Tech & AIBeginnerPreview
Using ChatGPT & AI Tools for Work
A practical, no-jargon guide to using ChatGPT, Claude, and Gemini to draft content, answer questions, and speed up the daily work of any role. You finish with a reusable prompt library and a clear sense of where AI helps and where it cannot.
For working professionals and beginners who want to use AI tools confidently in their daily job without a technical background.
Course content
Workbook & downloads
Put the course into practice — a printable workbook plus editable templates you can fill in and reuse.
Preview the workbook
This workbook turns the course into a working system you apply to your own job this week, not notes you forget. Keep one real recurring task in mind, an email you always rewrite, a report you always summarise, a meeting you always run, so every exercise produces something usable: a chosen tool, a reusable voice prompt, a growing prompt library, and safe habits you can defend to a manager or client. Work through one section per module and fill the templates with your real work.
Getting Started With AI Assistants
Pick the right tool for your work and get comfortable holding a real back-and-forth conversation.
Exercise: Run the Same Task Through All Three
Take one real, small task from today and run the identical prompt through ChatGPT, Claude, and Gemini. Compare the three results for accuracy, tone, and how much editing each needs. Note which felt best for this kind of task.
- Which tool gave the result closest to ready-to-use, and what specifically made it better?
- Did any tool invent a fact or sound noticeably more generic than the others?
- Based on where your work actually lives (Google, Microsoft, or neither), which tool will be your default?
Worksheet: Your Tool Decision Sheet
Fill this in once to decide which assistant you reach for by default and when you switch. Keep it near your desk for the first two weeks.
- Where most of my work lives (Google Workspace / Microsoft 365 / other)
- My default tool and the one-line reason
- My second tool, used for second opinions
- Tasks where I will prefer Claude (long documents, careful editing)
- Tasks where I will prefer Gemini (inside Gmail, Docs, Sheets)
- Tasks where I will prefer ChatGPT (general, custom assistants, images)
- Free tier or paid plan, and the date I will reassess
Checklist: First-Conversation Readiness
- I have created a free account on at least two of the three tools
- I have held a back-and-forth of at least three messages on one real task
- I have uploaded a document and asked a question about it
- I have turned memory on or off deliberately, knowing what it does
- I have started a fresh chat when switching topics instead of continuing an unrelated thread
Prompting That Gets Usable Results
Master the role-context-task-format pattern and capture your own voice so output needs only a light polish.
Exercise: Rebuild a Vague Prompt With Structure
Take a vague request you might normally type, such as write me a status update, and rebuild it using all four parts: role, context, task, and format. Run both versions and compare how much editing each result needs.
- What facts did the context line need to supply that the model could not have guessed?
- How much shorter was your edit on the structured version compared to the vague one?
- What format constraints (length, bullets, tone) made the biggest difference to usefulness?
Worksheet: Your Reusable Voice Prompt
Build the single most valuable prompt you own: a paragraph that describes how you sound, so you can paste it before any writing task. Fill each field, then assemble them into one block.
- Three real sentences I have written that sound like me
- My typical sentence length and rhythm (short and punchy / longer and detailed)
- Words and phrases I never use (e.g. leverage, delve, seamless, going forward)
- My default tone (warm, direct, formal, plain)
- How I usually open and close a message
- Any formatting I prefer (bullets, no jargon, one concrete example)
Worksheet: Worked Email Rebuild
Recreate the course walkthrough with your own material. Start from rough notes of a real recent conversation and shape them into a finished message.
- My rough bullet notes from the conversation
- The role I want the model to write as
- The single task and the one clear call to action
- Facts I must check in the draft (names, numbers, dates)
- The final word count and tone I want
Checklist: Prompt Quality Check
- My prompt states a role for the model to take
- My prompt supplies the context facts the model cannot know
- My prompt names one clear task, not several at once
- My prompt specifies length, structure, and tone
- I pasted my voice prompt before any writing task
- I read the result aloud and cut any sentence that sounded like a template
Summarising, Researching, and Meetings
Turn long documents into decisions, get research you can cite, and make every meeting end with clear owners.
Exercise: Summarise for a Decision, Not a Shrink
Take a real report, contract, or long email thread and ask for a decision-focused summary, then an action-focused one. Require the model to quote the sentence each point came from so you can trace it.
