The idea that the office is a specific place where our professional lives happen is becoming less universal, and less important. These days many knowledge workers can be productive anywhere, thanks to smarter, more numerous mobile devices, faster network access, and a growing number of online collaboration tools. And wherever the office may be, wider and better use of social networks, data analytics, and technologies such as voice recognition could be poised to increase productivity dramatically—meaning that both real and virtual offices may have fewer people in them.

But while the physical office is changing, certain connotations of the word “office” are not. I can think of at least two—“hierarchical organization” and “place for human interaction”—that are not becoming any less important. Even the most progressive high-tech companies retain organizational trappings of their industrial-age predecessors: full-time managers, org charts, job descriptions, and so on. And since humans remain social animals, gathering places will remain important in business. These spaces—whether they be conventional offices, temporary ones, or conference facilities—must be made conducive to collaboration. They must also become physically healthy places to spend hours of time, since sedentary work has emerged as a significant health threat.

As the office expands beyond its conventional boundaries, key challenges must be met, including the privacy and security issues posed by a distributed workforce of people who use multiple digital devices. New tools like cloud-based office productivity apps must be made not only user-friendly but resistant to attacks and data loss. And workers will need better tools—including improved voice-recognition software, e-mail-organizing technologies, and intelligent agents that help handle complex tasks once reserved for specialists—to streamline work processes and make sense of the overwhelming volumes of data besieging them.

To date, IT-driven productivity gains within the office have been modest, at least compared with those seen in manufacturing. In 1989 the U.S. manufacturing sector employed 18 million people; by 2009 that figure had declined to 11.8 million. But though the workforce shrank 34 percent, the value added by U.S. manufacturers—the value of their output minus the cost of raw materials purchased—surged 75 percent. We’ve observed white-collar productivity improvement as well, especially since the mid-1990s, but it hasn’t been as big.

That may soon change. Consider that people already routinely deal with computers rather than office workers when they make an airline reservation, buy products and arrange for delivery, or troubleshoot a problem with a product. If a task involves simple and predictable forms of communication without much nuance or emotion, computers can do just fine, leaving humans to handle an ever-dwindling number of exceptions to the usual procedures or questions. More far-out advances in artificial intelligence could push productivity even further. Voice recognition, speech synthesis, and automatic translation have improved significantly. And we’ve seen that computers can now accurately understand and reply to questions: IBM’s Watson supercomputer beat human competitors at Jeopardy! this year. Skeptics will point out that futurists have been promising an AI-driven revolution in knowledge work for decades. But by now even the skeptics are finding phone numbers with the help of computer operators. When the productivity enhancements from these innovations are tallied, I predict, they will be striking.

On top of this, software and social tools can boost the productivity of the remaining human office workers. For example, a customer-­service rep who answers technical questions can work with just one customer at a time on the phone, but it’s easy to handle two or more simultaneously if the medium is instant messaging. Whole office-based industries may become vastly more efficient. The legal profession, for one, may be in the early stages of a deep transformation. A new breed of legal outsourcing offers much cheaper ways to accomplish certain tasks: contract lawyers and digital tools scan documents during discovery processes, for example. Intelligent software will only get better at finding associations in those documents and mining meaning from it all.

One of the biggest barriers to higher office productivity was articulated best by the late Lew Platt when he was CEO of Hewlett-­Packard: “If only HP knew what HP knows, we’d be three times more productive.” Most organizations, in other words, do a lousy job of capturing relevant information and sharing it among all the people who could benefit from it. But the digital tools available to address Platt’s frustration have gone from inadequate to industrial-strength in the past few years. They include blogs and microblogs, social-networking software, and wiki-style tools that allow collaboration without tightly constraining it. They give individuals a voice, allow groups to form easily and spontaneously, and help knowledge spread. They will be a major force shaping office work in the coming years.

For that to happen, devices and data need to be secure. The prevalence of tablets and smart phones presents a double-edged sword. They help an employee get work done anywhere—but company data goes wherever the worker goes, and the company can’t easily control it. A manager wants to be able to lock you out of your mobile devices if you are fired, so you can’t pilfer anything. And a chief information officer doesn’t want you downloading malware.

For some companies, the iPad model solves the latter problem: no application can run unless Apple reviews and blesses it. Internet-freedom purists may not like that, but CIOs do; they want to sleep soundly at night. In the near future, however, the major mobile platforms will probably introduce some interesting solutions that preserve the benefits of mobile and social computing while imposing security constraints that limit the risk to company data. When the security issues get sorted out, we may finally achieve the full potential of the distributed workforce.

Even as technologies proliferate and their problems are overcome, offices—no matter how virtual—remain collections of people. In my work, I’ve seen a positive feedback loop between what we do when we get together face to face and the ways in which we reinforce those relationships digitally with new tools.

And it’s important to remember that even in this world of freelance and part-time contractors, companies are still desperate to hire good people and retain them. That’s not going to change anytime soon, no matter how many snazzy digital tools we get. The office of the future might have fewer people in it, but the ones who are there will matter more than ever.

Andrew McAfee is principal research scientist at the Center for Digital Business at MIT’s Sloan School of Management and author of Enterprise 2.0: New Collaborative Tools for Your Organization’s Toughest Challenges.