In learning and development (L&D), some projects take on a life of their own, especially when large portions of the organization are impacted and there are high levels of visibility. In those instances, the journey from ideation to presentation can feel daunting.

In 2025, my team, which provides training content to a segment of Avantor’s commercial and technical staff, took on several of these projects, and to help with the process, we turned to artificial intelligence (AI), especially Microsoft Copilot*. Doing so often resulted in a better than 25% reduction in our typical project cycle time from past years.

How We Did It

We used Copilot for a variety of tasks throughout the course of our projects, and our collaboration process generally followed this workflow:

1. Ask Copilot for help with clarity. We started with natural, conversational prompts, but quickly learned the importance of detailed, structured instructions to minimize back-and-forth.

2. Inspect the outputs Copilot offered. Thorough prompts yielded impressive results, but we always closely inspected the outputs – and questioned as needed, asking Copilot for explanations and deciding as a team whether to adopt or adapt its ideas.

3. Modify and iterate with Copilot. We frequently re-uploaded amended documents and requested further review, enhancing our content through iterative collaboration.

We found this process worked well with a wide variety of tasks, as one of our team members shared: “Following our workflow process, I was able to collaborate with Copilot for image generation and visual layout input; to check for consistency of wording and tone in slide deck content; to help break down more complex content into digestible points; to add interactive and multimedia elements to the collateral; and to brainstorm ideas for follow-up exercises.”

Sometimes, it only took a single iteration with Copilot to get the final output we were happy with; other times, it took several. Regardless, we looked at our interactions with Copilot as collaborative and let the process unfold as we went.

Here are some of the most impactful ways that Copilot influenced the development of our projects:

1. Initial Research

One of the most significant early wins with Copilot was felt during our research phases. While our team often knew the general direction we wanted to take with each project, there were often holes in our knowledge that needed plugging.

Copilot was invaluable during our research steps, quickly summarizing vast background content and clarifying complex concepts. That helped us build a solid knowledge foundation on which we could build our training content.

Note: When using AI as a research partner, we regularly cross-checked Copilot’s summaries with other trusted resources. When we had questions, we would thoroughly review the references it used as its sources. This step is critical for any team using AI as a collaborator.

2. Content Review

Throughout the projects, we found Copilot’s content review capabilities to be very helpful. For example, we uploaded each of our project roadmaps and prompted: “Check for flow and consistency — are any concepts missing or unnecessarily repeated?” Copilot quickly flagged redundant content, suggested tighter sequencing and provided a clear explanation for its suggestions that helped us more easily review its output and make final adjustments to our collateral.

When creating each of the slide decks for our projects, we were able to upload those to Copilot accompanied by a review prompt such as: “Review this updated slide deck and let us know if there are additional adjustments you would recommend for clarity and strength of message as well as overall flow.”

Within seconds, Copilot would return a solid summary of the content, highlight the areas that were done well and suggest ways we could tighten the flow and improve consistency. It even suggested tweaks in areas like font sizing, colors and graphics used throughout.

Once we had the first-pass version of our slide decks completed, we continued collaborating with Copilot to create interactive segments: things like group discussion points, exercises we could facilitate for applying concepts and even role-play activities. As we amended our content based on Copilot’s suggestions, we would upload the latest version for further insights, which enhanced the collaborative feel of our interactions.

3. Learning Enhancements

We had several “ah-ha” moments throughout the process, especially regarding ways the learning process could be enhanced with Copilot.

For instance, we often wanted to leverage video content to support the concepts covered. We used the prompt, “Suggest three-minute, safe-for-work, memorable clips for this topic,” and Copilot yielded instant, relevant suggestions with timestamps, saving us hours in background research.

Another moment was when we realized the power of Copilot to help participants with role-play interactivity. Copilot’s ability to simulate customer personas and provide feedback on role-plays gave participants a safe, effective way to practice the conversations they would ultimately have with their actual customers when talking about our products and services. When asked for feedback on the role-play interaction with Copilot, one participant rated it as a five-star experience and said practicing with Copilot helped him feel confident to “schedule a presentation with decision-makers.”

4. Miscellaneous Additional Benefits

Worth mentioning are two additional ways Copilot helped our team.

First, it helped with making content more memorable. Here’s one way we prompted Copilot: “We’re looking for an acronym to describe the content of our presentation. Any thoughts?” Copilot returned three solid options, one of which we incorporated into our material. We even used Copilot to help produce supportive imagery.

Secondly, it helped us create participant collateral. We found the ideas Copilot suggested for collateral were often fantastic, and we amended and adopted many for inclusion in what we provided to each participant. One of the specific ways we frequently collaborated with Copilot was in the creation of high-level summaries of core learning principles that participants could use for easy reference following our training sessions.

Final Thoughts and Key Takeaways

Using Copilot as a collaborator was eye-opening. Its quick, thorough feedback accelerated our training content development and helped ensure our content was polished before sharing with colleagues.

As mentioned above, to help get the best results from an AI tool like Copilot, we followed a standard workflow using the acronym “AIM”:

• A – Ask with clarity: We found through trial and error how important it is to craft detailed, structured prompts to guide AI effectively.
• I – Inspect outputs: To ensure we were providing the best content possible for our colleagues, we reviewed Copilot’s suggestions critically and fact-checked for accuracy.
• M – Modify and iterate: We didn’t limit the number of times we would interact with Copilot on a given topic; rather, we would refine the content based upon its suggestions and then often upload those revisions and keep improving.

Throughout the past year, the process of collaborating with Copilot to develop content helped us recognize what a tremendous help it can be for even complex projects. Each member of our team has come to depend upon Copilot for things we do every day, and we are confident that virtually anyone in the L&D space could benefit in similar ways as well.

*Microsoft and Copilot are trademarks of the Microsoft group of companies.*