Despite the fact that machines may perform most analytic functions at increasingly automated companies, investors will still need a human to put a brave face on a bad quarter and 'manage the messaging' to comply with various financial regulatory schemes.
While the 'science' around marketing has become increasingly automated in the form of marketing analytics, humans are not the 'rational actors' we learned about in basic economics -- especially when it comes to interpreting trends and purchase behavior (Pet Rock, anyone?). Humans will increasingly need to interpret product and sales data to make decisions about how to promote existing products and develop new ones.
There's not much room at the top, especially for automation. I've met a CEO or two with a robotic personality, but machines are unlikely to possess the leadership, interpersonal skills, and intuition that characterize the head of a successful company. These future leaders will, however, have a cadre of machines that provide rapid insights into the health of the companies they command.
Like a mini-CEO, project managers possess leadership and planning skills that automation will unlikely be able to duplicate in the foreseeable future. Rather than teams of hundreds of humans, this 'PM of the future' may have an army of machines at his or her command, ensuring that they're performing complementary tasks and that various interdependencies are kept on track.
In some companies -- and entire industries -- human lives are put on the line every day. Whether it's petroleum exploration or a busy warehouse, dangerous substances and fast-moving equipment can easily end a human life without proper safety procedures. A broken machine can be replaced, but a broken human is a different story. With increasing automation, designing safe ways for humans to interact with machines in these environments will likely require the insight and approval of a human.
As processes become more automated, from production to document handling, someone will need to design the machines, steps, and human/machine transfers to complete a multi-stage process. Machines and machine data may help, but a human will need to lay the groundwork for each new process.
The middle manager has become something of a dirty word, as organizations have attempted to streamline and adopt 'matrix' organizational structures. However, if a large part of your workforce becomes automated, new roles for humans who monitor and oversee groups of machines will be required. This person might do everything from creating budgets for machine upgrades to determining when to schedule downtime for various machines.
The ultimate dream for automation is machines that can learn, update themselves based on that learning, and design and build newer, better machines. Even with the success of this progression, a human will still be required to guide this process. This is similar to the model employed with humans today. Throw your best coder, your best infrastructure person, and your best designer in a room, and who knows what they'll come up with...unless a competent architect is helping to direct the process.
Big data and automated analytics are increasingly the 'detectives' in internal security investigations; however, a human will likely be required to accuse and investigate another human in a case of suspected wrongdoing. It will also be interesting to see how increasing automation affects everything from data security to corporate transparency.
This position obviously doesn't yet exist, and it will probably get a new title. But if wholesale adoption of 'intelligent' machines occurs, a human will need to serve as the final check and balance of a highly automated system. Will machines designed to efficiently share information send data to competitors? Could a machine lie? Do we need a kill switch to disable an automated refinery that could blow up a town if managed incorrectly or maliciously? These are all questions someone will have to ask in an increasingly automated future.