Prompts you can test, trust and keep working when the model changes
For professionals and students who use AI assistants for real work: a 12-chapter e-book, a 22-sheet workbook, a practice pack with a Python scorer that needs no API key, 5 fillable forms, 8 Word templates and 212 prompts.
- A Prompt Practice Master Workbook (22 sheets): prompt register with review flags, test set, evaluation scores, side-by-side comparison, regression log, injection tests, bias check and a dashboard with traffic lights
- A practice pack of 25 working files: a scorer that runs on your computer, a 20-case test set and rubric, versioned prompts, a JSON checker, a notebook and the 200-prompt library as Markdown and CSV
- Fillable forms for specifying a prompt, comparing two versions, checking a model change, checking before you paste and running a bias check
- A policy starter, a library standard, 31 messages, 9 scenario cases, a 45-minute briefing and a 14-question LMS-ready knowledge check
The Prompt Engineering Practitioner Kit · Bodhih Pro Kits · 349 by Bodhih Training, trainers to 2,000+ companies since 2008.
How do you write prompts that give reliable results at work?
Write each prompt with six parts (role, task, context, constraints, examples, output format), run it on a fixed set of real cases, score every output with the same rubric, compare versions side by side, and re-test whenever the model or tool changes. The Prompt Engineering Practitioner Kit from Bodhih Training gives you a file for every step.
Want the full free guide first? Read How to write and test AI prompts that work reliably.
You already feel the problem. Here is what it is costing you.
It worked on Tuesday
The same prompt gives a great answer one day and a wrong one the next, and nobody can say whether it got worse or was lucky.
Confident, wrong and sent
An AI draft promises a refund, invents a source or misses the one clause that mattered, and it reads just as smoothly as a right answer.
Prompts in a dozen places
Chat histories, screenshots and a document called FINAL v2. Nobody knows which prompt is current, who owns it or whether it was tested.
The vendor changed something
Replies get longer or stranger after an update, and you find out from a customer, not from a test.
Picture the next few weeks with the kit open beside you.
Imagine the next time someone says the AI assistant 'feels worse'. Instead of arguing, you open the test set, run the scorer and see the pass rate against last month's baseline in under five minutes, no API key needed.
By the end of the week your top three prompts each have an owner, a version, a test set and a rubric. As you approve the first one, the readiness check says GO WITH FIXES and tells you exactly what is left. Colleagues stop sending screenshots of prompts and start using the library.
Whether you start with the Before-you-paste check tonight or the Master Workbook tomorrow, every file is waiting with a filled SAMPLE from Aisha's team, so you can see what good looks like before you type over it.
27 files that do the work for you
The e-book shows you how. The tools make it happen. Here is exactly what you download.
The Prompt Engineering Practitioner Kit (e-book, PDF)
The guide: methods, worked examples, scripts and exercises
Start here: the prompt practice map
The job in eleven steps with the file for each, a 15-minute quick start, your first hour, your first week, the 30/60/90-day plan and the 'brand it in five minutes' box.
Situation router: which files to open, in which order
14 common situations, each with the files to open in order, plus the 15-minute quick start.
Prompt Practice Master Workbook
22-sheet workbook: Settings, 0-4 self-assessment and ranked gap plan, prompt register with review flags, test set, evaluation scores, run summary, side-by-side comparison, regression log, injection tests, bias check, data boundaries, decision log, roadmap, RACI, crosswalk, Tool notes, What changed, glossary, sources and a dashboard with traffic lights and five charts.
Business case calculator (Low, Expected and High cases with break-even)
Inputs labelled 'your estimate', three cases with the Low case first, break-even months, net after 12 months, the cost of doing nothing and a cumulative chart.
Readiness check: go/no-go for a shared prompt
15 criteria (critical and non-critical) with evidence, status and a GO / GO WITH FIXES / NO-GO verdict by formula, filled for the persona's support prompt.
Prompt specification form (the six-part anatomy)
Two-page fillable spec for any prompt your team will reuse: purpose, the six parts, data class and tool, human-review rule, test plan and sign-off.
Side-by-side evaluation scoresheet
Blind scoring of two prompt versions on the same cases, with the rubric reminder, totals, the decision and who reviewed.
