Will GenAI Agents and Copilots Bring People Back to the Office?

It is very likely that some companies will bring employees back to the office – at least partially and for a period of time – to observe, measure, and optimize productivity gains from GenAI Agents and GenAI Copilots. Working entirely remotely makes it hard for companies to understand how much productivity can be gained, how quickly it can be achieved, and how to ensure there is proper and secure usage along the way. 

Let’s first explain what a GenAI Agent and a GenAI Copilot are, and how each technology can boost employee productivity independently, and how the combined usage will enable us to maximize productivity:

A GenAI Agent is an autonomous AI system that can make decisions on its own, take actions, and execute multistep tasks independently with no (or minimal) human intervention.

A GenAI Copilot is an assistive AI system designed to work interactively alongside humans, enhancing productivity by providing suggestions, recommendations, or partial automation –  but always keeping humans in control.

So, if you need AI to fully automate tasks, you need a GenAI Agent, but if you need AI to assist you but you are keeping control, then you need a GenAI Copilot. Agents do things for you while Copilots help you do what you do.

Therefore, a human can use a GenAI Copilot to assist with thinking, writing, coding, or decision-making, and use a GenAI Agent to execute repetitive or autonomous tasks based on instructions.

Let’s now explain the potential productivity boosts expected with GenAI Copilots and Agents:

• GenAI Copilots – It will depend on the nature of the task, industry, level of AI integration and the experience and knowledge of the employee. So far, studies suggest that employee productivity can increase by 30 to 80% in certain cases. GenAI Copilots help humans think, create, and make better decisions faster.

• GenAI Agents – Productivity gains are expected to be greater than with GenAI Copilots, especially for processes that can be automated, repetitive, or data-driven tasks. While copilots assist humans, agents can fully automate workflows, leading to productivity increases of 50%–300% depending on the task and work ecosystem. GenAI Agents handle execution and automation, removing manual effort.

Using both GenAI Agents and GenAI Copilots together can create a massive productivity boost, potentially doubling or tripling efficiency (100%–300%); thus, working together in a coordinated way, they enable true human-AI collaboration:
- Copilots assisting in brainstorming, writing, and decision-making.
- Agents executing tasks, automating workflows, and taking actions.

The combined gains can represent 20-50-plus hours per week, or 2x-5x faster work while reducing manual effort. So, as an employee, will I keep working 40 hours per week but delivering 2X value (throughput, deliverables, among others), or will I work 20 hours only and deliver the same value I deliver today? Who will make those decisions in a remote work environment?

And if you wonder whether productivity improvement is only for new workers, it turns out that both experienced workers and new workers can gain productivity from GenAI Agents and Copilots, but the benefits vary based on their skill and knowledge levels, the nature of the specific processes as part of their workflows, and how they use AI:

• New workers: gain faster learning and skill-building, reducing the time needed to become proficient.

• Experienced workers: enhance efficiency by automating repetitive tasks and focusing on high-value work.

And of course there are risks! Deploying GenAI can bring significant benefits regarding employee productivity, but it also introduces some relevant risks. Allowing GenAI Agents to operate without human supervision comes with several risks, especially if they are making autonomous decisions. GenAI Copilots come with some risks as well, even though they are designed to assist humans rather than act autonomously. While they are generally safer than fully autonomous GenAI Agents, they can still introduce errors, biases, security concerns, and over-reliance issues. GenAI Copilots are powerful assistants, but they should be used responsibly to avoid risks. They work best when combined with human expertise and critical thinking.

So why would companies bring employees back to the office due to – or for –  GenAI adoption? Here are some very meaningful reasons why:

1. To measure productivity gains more accurately: remote work makes it harder to observe real-time AI usage and benefits.

2. To ensure proper AI adoption and training: bringing them back to the office allows for hands-on training and real-time coaching.

3. To prevent AI misuse or over-reliance: some employees might over-rely on AI, leading to lower creativity and collaboration.

4. To maintain data security and compliance: with more access to sensitive data, in-office AI usage ensures the right levels of control regarding security protocols.

5. To enhance team collaboration and innovation: brainstorming tends to be more impactful in the office and enables tighter collaboration.

I don’t think this will mean the end of remote work, but I think there will be a rebalance in terms of how and when hybrid models remain pervasive – again –  in the future. Companies will have to fine tune their right balance: office work for AI training, monitoring, and optimization; and remote work for deep-focus tasks, potentially adding AI analytical tools for remote productivity tracking.

All of this while GenAI will get cheaper, smarter, and more accessible beyond what we initially thought. Innovations like the one from DeepSeek will make AI in general more likely to be used across smaller organizations, not only the larger ones with big investment budgets.

As a final note, this is the first time that I have used GenAI to help me write an article. Being the first one, I can tell I gained some level of productivity (maybe 15-20%), but this time I did spend more time thinking and reflecting on the subject, so in the end, I think I did a better job!