Empowering Teams with Prompt Engineering for Copilot

Client Context & Challenge

A mid-sized organisation in the public services sector (10 users) had recently rolled out Microsoft 365 Copilot to support productivity, document automation, and content creation. While the platform was enabled, the team was struggling to get consistent, high-quality results from Copilot—responses were often vague, irrelevant, or required extensive editing. The leadership recognised that while software adoption was one thing, mastery of how to prompt the AI was equally important.

Key challenges included:

• Users lacked understanding of prompt design principles (how to provide context, goals, and constraints)

• Different people in the organisation used Copilot inconsistently; outcomes varied a lot

• Time was being wasted in trial-and-error prompting

• There was no internal competency or best-practice framework for “prompt engineering”

They turned to Microtek Learning to run a focused capability-building intervention: a one-day workshop on MS-4005: Craft Effective Prompts for Microsoft 365 Copilot (Microsoft’s official course).

Goals & Success Criteria

Before the engagement, Microtek Learning and the client together defined the following objectives and measures of success:

Training Design & Delivery

1. Pre-training Assessment & Customisation

• Microtek Learning assessed the participants’ baseline: which Microsoft 365 apps they often used (Word, Excel, PowerPoint, Outlook, Teams)

• They also gathered sample prompts the team had tried (and failed) to understand common errors or patterns

• Based on this, Microtek Learning slightly tailored examples and labs to relevant workflows (e.g. summarising meeting notes, drafting memos, transforming spreadsheet data, generating presentation outlines)

2. Workshop Structure (1 Day)

The training followed the structure of MS-4005, covering modules on summarization, drafting/creation, editing/transformation, and asking/analysis prompts.

• Introduction & Principles — What makes a “good prompt”: clarity of goal, context, constraints, source, and expectation

• Hands-on Labs — Each participant worked in a Copilot-enabled environment; exercises included:

  • Summarizing long documents

  • Drafting email responses

  • Rewriting or transforming content

  • Asking Copilot to analyze tables / compare data

• Live Coaching & Iteration — Trainer observed prompt trials, gave immediate feedback, asked “How would you tweak this prompt?”

• Best Practices & Templates — Shared prompt patterns, do’s and don’ts

• Wrap-up & Next Steps — Co-creating a mini “Prompt Cookbook” with sample prompts relevant to their business domains

3. Post-Workshop Reinforcement

• Microtek Learning provided participants with lab access (or access to sample prompt libraries)

• They offered a follow-up “office hours” slot over the next week to help refine real prompts

• They encouraged users to bring their real tasks into the Copilot environment, try the prompts, and share experiences

Outcomes & Impact

Within a few weeks, the organisation began to see tangible improvements:

1. Higher-quality outputs
   Many users reported that Copilot drafts now required far less manual editing. Some prompts gave near–final versions straight off.

2. Time savings
   On average, users estimated they were saving around 20-30% of editing or iteration time after adopting improved prompt patterns.

3. Greater consistency & confidence
   The prompt guidelines (“Prompt Cookbook”) helped the team converge on a shared style. In post-training feedback, the average confidence score in prompt design was 4.7/5.

4. Better adoption across roles
   Even team members who were hesitant initially began to experiment with Copilot for drafting emails, memos, meeting summaries, or turning spreadsheet data into narrative summaries.

5. Cultural shift toward experimentation
   The training planted a mindset of “prompt → review → refine” rather than ad hoc trial and error. People became more willing to iterate prompts, share better prompts with peers, and treat prompt design as a craft.

Lessons Learned & Best Practices

• Real tasks are the best prompts
  Allow participants to bring in their own work (reports, emails, spreadsheets). Training feels more relevant and stickier.

• Iterative feedback loop matters
  It’s not enough to teach prompt theory; live coaching and prompt iteration accelerate learning.

• Shared prompt artifacts help scale learning
  The Prompt Cookbook or template library becomes a living reference that ensures consistency.

• Ongoing reinforcement is key
  Post-course support (office hours, peer reviews) helps cement behaviors.

• Expect variation across roles & apps
  Some prompts work in Word but not Excel; context matters. Encourage people to tailor patterns.

• Start simple, then deepen
  Begin with summarization or drafting prompts; then move to transformation, comparative, and analytic prompts.

Conclusion & Future Plans

By working with Microtek Learning on the MS-4005 workshop, the organisation accelerated its Copilot adoption journey. Rather than leaving users to “figure it out themselves,” this structured, hands-on capacity building created real improvements: more usable outputs, less wasted time, greater confidence, and a shared prompt design culture.

To continue momentum, the client plans to:

• Expand prompt training to more employees

• Periodically run “prompt clinics” or peer-review sessions

• Curate a company-wide prompt repository aligned with domain use cases

• Consider advanced training (e.g. prompt + plugin/connector extension) once baseline mastery is achieved