Model-change regression check
What to do when a vendor updates a model, you switch tools or edit a prompt: re-run, compare with the baseline, decide, notify.
Before-you-paste data boundary check
A one-page check to run before pasting anything into an AI tool: data class, personal data, approved tool, minimise, output use and disclosure.
Bias and fairness check (name-swap test)
Plan and record a name-swap test for any prompt whose output mentions or affects people: design, minimum runs, results, reading the gap, fix and re-test.
Leadership sign-off memo: prompt practice
One-page decision memo for the executive sponsor: the decision, why now, the 12-week plan, the business case with the Low case first, risks and controls, and a signature block.
Responsible AI use and prompting policy (starter)
A plain-English policy starter: scope, approved tools, data boundaries, prompting and review rules, disclosure, prohibited uses, the shared library, incidents, training and review. For your legal and HR teams to adapt.
Prompt library standard and change procedure
How shared prompts are named, described, versioned, tested, approved, re-tested and retired, with roles and a change procedure.
Message library: 31 emails and messages
31 ready-to-adapt messages: launching the practice, day-to-day prompting, evaluation and regression, safety and injection, leadership, and students' messages to tutors.
Worked example pack: Aisha's filled templates
The persona's completed versions of the three most-used templates: the prompt specification for P-SUP-01 v3.1, the side-by-side scoresheet (v2 vs v3) and the model-change regression check, with a register excerpt.
Scenario casebook: 9 cases with model answers
9 realistic cases from the US, UK, Canada, Australia, Ireland, New Zealand and India, each with facts, questions and a model answer citing the kit's method and files.
45-minute team briefing plan and knowledge-check answer key
A timed 45-minute session plan with materials and activities, plus the answer key for the 14-question knowledge check.
AI prompt pack: 200 prompts by job plus 12 practice prompts
212 tool-agnostic prompts with placeholders: 12 for building and testing prompts, and 200 across ten jobs (support, marketing, sales, HR, finance, operations, data, product, leadership, students), each with what to paste and the check before you use the output, under five human-review rules.
Prompt anatomy and patterns card (with an if-this-then-that decision table)
One-page card: the six parts with examples, six patterns that work across ChatGPT, Claude, Gemini and Copilot, a fix-it decision table and the anatomy as a mini-infographic.
Prompt injection and data safety card (mistakes to avoid)
One-page card of the mistakes to avoid: the data traffic light, ten common mistakes with fixes, direct and indirect injection, and the design rules that limit damage.
Numbers that matter: prompt quality KPI card
The numbers that show whether your prompt practice works: definitions, formulas, where each lives in the workbook and what to do if it moves, with Northpeak's SAMPLE values.
Correct at November 2026: tool notes and rules by market
Dated reference: version-specific notes for ChatGPT, Claude, Gemini and Copilot (projects, data controls, prompting guidance) and the rules that touch prompting at work in the US, EU, UK, Canada, Australia, New Zealand and India, with official sources in the workbook.
Knowledge check: 14 scenario questions (LMS-ready CSV)
Scenario-based multiple-choice questions with the correct option and an explanation; imports into Bodhih LMS or any LMS that accepts CSV.
Calendar: 30-day practice plan and 12-week rollout
30 daily 20-minute practice sessions plus the 12 weekly rollout milestones; imports into Google Calendar, Outlook and Apple Calendar.
Task board import (36 tasks)
The 12-week rollout as tasks with week, due date and status, ready to import into Trello, Asana, Notion, Microsoft Planner or Jira.
Prompt practice files (scorer, test set, outputs, 200-prompt library, templates, notebook)
25 working files: a Python scorer that runs with no API key, the 20-case test set and rubric, SAMPLE outputs for five runs, four versioned prompts, the 200-prompt library as Markdown and CSV, templates, an injection test pack, a JSON checker, a Jupyter notebook and tests.
Chapter by chapter
Each chapter opens with a real situation, gives you a method and ends with something to do today, using the tools in your kit.
- Format
- PDF e-book + XLSX + fillable PDF + DOCX + ICS + CSV + ZIP of practice files
- Files
- 27 files plus a one-click ZIP
- E-book
- 66 pages
- Works with
- Excel, Google Sheets, LibreOffice, Word, Google Docs, Acrobat, Python 3.8 or later, Jupyter, Notion and any LMS that imports CSV
- Best for
- Professionals, teams and students using ChatGPT, Claude, Gemini or Copilot for real work
How language models work, in plain words
Tokens, context windows, probability and why models make confident errors, with the evidence on benefits and its limits.
