beware, too. A novella, co-written by an AI and human assistants nearly won a Japanese literary prize, showing that it’s only a matter of time before AI can write creatively.

How long before an AI writes way better articles about AI? Not long. People like me are doomed.

Counterargument: AI Will Enhance Jobs

Despite evidence that AI will take away jobs, it is naive to assume that AI and automation will lead to immediate and wholesale job losses. The truth, as with most things in life, is more complicated.

Truckers and data analysts probably won’t get replaced overnight. Trucking companies can’t replace their fleets with self-driving trucks that quickly. Human drivers will still need to at least sit in the truck because people and regulators don’t trust self-driving tech yet.

Corporations have convoluted and evolving data processes that require human analysts to work with AI tools. Before they can automate data processing and analytics, companies must undertake the lengthy tasks of organizing and cleaning their data. Most of this data is distributed across old systems and incompatible file formats.

There will be a transition period away from these jobs. New tech always comes with bugs and takes time to integrate into work and life. Not all firms have the money to fast-track automation.

There is also the valid argument that automation tools make work less mundane, allow us to be more productive, and find meaning in our work.

For employees and managers, AI-powered analytics can lead to better decision making, better job performance, and better rewards. Employees can focus on more meaningful and impactful work, which benefits mental health and self-worth.

Socially, AI tools can improve the average person’s life. This can be as simple as Google Translate improving your travel experience, or a mobile AI tool that helps African farmers detect diseased crops using only a smartphone camera. AI-enabled medical imaging is already better at identifying tumors and other diseases than human doctors, which is great news for patients.

Proponents of AI argue that technology will get integrated into different careers in different ways. Some jobs will be eliminated while brand new jobs will be created. On the whole, they argue, AI will become a productivity aid. But while there will be new jobs created, workers displaced by automation may not qualify for these new jobs.

Some also suggest that wholesale automation jobs aren’t reflected in the data yet. However, these data metrics are lagging indicators. By the time the data does reflect widespread automation, it’ll be too late.

Overall, promoting only the benefits of AI misses the wider point. Yes, AI & automation will improve jobs and lives. At the same time, it will create a painful transition period where many common jobs will dry up. This will affect workers who may find it hard to retrain themselves. This outcome is what we must prepare ourselves for.

Protecting Ourselves from AI and Automation

Understanding the threat to jobs from automation is fine and dandy, but how can we protect ourselves, and even thrive in this new world?

While everyone’s situation is unique, there are general actions that most of us can take. We need to understand how automation will affect us personally and develop skills that are hard to automate. Luckily, none of this requires learning to code or getting an advanced degree, because not everyone has the time or interest.

Understand how automation affects you

How will AI & automation affect your job and career path in the near future? How is your company/industry adopting automation? Which types of employees will it affect? How automatable is your current job?

If your job is automatable, what can you do to either change functions or embrace technology to come out ahead? Perhaps your company offers training to use software and perform your job at a higher level, thereby becoming more immune to automation.

If you understand your role and function well, perhaps you can help your employer implement automation tools. That way, you can control the pace of change for you and your co-workers. In general, companies value employees that improve efficiency, and this is a great way to maintain job security and even get promoted.

If your company is implementing new technology, try to get involved in these projects even if you are not a technical expert. Perhaps these projects need project managers and business analysts to coordinate between the business & tech teams. You can acquire first-hand knowledge of the tech being implemented, which makes you valuable in your industry.

Listen to what business media and CEOs are saying about how automation is being adopted in your industry. Keep tabs on tech startups that are raising serious venture capital money — and see if these startups’ products will impact your job. This can help you forecast what your future job will look like.

Finally, stay aware of what’s happening at your own company across departments. For example, in a previous job at a large insurance company, I saw that we were implementing digital transformation projects — enabling customers to buy and manage insurance policies online and through mobile devices. This is an industry-wide trend and has clear implications for insurance sales agents, policy administration staff, and customer support.

Build skills that are hard to automate

Studies show that work requiring strong communication, relationship building, trust, empathy, critical thinking, persuasion, and people management are harder to automate.

Think about a great teammate or manager you’ve had in the past. Perhaps they were great at motivating you and your team to work towards an objective. Perhaps they were great coaches who made others better. They probably made great decisions by looking at the big picture and thinking critically. Could an AI do all of this? Probably not. AI is great at a narrow range of tasks, but humans will always beat AI when it comes to skill variety.

Jobs that require building human connections are harder to automate. Few customers will purchase goods or services worth millions just by talking to a chatbot. They would rather trust a good relationship manager or salesperson who understands their unique needs and make them feel comfortable.

Finally, using plain old common sense is a great way to set ourselves apart from AI. At this stage, AI can surprisingly dumb and is only as good as the data it is fed. It still needs a human to perform sanity checks and act as the ultimate decision-maker. Thinking critically and being more than a cog in the machine is a good way to protect ourselves from automation.

Become a great storyteller and influencer (not on Instagram)

Persuading people, keeping them engaged, and motivating them to action are things that no AI can do.

The best TED Talks captivate your attention and fill you with a sense of optimism or purpose. That’s because TED Talks are stories. They convey a clear message and get the audience to empathize and relate to the storyteller. Steve Jobs did this at every Apple presentation and that’s probably why you own an iPhone or an iSomething.

These skills are invaluable at work. Whether you’re presenting to a customer, manager, or a room full of executives, cold facts and logical arguments only work sometimes. A compelling story that includes facts and logic goes even further because it engages listeners and makes them wonder what comes next.

Employees that make convincing arguments and articulate solutions using compelling stories are seen as problem solvers and ‘doers.’ They are trusted with bringing in new customers, business, or funding. They are often promoted and given raises, not automated away.

All good presentations contain storytelling elements. They talk about a great future and how to get there (e.g. the project you’re requesting funding for will open up new markets). They contain elements of conflict (e.g. the current budget won’t allow you to staff your project). They regularly remind listeners of the stakes and benefits (facts, figures, and logic). Nancy Duarte’s Guide to Persuasive Presentations is a short read that covers these concepts well.

It is very hard for an AI to tell a story as Steve Jobs could. You don’t need to work at a Fortune 500 firm to develop these skills, either. Think about the best presentations you heard in your company, on a TED Talk, or somewhere else, and incorporate these elements into your own presentations.

A side hustle that makes you happy

The ultimate insurance against job automation is being able to make money in different ways. You don’t need to launch a VC-backed startup. A small side business doing something you’re passionate about can bring in extra income and increase your happiness and self-worth.

Never stop learning

Even a little knowledge can make a difference. Knowing about things outside your job brings opportunities to move to a new project, team, or even a career.

Remember that AI is more likely to automate highly specialized tasks. If your knowledge is limited to your own job, you are at a higher risk of automation. The remedy is to acquire a wider skill set, especially since knowledge and skills quickly become outdated.

Learning new skills doesn’t always require spending money going back to school (guilty) or getting professional certifications (also guilty). Skills that are fun to learn can be valuable as well, and not all skills need to be applied in an office setting. Mixing cocktails, DJ’ing, or learning how to produce a vlog can bring both profit and popularity. Did you know that Goldman Sachs’ CEO is a part-time DJ?

Try free online courses on Coursera or edX. You can also learn almost anything on YouTube. If you want industry knowledge, try attending industry events and conferences.

Final Thoughts

AI and automation will change the way people work. Today’s popular jobs won’t be so hot tomorrow. While most jobs won’t disappear entirely, automation will displace a large number of us. Fortunately, we can prepare ourselves by augmenting natural skills that are hard to automate.


Created by

Raj Shroff

I write about how AI and blockchain can help business and society.







Related Articles