Robotic Resources

As organizations across the globe integrate AI into their workflows, their jobs are becoming slightly easier. So far, AI has mostly played a supportive role—suggesting, aggregating, filtering, or sorting through data that would otherwise be too time-consuming compared to other tasks. Still, its impact is significant enough to warrant continued use and application across a wide range of jobs. Perhaps the only thing keeping it from taking over more tasks is just a matter of quality.

If an AI bot could handle that one chore no one on your team is eager to take on, the proposition would likely become very appealing, very quickly. Now, what’s the one thing standing in the way of better results from AI—and by extension, from your organization? It’s data, mostly. As enterprise software becomes enriched with AI features and gains access to increasing volumes of data, it will improve. Repeat that cycle, and what emerges is an economic flywheel: one that rewards automation over manual labor.

So, what comes next? At some point, most of your organization’s proprietary online data will have been consumed, and the quality curve of AI’s return will inevitably start to flatten. But then there’s the offline data. I’m not talking about robots—though they may play a role someday. I’m referring to AI bots listening in on every conversation during meetings, identifying each participant and their context. Soon, it may develop a complete understanding of what makes your organization tick—and how it might be improved.

Once upon a time, companies used clock-in machines to measure worker productivity in strictly quantitative terms. Fast forward a few decades—or a century—and we began managing this information more comprehensively, tracking other characteristics and data points that could help assess job quality. Think of customer surveys after service interactions, or the humble security camera watching over a factory floor. When COVID hit, organizations had to level up their game in monitoring, too. Still, it largely relied on qualitative data that wasn’t always available. Employees couldn’t be expected to document every single thing they did, and managers didn’t have the time to sift through all those responses.

Then came AI—with the ability to do both: capture the information and process it, then allow others to query it. It makes perfect sense. Tied into calendars, email, and other platforms, AI can provide insights into what was achieved during the week, which deadlines were missed, or any other relevant markers of growth or turnover. But what if that same system could tell you who’s underperforming? Or who made a comment that may have scared off a client? After all, it’s all there—in the data.

Concerns about hallucinations can be addressed by mapping insights to specific moments in time. In essence, people would be monitored—not in the classical sense by a manager or by tracking thoughts and feelings—but by what they say in physical spaces, across meetings and other relevant moments throughout the day. Digital signals would complement that, enabling AI to know exactly what someone did simply by analyzing their patterns and statistics that can be easily retrieved.

Is this the end of privacy? Not necessarily. Some organizations might argue it’s simply capturing business data. Employees might retain a degree of control over what’s recorded in the physical world. Then again, muting a conversation might raise more red flags than simply allowing it to be recorded. Employers may ask why someone wasn’t involved, framing it as a personal issue—or worse, as a sign of unprofessionalism.

Of course, all of this depends on whether organizations are able to combine such data in the first place, are incentivized by competitive pressure, or are simply pragmatic and eager to streamline. These are not unrealistic conditions. Some companies may already be experimenting with such practices behind the scenes. It could very well be seen as a dark pattern of AI—one we should watch closely before it becomes the default.

If societal norms shift to allow these practices—perhaps because they make our jobs easier—we’ll be faced with a real dilemma: do we accept this new way of working as merely more efficient, or is there more at stake? What are the alternatives, and can we strike a balance that gives organizations the insights they need without subjecting employees to constant monitoring and feeding that data back into the system just to make it all seem more tangible?

It’s a question that struggling companies—and, in truth, all of us—will need to confront. And it’s not a choice we’ll enjoy making.