What you may paste, and where
Data classes, work versus personal accounts, what regulators in five markets advise, and a thirty-second check.
The anatomy of a strong prompt
Role, task, context, constraints, examples and output format, and how they map onto each vendor's guide.
Examples and structured outputs
Few-shot examples that teach format, structured outputs you can check in code, and saved instructions.
Chains of steps and decomposition
Breaking big tasks into checked steps, with a three-step chain for release notes.
Long documents and your own files
Why long inputs hide the middle, how retrieval works, and four grounding rules with quotes and NOT FOUND.
Test sets, rubrics and scoring
Building test sets, writing rubrics with critical criteria, and grading with code, people or a model.
Regression testing when the model changes
Baselines, a regression threshold and a five-step routine for the day the model changes.
Prompt libraries with versions and owners
A register, statuses, versions, change notes and the 200-prompt library as starting points.
Prompt injection for business users
Direct and indirect injection, why a better prompt is not enough, and design rules that limit damage.
Bias and fairness checks
The name-swap test with minimum runs and thresholds, and a worked fix.
Make it a team practice
RACI, roadmap, briefing, knowledge check, the business case with the Low case first, readiness and students.
Not just reading. Tools you will actually open again.
Start here
- Start-here map: the job in eleven steps, a 15-minute quick start, your first hour, first week and 30/60/90 days
- Situation router: 14 situations, each with the files to open in order
Workbooks
- Prompt Practice Master Workbook (22 sheets): self-assessment and gap plan, prompt register with review flags, test set, evaluation scores, run summary, side-by-side comparison, regression log, injection tests, bias check, data boundaries, decision log, roadmap, RACI, crosswalk, Tool notes, What changed, glossary, sources and a dashboard with five charts
- Business case calculator: Low, Expected and High cases with break-even and the cost of doing nothing
- Readiness check: 15 criteria with a GO, GO WITH FIXES or NO-GO verdict
Practice files (ZIP)
- 25 working files: a Python scorer that needs no API key, a 20-case test set and rubric, SAMPLE outputs for five runs, four versioned prompts, the 200-prompt library as Markdown and CSV, templates, an injection test pack, a JSON checker, a Jupyter notebook and tests
Fillable PDF forms
- Prompt specification form
- Side-by-side evaluation scoresheet
- Model-change regression check
- Before-you-paste data boundary check
- Bias and fairness check
Word templates
- Leadership sign-off memo
- Responsible AI use and prompting policy (starter)
- Prompt library standard and change procedure
- Message library (31 emails and messages)
- Worked example pack: the persona's filled specification, side-by-side scoresheet and regression check
- Scenario casebook (9 cases with model answers)
- 45-minute team briefing plan and answer key
- AI prompt pack: 200 prompts by job plus 12 practice prompts, with human-review rules
Cheat sheets
- Prompt anatomy and patterns card (with an if-this-then-that table)
- Prompt injection and data safety card (mistakes to avoid)
- Numbers that matter: prompt quality KPI card
- Correct at November 2026: tool notes and rules by market
Calendar, imports and training
- Knowledge check: 14 scenario questions as an LMS-ready CSV
- Calendar: 30-day practice plan and 12-week rollout (ICS)
- Task board import (36 tasks) for Trello, Asana, Notion, Planner or Jira
When you have worked through it, you will be able to…
Every outcome is practised with a worked example and a tool, not just described.
- You will be able to explain in plain words why models make confident errors and what that means for checking
- You will be able to set data boundaries and run a thirty-second check before pasting anything
- You will be able to write prompts with the six-part anatomy that work across ChatGPT, Claude, Gemini and Copilot
- You will be able to build a test set and rubric and score outputs with code and a person's spot-check
- You will be able to catch a regression after a model change by comparing with a saved baseline
- You will be able to run a versioned prompt library with owners and review dates
- You will be able to test prompts for injection and for unfair treatment of people by name
- You will be able to brief a team, run a knowledge check and present a business case with the Low case first
Know your starting point. Prove your progress.