- What decisions did the document actually ask the reader to make?
- Did any extracted action item have no clear owner, and who should it be?
- When you checked the quoted sentences, did the summary invent or distort anything?
Worksheet: Sourced Research Log
Use a web-connected tool (ChatGPT with search, Gemini, Copilot, or Perplexity) to answer one real work question, and record your verification so the answer is defensible.
- My research question, stated with the form I want (e.g. three figures with sources and dates)
- Tool used and whether it returned clickable links
- Claim 1 and its source URL
- Claim 2 and its source URL
- Which sources I actually opened and confirmed
- Any claim I could not verify and chose not to use
Checklist: Meeting Capture Discipline
- Before the meeting I asked AI for likely questions and objections
- I told participants if a notetaker was recording, and got their agreement
- I asked for action items as owner, task, and due date
- I asked the model to list anything discussed but left undecided
- I reviewed the AI summary for accuracy before sending it
- I sent the recap to attendees the same day
Everyday Workflows and Safe Habits
Automate the repeatable work, protect confidential data, and lock in a daily AI habit that compounds.
Exercise: Find Your First Automation Candidate
List the tasks you do at least weekly that follow a predictable shape. Pick one where a wrong output is annoying rather than dangerous, and sketch how a no-code tool (Zapier or Make) or a custom assistant could handle the AI step with you reviewing before anything sends.
- Which weekly task is the most repetitive and the lowest risk to automate first?
- Where exactly does a human need to review before the output goes anywhere?
- How much time does the task take now, and what will you compare against after?
Worksheet: Data Safety Decision Sheet
Before pasting anything sensitive, run it through this sheet. Complete it once for the kinds of data you handle so the rules are decided in advance, not in the moment.
- Information that is always safe for me to paste
- Information I must anonymise before pasting
- Information I must never paste unless on an approved tool
- Whether my account has training-on-my-data turned off
- The approved or business-tier tool I use for internal work
- My employer or client policy on AI use and disclosure
Worksheet: Two-Week Habit Tracker Plan
Set the concrete target that turns AI from occasional to automatic, and plan how you will capture what works.
- My daily target (e.g. three real tasks per working day for two weeks)
- Where my prompt library lives and how I open it
- The day and time of my weekly five-minute review
- One new task type I will try with AI next week
- How I will record wrong or weak answers to stay calibrated
Checklist: Safe and Sustainable Use
- I never automate anything that sends to a customer without human review
- I would be comfortable if my manager, client, and the provider saw what I pasted
- I anonymise inputs when unsure rather than pasting raw sensitive data
- I am honest about significant AI use when someone relies on my judgement
- I add a new prompt to my library whenever one works well
- I run a weekly review to widen my AI use and check its limits
Your Action Plan
- Create free accounts on two of ChatGPT, Claude, and Gemini, and confirm which one sits closest to where your work already happens.
- Complete your Tool Decision Sheet so you have a default tool and a second-opinion tool chosen in advance.
- Build and save your reusable voice prompt, then paste it before every writing task for the next two weeks.
- Rebuild three vague prompts you use often into the role-context-task-format pattern and save the strong versions.
- Run one real document through a decision-focused and an action-focused summary, requiring quoted sources for each point.
- Answer one real work question with a web-connected tool and verify every claim against an opened source.
- Set up one AI-assisted meeting recap with owner, task, and due date, and send it the same day with participants informed.
- Complete your Data Safety Decision Sheet and confirm training-on-your-data is turned off or you are on an approved tool.
- Build one small no-code automation or custom assistant for a weekly low-risk task, keeping a human review step.
- Start your prompt library and run a weekly five-minute review for two weeks until daily AI use becomes automatic.
Pairs well with
Courses members commonly take alongside this one.
Flagship CoursePreview
Freelance Business Foundations: Position, Price, Sell, and Deliver High-Value Services
Freelancing · Beginner · 16h
Self-pacedPreview
Client GrowthPreview
Freelance Client Acquisition: Outreach, Leads, Referrals, and Deal Flow
Freelancing · Beginner · 15h 30m
Self-pacedPreview
Sales SystemPreview
Freelance Sales & Proposals: Discovery Calls, Scoping, Objections, and Closing
Freelancing · Intermediate · 16h
Self-pacedPreview