Email tone takes too many revisions
Provide the context, reader, purpose, and tone once, then ask AI for an editable first draft.
An email checked for facts and tone
Start with email Do not start with a tool list
Which part of the day do you want back?
Choose a recurring task with clear inputs and outputs. Change one thing in the first week so you can see whether AI genuinely helps.
Every meeting creates another half hour of admin
Turn a transcript into decisions, actions, owners, and due dates—not a wall of notes.
Minutes ready to confirm and assign
Process one meeting Every slide deck begins with a blank page
Define the audience, conclusion, and evidence before asking AI for the narrative and slide outline.
A report structure built around one conclusion
Build a deck outline Cleaning data and formulas drain attention
Ask AI to explain formulas, identify anomalies, and suggest steps while keeping source data and spot checks.
A data result with a review procedure
Work on a spreadsheet One controlled workflow
AI productivity is five checkable stages, not one prompt
Select a stage to see what the person owns and what AI handles. Without context, review, or reuse, the next task still starts from zero.
Measure with your own data
How much time does this workflow actually save each week?
Ignore platform marketing figures. Record the original time, the real AI-assisted time, and the weekly frequency for your task.
- Output quality did not fall
- Sensitive information follows company policy
- Human review remains
- The template works on the next task
minutes
minutes
times
Based on your inputs, you reclaim2.9 hours per week
This is a process metric from your inputs, not a guaranteed outcome. Record two weeks before expanding usage.
Your first hour
Finish one reusable template today—do not read all 23 lessons first
Start from real work and open learning content only when you need it. These four steps already leave you with something worth improving.
Choose one repetitive task due today
An older version makes comparison easier.
Add context, constraints, and output format
Do not stop at “make this better”.
Review the AI draft item by item
Check facts, omissions, tone, and sensitive information.
Save the template and review checklist
Record when it can and cannot be used.
Continue by work outcome
Do not read in order—solve the next real task
The 23 lessons cover communication, documents, data, automation, and team adoption. These are the eight most useful entry points.
Communication
AI email and messaging
Draft, edit, translate, and control tone and format.
Meetings
AI meeting notes
Extract decisions, owners, and actions from transcripts.
Reporting
AI documents and slides
Build outlines and page content from conclusions and evidence.
Data
AI spreadsheets and cleaning
Generate formulas, cleaning steps, and review methods.
Knowledge
AI research and knowledge bases
Compare material, preserve sources, and build SOPs.
Automation
AI workflow collaboration
Connect email, calendars, knowledge, and automation tools.
Safety
Security, compliance, and quality
Handle privacy, fact checks, and human review.
Adoption
Team adoption and measurement
Move from personal templates to pilots and metrics.
When you need to go deeper
Choose a specialist course or credential after the first workflow
Courses and credentials should not block the first practice. Build one reusable workflow, then choose for your work environment.
Claude Code Workshop
For people ready to connect batch files, spreadsheets, email, and slides into deeper automation.
Explore the workshopClaude Associate Foundations · CCAO-F
For validating Claude workplace use, prompting, context, and safety foundations; it does not replace real work evidence.
View the credential pathAI Data Analysis
Move here when the job has shifted from organising spreadsheets to analysing data and communicating insight.
Open data analysisFAQ
Common concerns before starting
Can I learn AI for office work without coding?
Yes. Email, meetings, documents, slides, and basic spreadsheet tasks can start in conversational tools. Add technical skills only when you need batch files, system integrations, or automation.
Can I upload company material directly to ChatGPT or Claude?
Do not assume so. Follow company AI, privacy, and data-classification policies; remove unnecessary personal or commercially sensitive information; and check the account and model data settings. Practise with de-identified samples when uncertain.
How do I prove AI genuinely improved efficiency?
For the same task type, record time before, time after, rework, and errors for at least two weeks. Generation speed alone is insufficient; extra review and rework can make the total process slower.
Do I need subscriptions to many AI tools?
No. Start with one tool already available to you or your company. Compare another only when you encounter a specific context, file-format, integration, or compliance limitation.
Reclaim one hour today
Start with the next email, meeting, or spreadsheet
Choose a real task, record the original time, complete one AI-assisted pass and human review. You do not need to become an AI expert first.