People who measure before and after improve faster and can show it. AssessAll, our sister assessment platform, has AI-graded assessments that pair with this kit. Take one now, work through the kit, then re-take it.
Task Specification and Prompt Writing Work Sample for Knowledge Workers Using Generative Tools
Take the assessment →AssessAll assessmentAI Draft Supervision: Reviewing What the Machine Wrote
Take the assessment →AssessAll assessmentEveryday AI Judgment at Work
Take the assessment →JobularyTurn what you learn into a measured development plan (Jobulary)
Make it a growth plan →A taste of Chapter 3
Northpeak's support reply prompt, P-SUP-01, went through three versions in four weeks, each run on the same 20 test cases and scored the same way. Version 1 was the one line "Reply to this customer." It passed 6 of 20 cases (30%), with 9 critical failures: invented promises, no escalation on cancellations, no source cited. Version 2 used the six parts in a simple form. It passed 13 of 20 (65%). Version 3 added tags around the ticket and article, an explicit list of topics that must go to a specialist, a rule against repeating personal details, one example and a 120-word limit. It passed 18 of 20 (90%), with no critical failures. The practice files hold every version as a text file with its scores in the header.
The honest comparison
| Free templates and typical advice | This kit | |
|---|---|---|
| Method | Lists of magic prompts that change with each model | A practice: six-part anatomy, test sets, rubrics, baselines and a versioned library |
| Testing | Try it once and see if it looks right | A 20-case test set, a five-criterion rubric and a Python scorer that needs no API key |
| Model changes | Find out from a complaint | A regression check against a saved baseline with a written threshold |
| Data safety | A policy nobody reads | Data classes by tool, a thirty-second check and a policy starter for your advisers to adapt |
| Security and fairness | Not covered | Injection tests and design rules; a name-swap bias check with minimum runs |
| Compared with hiring a consultant or buying software | Can help, but you still need your own test cases, owners and routine | Gives you the routine, templates and files to run it yourselves, and works alongside any tool |
Made for you if…
- Professionals who use ChatGPT, Claude, Gemini or Copilot for real work and want results they can rely on
- The person who has become 'the AI person' in a team and needs a method, not tricks
- Managers setting rules for AI use and a shared prompt library
- Operations, support, marketing, HR, finance and analyst teams with prompts other people rely on
- University students who want to use AI within their institution's rules
- Teams in the US, Canada, the UK, Ireland and the EU, Australia, New Zealand, India and other English-speaking markets
Probably not for you if…
- Developers building AI applications or training models: this is a practice for people using assistants, though the evaluation method still applies
- Anyone looking for legal advice or a certification: the dated sheet points to official sources and your adviser
- Anyone hoping for magic phrases: the kit teaches testing and maintenance, and says plainly that no prompt is guaranteed
- Students looking for help to have AI write graded work: the kit puts your institution's integrity rules first
Built by trainers, tested in real rooms.
Bodhih Training has designed and delivered corporate programmes since 2008. Bodhih Pro Kits package what works in those workshops into a guide plus the exact tools we use, so you can apply it the same day without a facilitator in the room.
From here to using it in three steps
- STEP 1
Tap Get the kit
Name, email and country. UPI, card or netbanking in India; international cards everywhere else. No account to create.
- STEP 2
Download instantly
Your download page opens at once, with every file and a one-click ZIP. A permanent link lands in your inbox too.
- STEP 3
Open the first tool
Start with the quick-start chapter, then the main spreadsheet or planner. Most buyers use something from the kit the same day.
Make your next prompt one you can test
Open the Prompt specification form on your most-used prompt and see which of the six parts are missing. Whether you start there tonight or with the scorer tomorrow, the files are ready, each with a filled SAMPLE to show what good looks like.
Personal licence for the buyer's own use (and use with your own team or students where the kit says so). Please don't share or resell the files. Team and company licences: solutions@bodhih.com.
What buyers say
No ratings yet. Only people who have bought this kit can rate it - every buyer gets a one-click rating link a couple of days after their purchase.
The Bodhih family, when you are ready for more
The Build Your Own AI Assistants Kit
Design, build, test and share custom AI assistants for repeatable work with skills, plugin.
