Last year I had the opportunity to work in the executive board of a non-profit organisation as the Secretary and Head of IT. Due to the size of the organisation and governance needs, I led a full migration from Google Workspace to Microsoft 365. I have always preferred Microsoft’s ecosystem so was very excited to spearhead this migration and use all the new tools! This is how I discovered Copilot.
What is Copilot?
Copilot is Microsoft’s family of AI assistants embedded across tools you already use:
- Microsoft 365 Copilot lives in Outlook, Word, PowerPoint, Excel and Teams
- GitHub Copilot is an AI programmer in your IDE (Integrated Development Environment) that suggest code, explains snippets and any other programmer related functions
- Copilot in Bing does text-to-image and creates drats for marketing assets and visuals
In practice I used Microsoft 365 Copilot and GitHub Copilot daily. One to manage communications and documents, the other to accelerate small scripts and automations.
How Copilot changed my life
As the secretary my inbox was the nerve centre of the organisation and Copilot became my triage layer. I would open a 20-email chain, hit summarise, requesting the output of the dates, owners and deadlines in the email, ensuring to make the list bullet format as well. The result doubled as a handover note on my planner and curate a reply if necessary. The drafts came back 80-90% ready but I had to learn to use the correct prompts as I quickly realised that Copilot rewards constraint rich prompts.
Data in the organisation was moved between forms, sheets and reports frequently and Copilot helped me transform messy reports into clean models. This is what I saw Copilot thrive in because it was great at translating what I meant into table structures and formulas in excel, and in turn explaining them in plain English so teams can maintain them. A simple data clean, such as creating a clean table in Excel, and Copilot would offer Power Query steps and dynamic array formulas so anyone on the team could refresh without me.
How GitHub changed the game
My favourite Copilot tool was GitHub. Due to the nature of this organisation, everyone who applied had a chance to be interviewed. We had ~200 applicants, 9 interviewers, and 3 parallel interviews per time slot. Each applicant needed exactly one slot; each interviewer could conduct at most one interview per slot. Applicants listed preferred time windows and sometimes preferred interviewers. We wanted fairness and minimal manual edits.
With GitHub Copilot in VS Code, I built a pragmatic greedy + repair algorithm:
- Model all tracks as (slot, room) where room ∈ {0,1,2}.
- Score candidate matches (preference fit, interviewer load balance).
- Assign most-constrained applicants first (fewest viable slots).
- If a conflict emerges (double-booked interviewer), repair via simple swaps.
- Export to CSV for Excel + mail-merge.
Copilot proposed helper functions as well as docstrings and generated the CSV report with Power Query instructions for Excel. This allowed me to create a complete schedule quicker, with fewer errors and keeping the script allowed it to be reusable.
Conclusion
My takeaway is simple: if I can specify it, Copilot accelerates it and even when I can’t, writing the prompt clarifies my thinking. Outlook and Excel became force multipliers, and GitHub Copilot turned a 200-interview puzzle into a manageable workflow. Crucially, Copilot is embedded in the organization’s Microsoft 365 environment, so it inherits the access controls and compliance policies. That means sensitive emails, files, and calendars stay within the tenant while Copilot works on top of them no context switching to third-party sites or risking data leakage. Human review still matters, but the baseline is faster, clearer work. Drafts arrived faster, numbers were cleaner, and the scheduler went from “painful” to “repeatable.”
