Tel. 02-570-8449 Mobile 088-807-9770 Payment
Mon–Fri 9:00–17:00 @itgenius
IT Genius Institute
IT Genius Institute
CoursesWorkshopSchedule
For business
In-house trainingAI Transformation (TH)Government training (TH)Development servicesTraining room rental
Articles
About us
About IT GeniusInstructorsPortfolioOur clientsPayment methods
Contact usOnline
Theme
Language
THENLA
Ask about a coursePick whichever channel suits you
LINE @itgeniusFastest reply · Mon–Fri 9:00–17:00
Call us
02-570-8449Office088-807-9770Mobile
Email & social
contact@itgenius.co.thFor quotations and documentsFacebookitgeniusonlineYouTubeitgeniusengPlan training for your team
Home / Courses / AI / Detail
AI · AIC-60

Prompt Engineering for Business

Prompt Engineering for Business is a course in writing structured prompts with Role, Context, Task, Format, Constraints and CO-STAR, plus few-shot, iterative prompting and prompt chaining, for use with Copilot, ChatGPT, Gemini and Claude. It suits office staff, engineers, supervisors and AI champions, and you leave with a prompt library for your recurring tasks.

Updated
Prompt Engineering for Business
From 4,050 THB / person 4,500 −10% excl. VAT 7% · group rates available
Corporate training quote Join the waiting list
PDFDownload the course outline
  • Duration6 hours · 1 days
  • FormatOnsite / live online
  • Next roundOn request
  • CertificateIncluded

Course overview

Many people already use AI in their daily work, yet the results still miss the mark. They ask for revisions again and again, or they are not sure the answer is accurate enough to pass on. Most of the time the problem is not the tool but the way the work is requested. A good prompt is structured communication, a skill that works with every tool, whether Microsoft Copilot, ChatGPT, Gemini or Claude, and one that keeps its value as the tools change. This course teaches tool-agnostic principles, starting with how an LLM works from a working professional point of view, the Role, Context,

Task, Format, Constraints structure and the CO-STAR approach, zero-shot and few-shot techniques, step-by-step reasoning, output formatting and iterative prompting. Learners then apply them to real business work: drafting emails, summarizing reports and meetings, translating documents, and analyzing problems with 5 Why, a Fishbone Diagram, SWOT and a Decision Matrix. The course builds on this with prompt chaining and prompt templates for repetitive work, plus a checklist for checking the quality and risk of AI answers. It closes with a workshop that compiles a tested prompt library for your unit, ready to use the next day. (1 day, 6 hours, Beginner level.)

What you’ll gain

  • Understand how an LLM works and why the prompt shapes the quality of the answer
  • Write structured prompts with Role, Context, Task, Format, Constraints and CO-STAR
  • Use zero-shot, few-shot, step-by-step reasoning and output formatting for accurate, consistent results
  • Refine prompts iteratively, having AI critique its own answer and ask for missing information first
  • Apply the same principles in Copilot, ChatGPT, Gemini and Claude while understanding how each tool differs
  • Apply prompts to documents, communication, analysis and problem solving in real work
  • Design prompt chains and prompt templates for repetitive work
  • Check and evaluate the quality of AI answers responsibly before using them

Who this course is for

  • Office staff in every function who write, summarize or analyze information regularly
  • Engineers and supervisors who prepare reports, analyze problems and communicate across departments
  • Managers who want AI to help prepare information for decisions
  • AI Champions who support AI use in their units
  • People already using AI who still get off-target results and repeated rework

Prerequisites

  • Has completed AI Literacy for Everyone or has some experience with generative AI tools
  • Comfortable using a computer and typing in Thai and English
  • No programming background required
  • Learners may bring examples of their own repetitive work to the workshop, with confidential data removed

Curriculum

Course Details

A 1-day course of 6 hours (09:00-16:00) that alternates short lectures with exercises, with roughly 70% hands-on time. Beginner level, and no programming background is required. The course focuses on tool-agnostic principles. Hands-on practice uses Microsoft 365 Copilot Chat as the main tool for data security, while ChatGPT, Google Gemini and Claude are demonstrated side by side by the instructor. Learners bring a Microsoft 365 account with Copilot Chat access, in line with their own organizational policy. App-specific use

in Word, Excel, PowerPoint, Outlook and Teams is covered in depth in the separate Microsoft 365 Copilot and Copilot Studio course. Assessment includes a 15-question pre-test and post-test, an evaluation of each prompt library for clarity, consistency and safety, and a post-course satisfaction survey. Learners take home a prompt cheat sheet, a checklist for checking AI answers and a tested prompt library, which can be stored in Microsoft Excel, a SharePoint List or Copilot Prompt Gallery, depending on what the organization has enabled.