The Claude Code and AI Coding Assistants Playbook
Ship better software faster with AI coding assistants as a professional developer or team .
The Microsoft 365 Copilot at Work Kit
Use Microsoft 365 Copilot well every working day in Word, Excel, PowerPoint, Outlook and T.
AssessAll
250+ AI-graded assessments to measure skills and earn verified credentials.
Jobulary
Individual development plans that start from evidence and stay yours.
Bodhih Training
Run this topic as a live workshop for your team, in person or online.
Pewple
Book a vetted coach for career, leadership or communication. Pay per session.
Bodhih.org pathways
Certification pathways in Applied AI, Workplace English and management.
The Prompt Engineering Practitioner Kit: FAQs
Read the free guideAlready bought? Resend my linkWhat is prompt engineering, in plain words?
It is the practice of writing instructions for an AI assistant so that it gives useful, correct and consistent results, and then checking that it does. This kit treats it as a routine: write with six parts, test on real cases, compare versions and re-test when anything changes.
Do I need to code to use this kit?
No. The workbook does everything the Python scorer does, from scores you type in. The scorer is there for people who want speed: it runs on any computer with Python 3.8 or later, with no API key, and the book shows each command.
Does it work with ChatGPT, Claude, Gemini and Copilot?
Yes. The method and prompts are tool-agnostic. Version-specific facts such as projects, skills, plugins, file limits and data settings are on the Tool notes sheet with official links and a checked date, because they change often.
How do I stop AI making things up?
You cannot stop it completely, but you can make errors rarer and visible: give the model its sources, tell it to use only them and to say NOT FOUND, ask for quotes, set constraints, and score outputs against a test set with critical criteria.
How long does it take to set up?
The 15-minute quick start improves one prompt today. A full team practice follows the 12-week roadmap in the workbook and calendar, at about a few hours a week for the practice owner.
What should we never paste into an AI tool?
Check your own rules first. As a starting point the kit treats health data, biometrics, payment card numbers, passwords and security keys as never, and personal and confidential data as approved work tools only, with the details minimised. Regulators in several markets advise against personal information in public tools.
Is prompt injection a risk for ordinary business users?
Yes, when an assistant reads emails, web pages or files written by someone else. The kit explains direct and indirect injection, gives eight test inputs, and sets design rules such as requiring a person to approve any action.
Can I use this with my team or students?
Yes, within your own organisation or class under the personal licence. Company, multi-site and LMS licences are available from solutions@bodhih.com.
Does it work in my country?
The method is universal. The dated sheet and the workbook's Crosswalk cover the US, the EU, the UK, Canada, Australia, New Zealand and India with official sources; laws change, so check locally with your adviser.
Is this legal advice or a certification?
No. It is educational and helps you organise a practice. The policy starter and memos are drafts for your advisers to adapt, and nothing in the kit guarantees compliance or results.
How can I measure my starting point?
Score the self-assessment in the workbook first. For an independent measure, AssessAll's Task Specification and Prompt Writing Work Sample, AI Draft Supervision and Everyday AI Judgment at Work assessments test the skills this kit builds.
How do I turn this into a development plan?
Use the 30-day practice plan in the calendar, then record goals and review dates in a development plan; Jobulary's individual development plan works well for this.
How much does The Prompt Engineering Practitioner Kit cost?
₹1,299 including 18% GST for buyers in India, or $34 for buyers outside India. One-time payment, no subscription, with an automatic invoice.
How do I get the files after paying?
Instantly. Your download page opens the moment payment succeeds, and a permanent download link is emailed to you. Download each file separately or everything in one ZIP.
What software do I need?
Any of the usual tools: Excel or Google Sheets for spreadsheets, Word or Google Docs for templates, Adobe Acrobat Reader, Preview, Xodo or GoodNotes for the fillable PDFs, and Google, Outlook or Apple Calendar for the calendar plan.
Can I get a refund?
Because it is a digital download, kits are not refundable once downloaded. We always refund a duplicate charge or a failed delivery, and if a file is corrupted and we cannot fix it within 3 working days. Full terms: bodhih.org/refunds.
Can my company buy this for a team?
Yes. Team and company licences, and workshops on this topic run by Bodhih Training, are available. Write to solutions@bodhih.com or call +91 99000 11601.





