Microsoft Agents: Your new AI assistant

It’s the precise moment we had been fearing all these years. Machines are taking over. Robots, androids and artificial intelligence are beginning to creep into the workforce. Yet, despite everything you may have heard of in your classic sci-fi films, the scenario is nowhere near the apocalyptic descriptions in Terminator or Blade Runner.

The truth is that modern day workforces have a problem at hand; a heavy imbalance between business demands and what humans can sustainably deliver. While 53% of leaders believe that productivity needs to improve, 80% of the global workforce (Work Trend Index Annual Report, 2025), including both employees and leaders, report that they lack sufficient time or energy to complete their work. Furthermore, employees in a modern day-to-day routine will spend a significant portion of their time completing unimportant tasks, with 76% spending up to 3 hours a day on data entry and 90% spending up to 5 hours a day checking messenger apps (Zapier, 2021). People spend time completing tasks that aren’t core to their job function, and which on top of this could be completed by automated processes.

This is where agents come into play. Almost like a personal assistant, Agents are AI-powered tools that help with various tasks, such as providing instant help for a specific problem, automating repetitive processes, and providing real-time insights. These virtual buddies are tailored to particular use cases, meaning that specific topics, tasks or assignments are assigned to them upon creation. By incorporating these Agents into their workforce, companies have successfully tackled the disparity between business demands and human capabilities. Among Frontier Firm employees - those working at companies with organization-wide AI deployment - 71% report that their company is thriving, compared to just 37% globally. Additionally, such employees have reported being able to take on more tasks and are more likely to find opportunities for meaningful work. Employees are essentially having unimportant workload removed from their responsibilities, allowing them more time and energy to dedicate to their core job functions, leading to improved productivity and job satisfaction. As a result, nearly half of leaders (45%) prioritise expanding team capacity with digital labour over the next 12-18 months (Work Trend Index Annual Report, 2025).

Our Agent Story...

Inspired by this shift towards agent-based support, we wanted to become acquainted with these virtual friends. For this reason, we embarked on a journey of our own: creating our first OrangeTrail agent…

The idea was to craft an agent capable of onboarding new employees, potentially integrating it into the introduction process. We envisioned a virtual companion that could guide newcomers through their first weeks – answering questions, surfacing relevant policies, and helping them navigate our culture. Instead of having to sift through extensive administrative content, our agent would deliver and clarify onboarding information in an engaging and simplified manner.

Using Copilot Studio, we built BuddyBen – our new addition to the team - as a retrieval agent. With a simple drag-and-drop interface and natural language, Copilot Studio enables you to build virtual agents in the easiest way, without requiring any coding. By simply providing it with a short description, general instructions and specific knowledge sources, BuddyBen was able to retrieve information from grounding data, whilst reasoning, summarising and answering user questions. To ensure an accurate and contextual delivery of information, we carefully defined its scope, grounding BuddyBen’s knowledge base in our Onboarding sharepoint folder. Soon enough, BuddyBen - our brand new agent - was brought to life.

One of the main benefits that stood out when creating this agent was how simple it was to make. Without any coding or programming knowledge, Copilot Studio simplified the entire process in an easy-to-follow and guided flow. This perk helps to bring automation of workflows to a larger range of people, giving anyone the opportunity to personalise and configure their own personal assistant to help them with their work.

The Journey of Agent roll-out

Whilst the ease-of-access of agents extends their reach to anyone in need, this benefit also triggers challenges businesses may face in the future:

The real question therefore becomes: how do we regulate the proper creation and usage of said agents?

1.      Create the right opportunities for agents

With the simplification of agent building achieved by Copilot Studio, any employee, even those without coding knowledge, can create agents nowadays. Yet, without the right guidance, an increase in demand for agents with no appropriate purpose can result in clutter or unproductivity. The real challenge now instead lies in spotting the appropriate opportunities for agents to add value, instead of risking a roll-out of inefficient agents that waste time. The right “appetite for agents” needs to be set within organisations, by uncovering the appropriate scenarios in which agents can fluidify and automate workflows. Which are the most recurrent tasks? Where is it that time is lost? Where can agents assist with this? The right questions need to be asked to define the appropriate adoption scenarios.

2.      Pilot a number of agents

Once identifying appropriate opportunities for roll-out, it’s important to “pilot” a number of agents. Although made simple, agent building requires time and practice, in addition to a lot of trial and error to get it right. Immediate organisation-wide deployment can still be premature, and failing to perfect these virtual assistants can result in inaccurate information. Testing agents out at a smaller scale is therefore recommended, focusing on making them work “perfectly” as they unlock new opportunities.

3.      Agent Adoption

Introduce these agents as “AI services” into the organisation,  ensuring wide usage, and ensuring positive experiences. To achieve this, it is essential to communicate the benefits and potential of these agents clearly to all employees. Organize workshops and training sessions to demonstrate how these agents can simplify tasks and improve productivity. Encourage feedback from users to understand their experiences and make necessary adjustments. By fostering a culture of openness and support, employees will be more likely to embrace and utilize these AI services effectively.

4.      Codify creation process

Create a scalable process that will allow people to create agents themselves based on success stories. Create governance on how to request agents themselves based on success stories. Document the steps and best practices involved in building and deploying agents. Develop templates and guidelines to standardize the creation process, making it easier for employees to follow. Establish a governance framework to manage requests for new agents, ensuring that they align with organizational goals and priorities. This structured approach will help maintain consistency and quality in the agents being developed.

5.      Facilitate and coach

Train people on using and building efficient agents into workflows. Train people on using and building efficient agents into workflows. Provide ongoing support and resources to help employees integrate agents into their daily tasks. Offer coaching sessions to address any challenges or questions they may have. Encourage collaboration and knowledge sharing among teams to leverage collective expertise. By empowering employees with the skills and confidence to use agents effectively, the organization can maximize the benefits of these AI services.

6.      Monitor and assess

Keep track of the working agents and have agile improvement cycles. Regularly review the performance and impact of the deployed agents to identify areas for improvement. Collect data and feedback to measure their effectiveness and make data-driven decisions. Implement agile methodologies to iterate and enhance the agents continuously. This proactive approach ensures that the agents remain relevant and valuable, adapting to the evolving needs of the organization.

As collaborative work continues to evolve, the effective integration of AI agents into the workforce becomes increasingly crucial. Agents can become powerful tools in the future, and the increased efficiency of human collaboration with AI is an opportunity that cannot be left behind. Yet, without the right guidance to accompany adoption, workflows can rapidly become overwhelming and disorganised. Employees will need to understand these changes and adapt to become Agent bosses; building, delegating to, and managing agents to enhance their impact. A new way of working is becoming apparent, and employees are growing accustomed to the fact that artificial intelligence is working with them rather than against them.