AI for Productivity at Work: Email, Meetings, Reports and Excel

The fastest way to feel the value of AI is to apply it to the tasks you already do every week: email, meetings, documents, spreadsheets and planning. In this lesson you'll see concrete cases with copy-and-paste prompts, plus a method for deciding what to delegate and what to keep.

What to delegate: the decision matrix

Not everything is worth doing with AI. Ask two questions about each task: how bad is a mistake, and how easy is it to check?

Email

For most professionals, this is the use case with the biggest payoff.

• Drafting from rough notes: "Turn these notes into a professional email to my manager: supplier meeting moved to Thursday, they want a 5% increase, I suggest negotiating payment terms instead of price."
• Answering a difficult email: paste the email you received and ask, "Write a firm but polite reply turning down the extra discount and offering free shipping instead."
• Adjusting tone: "Make this email shorter and less defensive."
• Summarizing long threads: "Summarize this 15-email thread: what was decided, what's still open and who owns each item."

Here's a complete prompt for the trickiest case, the difficult email:

You're the sales manager at a restaurant supply distributor in
Dallas. A long-time customer is asking for an extra 15% discount
on their next order. We can't do it: our margin is already thin.
We can offer free shipping on orders of $2,500 or more, or net-30
payment terms. Write the reply: firm but warm, acknowledge their
loyalty, explain without long excuses, offer both alternatives and
close with a question. 130 words max.

<customer_email>
[paste the email here]
</customer_email>

Expected result: an email that doesn't read like a flat "no," with both alternatives spelled out and a closing question along the lines of "Which of these options works better for you?" Always check that it didn't promise anything you didn't authorize.

Meetings

1. Before · 5 minAn agenda with goals"Draft a 30-minute agenda to review last quarter's results with the sales team. Include goals, time blocks and the questions worth asking."
2. DuringTranscription and summaryMany video-call platforms and note-taking apps transcribe and summarize automatically. Always tell participants, follow your company's recording policy and remember that some states, such as California, require everyone's consent to record a conversation.
3. After · 10 minMinutes and follow-upPaste the transcript and ask: "Write minutes with decisions made, action items (owner and due date) and open issues. Then draft the follow-up email to attendees."

Documents and reports

• Structure: "Propose an outline for a report on last year's employee turnover to present to leadership."
• First draft built from your data and notes, never from nothing.
• Review: "Review this report the way a skeptical CFO would: what questions would they ask me, and which arguments are weak?"
• Repurposing: "Turn this 8-page report into a one-page executive summary" or "into 5 slides with a title and 3 bullets each."

Excel and spreadsheets

AI is excellent at writing and explaining formulas, even when you don't know the function's name:

I have a sheet with column A (date), B (sales rep), C (region)
and D (amount). I need a formula that adds up the amounts for the
rep in cell G2, only for the "North" region and only for the month
of the date in H2. I use Excel for Microsoft 365.

Expected result: a formula like this one, with an explanation of each part:

=SUMIFS(D:D, B:B, G2, C:C, "North",
   A:A, ">="&DATE(YEAR(H2), MONTH(H2), 1), A:A, "<="&EOMONTH(H2, 0))

It adds column D when the rep matches G2, the region is "North" and the date falls between the first and last day of the month in H2. If the AI gives you a different version, that's not necessarily wrong: there are several correct ways to solve it. What counts is testing it on cases you know.

It can also explain formulas you inherited ("walk me through what this formula does, step by step"), help you find errors, suggest how to structure your data or write macros. Always test the formula on a few rows you can calculate by hand. If you want to understand the functions it suggests, the Excel and Google Sheets for data analysis lesson of our data analytics course covers them.

Organizing and planning

• "Here are my 14 tasks for this week [list]. Sort them with the urgent/important matrix and propose a day-by-day plan, keeping in mind that Tuesday and Thursday afternoons are booked."
• "Break this project into phases, tasks and a 6-week timeline." (For a solid method behind the plan, see our lesson on project scheduling and Gantt charts.)
• "Help me prepare a conversation with a team member who keeps missing deadlines: the goal, how to open and how they might respond." (Pair it with our lesson on difficult conversations.)

Learning faster

One of the most underrated uses: a tutor available at any hour.

• "Explain EBITDA with an example from a bakery."
• "Quiz me with 10 multiple-choice questions on [topic] to prep for my exam, one at a time, and correct me."
• "I have an interview for a data analyst role. Run a mock interview: you're the interviewer, ask one question at a time and evaluate my answers at the end."

How much time you can save

Run this simple calculation on your own tasks:

These figures are illustrative: your savings will depend on your tasks and how much you practice. The point is to measure. The Time saved tab of the workbook already has these four rows: 215 minutes saved a week (59.7%), which adds up to about 186 hours over a year. Replace them with your own tasks for one week.

Common mistakes that waste time

Meeting minutes. From the transcript I'll paste, produce:
1) Context in one sentence. 2) Decisions made. 3) A table of action
items with owner and due date (if no date was mentioned, write
"TBD"). 4) Open issues. Don't add anything that isn't in the
transcript. Then write a short email to send the minutes to attendees.
<transcript>...</transcript>

Summary

• Delegate what's easy to check or low-impact; always review what matters.
• Email, meetings, documents, Excel and planning are the highest-payoff uses.
• Teach AI your style and save your reusable prompts.
• Get consent before recording meetings; some states require everyone's permission.
• Measure the time you save, review included, to know where AI adds real value.

Next lesson: Advanced Prompt Engineering: Prompt Chains, Structured Outputs and Fewer Hallucinations. The Intermediate level starts with chaining prompts, asking for tables and JSON, grounding answers in documents, critical roles, metaprompting and testing.