In 2024, there will be more than 2500 passenger aeroplanes on order by airlines around the world. Customer care centres around the world are seeing an increase in call volumes and are finding it difficult to staff positions. The National Health Service (NHS) in the UK is seeing a median wait time of fourteen-and-a-half weeks. Robots do not seem to be replacing humans in any of the three examples we covered.
A 2020 report from the World Economic Forum states something interesting: while 85 million jobs may be lost by 2025, about 97 million new jobs will emerge. So, what we must worry about is not about the robot taking away our current job, but about how we can remain relevant and useful without necessarily being tied to the requirements of our current job.
To understand this better, let’s look at how technology works at the interface of the human and the office. If you walk into any office environment, you will see technology deployed for a variety of activities. The office you visit might have an automated entry system or it might be using AI-based virtual office assistants. Using technology, organizations have been able to improve the productivity and efficiency of employees while also bringing in the benefits of better work-life balance and experience. Employees can now focus on more important tasks and spend more time with fellow workers on core activities. The rise of AI and the benefits it brings are another chapter in the office’s progress.
Generative AI has been in development for a while and will continue to get better. It is built on what is known as a foundation model. In simple terms, this means training the AI system with large amounts of data, including text and images, so that it slowly starts to recognise patterns in the data and make inferences and predictions. Once it has reached a certain level in terms of learning, the system is finetuned for accuracy using more precise data sets. To give an example, think of creating an AI model to recognize a shark when presented with an image containing different ocean creatures. The model is first trained on all the information about the living and non-living things in the ocean. Then the model is fine-tuned by showing it a variety of sharks until it can pick out a shark from any image shown to it. The best part is that the model improves with increased use and interactive learning. The initial training phase is time-consuming and intensive, but after the machine is trained, the benefits are exponential.
Interestingly, early AI inventors worked on mirroring how the human brain worked. The brain is the most powerful computational engine even today, and many machine learning techniques come from the way our brains operate. The billions of neurons in our brains form a network, facilitating our learning and thought processes. The scientists who worked on artificial intelligence have tried to mirror the networks in our brains to build on foundational models to give us deep learning machines that are close to human intelligence. Today, we can have machines process large and unstructured data to perform several activities that were once in the domain of humans – answer questions, make meaningful inferences, suggest actions, interpret images, guide vehicles safely and so on.
Soon, many office tasks that have been so far performed by humans will be fed into machines, who, over time, will learn to first duplicate and then improve these tasks as they learn. As a human, the advantage you have is that you can keep ahead of the machine, as your brain is at a level of sophistication that is difficult to replicate in a machine. You cannot compete with the machine by doing the jobs of the past, but you can be ahead in the jobs of the future.
Office workers can segregate their work in an office environment using multiple dimensions such as skill, complexity, interface, value, etc. To exemplify this, let’s categorise the jobs performed at a hotel using the dimension of complexity. At Level 1 of complexity, we have what are known as interface jobs that involve processing routine transactions like check-in or answering queries from customers and providing a resolution. At Level 2, you have jobs that require an added dimension of decision-making. For example, if you are a frequent guest or member of a hotel chain and the room you booked is not available, you will need some assistance to resolve this. The front desk will probably call a supervisor, who will then make a decision that could either get you an upgrade or a drink voucher that you can use while you wait for your room.
At Level 3, you have ambiguous situations that require judgement and decision-making based on experience. For example, the hotel might want to increase the room tariff, knowing that a large convention is scheduled to take place in the city. This is a judgment call, as increasing the tariff by too much would annoy potential guests, yet keeping the hike low would mean losing some profits.
In the past, a hotel employee would start at a Level 1 job, then get promoted to the role of supervisor and eventually move to becoming a general manager. Technology has now altered the skill set required at each level. Let’s see how.
At all three levels, technology can benefit us in different ways. At Level 1, technology can eliminate the need for a front desk, as some hotels have already done in the past.5 On arrival, you could swipe in your credit card at a kiosk and receive your key from the machine. Present your key to the machine at the time of leaving and you will be checked out automatically. However, people don’t normally like checking into a hotel where they deal only with machines. Eliminating the job doesn’t mean eliminating humans. Hotels have realised that while technology replaces humans in some ways, it still needs them. The person manning the front desk still appears and has a chat with you, but the skills required have changed. What are the new skills required? The ability to speak multiple languages, engage in meaningful conversations, suggest activities, cross-sell some packages and most importantly, bring in the human element to the interaction that will make the guest want to come back to the hotel. The job has in fact been split; the machine has taken over the uninteresting part and the human is handling the interesting one.
At Level 2, the supervisor could be supplied with guest profiles and pertinent information based on their previous stays and create a suggested list of actions they could take without needing to make the guest wait. For example, if the guest is a business visitor, an offer of a chance to refresh in an available room and a taxi ride to a business meeting would be more useful than a voucher at the bar. The supervisor could instruct the Level 1 worker at the front desk in advance, avoiding any unforeseen requests or problems that would require the supervisor’s assistance.
At Level 3, the system could intelligently analyse the data available from previous events to present a “what if” analysis on room rates. The general manager could then use these inputs to price the rooms in a range that appears reasonable and does not leave the guests with a feeling of being cheated, while at the same time optimizing the revenue for the hotel. The judgement of the human would still be very vital in deciding the final pricing, as only someone with ample experience can make the right call in this situation.
We can see that at all three levels – front desk, supervisor and general managers – the deployment of technology has enabled them to move to a higher level in their jobs and allowed them to learn new skills to make them more productive.
This can be employed in other businesses, such as airlines, supermarket chains, professional service firms and so on. It is clear that technology can enhance operations by automating menial tasks and allowing humans to focus on quality work and decision-making, no matter the industry or the job level.
Humans are better at making decisions involving ambiguity. Ambiguous situations require decision-making based on intuition, and humans are good at this. Machines are far superior at making decisions that require deductions involving data and logic, especially data of considerable volume. Merging the two capabilities of machines and humans makes for a powerful outcome.
Excerpted with permission from Human At Work: Arm Yourself To Thrive In a Fast-changing Workplace, Richard Lobo, Penguin Business.
