Companies and their fintech partners are rolling out a new generation of agentic AI agents capable of independently completing assigned tasks. The proliferation of these agents across the business world is bringing new urgency to the question: Will AI replace human workers?
As an executive of a fintech company that just launched a suite of agentic AI agents in capital markets and wealth management, you can believe I’ve given this issue a lot of thought. As I’ve considered it, I’ve come to realize we’ve been asking the wrong question. Instead of asking if AI will replace human workers, we should be asking how we can use AI to create bionic workers that supplement their own intelligence and common sense with the powers of AI.
In the 1970s, one of my favorite television shows was The Six Million Dollar Man. The show followed the adventures of U.S. Air Force pilot Steve Austin. After Austin crashes in a test flight, government scientists rebuild him with bionic implants that give him superhuman strength, vision and speed. The procedure, and the technology, cost $6 million.
AI is giving us equivalent powers in business. Equipped with “bionic” tools that automate research, writing and communication, data analysis, and administrative tasks, the typical worker today might just be the equivalent of Steve Austin in the 1970s.
At the same time, the rise of agentic AI appears to have created competition for our workforce of bionic people. Agentic AI allows companies to create agents that, when assigned an objective, can plan the steps needed to accomplish it, access the required tools and autonomously complete the task. AI agents can (with proper permissions and controls) independently access databases, customer resource management platforms, e-mail, calendars and even the enterprise resource planning systems that help run companies.
Jobs Are More Than Tasks
Some large companies have already deployed swarms of AI agents to automate regular tasks. AI agents are answering customer questions and processing refunds. Some of my personal experiences have been awful, and others impressive. AI agents are troubleshooting IT glitches for help desks, they are generating purchase orders, processing invoices, launching marketing campaigns and producing sales leads.
As I’m sure you’ve noticed, these are all tasks that, until this moment, have been completed by human workers. With AI agents taking on all this workload, there will be fewer simple mechanical tasks for humans to do in these specific functions.
The key phrase here is “specific functions.” While AI agents can be highly effective, their work is also narrow in scope. In each of the functions listed above, the AI agent accomplishes exactly the assigned task — nothing more, nothing less.
But anyone who has ever worked in business knows that there is much more to every job than just a list of core tasks. In modern business management, that observation is known as the “Doorman Fallacy,” or the mistake of judging a role by its most visible task while overlooking other hidden work and value it creates.
The term, of course, comes from building doormen, who on the surface seem to get paid to open doors for residents and guests, but in reality perform many other functions, helping tenants manage their lives and providing a sense of warmth, comfort and security for the building. In the same way, professionals in procurement, purchasing, accounts payable, finance, customer service, sales, marketing and all other business departments perform a host of functions and provide value well beyond their lists of core tasks.
When you zoom out to take in everything these jobs really entail, it becomes clear that AI agents probably won’t cut it as a pure replacement that could totally eliminate them. On the other hand, AI agents will be able to ratchet up productivity in these jobs to unprecedented levels — if they are managed by bionically enhanced human workers.
More Demand For Bionic Workers
Those massive productivity enhancements lead us to another principle of business management and economics: Jevon’s Paradox. Jevon’s Paradox is the idea that efficiency increases aimed at reducing resource consumption can make a product or service so much less costly that it increases demand, ultimately leading back to more resource consumption.
One of the clearest examples of this phenomenon comes from the computer industry. Over the past 50 years, computing power has become exponentially more efficient. Those massive efficiency gains might lead you to expect that companies’ costs for computing would have fallen dramatically, leading to lower costs for companies. But as we all know, that’s not the case at all. Today, companies’ computing budgets are skyrocketing due to soaring demand for the cheap “compute” they have adopted to run virtually every aspect of their businesses, including AI.
I believe the same thing will happen with human jobs and AI. As AI makes certain work cheaper, companies and consumers will use more of that work, leading to increased demand for the work overall. We are already seeing this play out in certain fields.
For example, just a few years ago, AI was getting so good at reading results from CT scans, MRIs, ultrasounds and other tests that people were predicting the demise of radiologists. What happened instead is that AI made this testing cheaper, leading doctors to order more tests, which has increased demand and the number of radiology jobs. (The Doorman Fallacy also played a role here, since radiologists do far more than just read test results.)
Anecdotally, I have seen the same process play out in my own work and among my clients. As AI lowers costs, companies are developing and deploying more software, using more financial research and running more marketing campaigns. In these and other areas, the only way for companies to keep up with this surging demand will be to pair powerful AI with the valuable contributions of bionic human workers.