Day 1: From Principles to a Department Prompt Library

Section 1 How LLMs Work, for Working Professionals
  • An overview of how large language models work, from a working professional point of view
  • Tokens and the context window, and how much information AI can take in at once
  • Next-word prediction and why AI answers differently each time
  • Why clear context matters more to answer quality than the tool itself
Section 2 The Structure of a Good Prompt
  • The Role, Context, Task, Format, Constraints framework and what each part does
  • The CO-STAR approach: Context, Objective, Style, Tone, Audience, Response
  • Choosing the framework that fits the nature of the work
  • Common mistakes in short, ambiguous prompts
  • Exercise: write an unstructured prompt and a structured one, then compare the results
Section 3 Core Techniques in Everyday Use
  • Zero-shot: direct instructions without examples, and the work it suits
  • Few-shot: giving examples to control the format and tone of the answer
  • Asking AI to reason step by step on tasks that need logic
  • Using delimiters to separate instructions from the content to be processed
  • Specifying the output format, such as a table, bullets or JSON
  • Exercise: refine a prompt with each technique and compare before and after
Section 4 Iterative Prompting
  • Refining a prompt step by step instead of rewriting it from scratch each time
  • Asking AI to critique and improve its own answer
  • Having AI ask for missing information before answering, so it guesses less
  • Exercise: refine a prompt for real work over several rounds until the output is usable
Section 5 One Principle, Many Tools
  • Demonstration of the same prompt in Copilot, ChatGPT, Gemini and Claude
  • Differences that affect everyday use: file attachments, web search and context size
  • Data considerations for each tool and choosing tools according to organizational policy
  • Why hands-on practice runs on Microsoft 365 Copilot Chat for data security
Section 6 Prompts for Documents and Communication
  • Drafting emails to customers and suppliers in a tone that suits the recipient
  • Condensing long reports and daily reports down to what readers need to know
  • Summarizing meetings into key points, owners and follow-up actions
  • Translating between Thai, English and other languages while keeping technical terms accurate
  • Exercise: draft emails and summarize documents from real work scenarios
Section 7 Prompts for Analysis and Problem Solving
  • Using AI to trace the real root cause with 5 Why
  • Grouping the causes of a problem with a Fishbone Diagram
  • Assessing situations and options with SWOT and Pros and Cons
  • Comparing options against clear criteria with a Decision Matrix
  • Exercise: analyze a quality problem on a production line or a case study prepared by the instructor
Section 8 Prompt Chaining and Prompt Templates
  • Splitting large work into chained steps where each output feeds the next step
  • Designing prompt templates with variables that colleagues can reuse
  • Writing usage notes so others can use a template without having to ask
  • Exercise: turn one repetitive task into a prompt template
Section 9 Checking the Quality and Risk of AI Answers
  • A checklist for verifying accuracy before an answer is used
  • Checking sources and handling information the AI has made up
  • Spotting bias in answers and its effect on decisions
  • Information that should never go into AI and removing confidential data before use
Section 10 Workshop: Department Prompt Library
  • Select 5-10 repetitive tasks in your unit, from your own work or a case study prepared by the instructor
  • Write a prompt template for each task using the structures and techniques from the day
  • Test the templates in Copilot Chat and score the quality of the output
  • Evaluate the prompt set for clarity, consistency and safety
  • Workshop: compile a central prompt library for your unit to share across the organization

Schedule & training options

For individuals — public rounds

No public rounds are open right now. Join the waiting list and we will contact you first when the next round opens, or ask us on LINE. Or call 02-570-8449 or 088-807-9770

Join the waiting list Ask about rounds

For organisations — in-house / private

  • Tailor the content to your team’s tools and projects
  • Your dates, at your office or live online
  • Quotation with tax ID for procurement
Corporate training quote

Instructors

Frequently asked questions

What is CO-STAR and how do you use it to write a prompt?

CO-STAR is a six-part prompt framework: Context describes the situation, Objective states the task, Style sets the writing style, Tone sets the voice, Audience says who will read it, and Response defines the output format, such as a table or a summary of no more than five points. Filling in every part means the AI does not have to guess, so you get an answer that fits the job first time.

What is few-shot prompting and how does it differ from zero-shot?

Zero-shot means giving an instruction with no examples, so the AI has to infer what the output should look like. Few-shot means including two or three examples of the input and the answer you want in the prompt, so the AI copies their format, tone and level of detail. It works well for tasks that need consistent output, such as categorising complaints or writing emails in your department's house style.

What is prompt chaining and how do you split a large task across several prompts?

Prompt chaining means splitting a large task into smaller steps and feeding the output of one step into the next. For example, first ask the AI to summarise the facts in a document, then to analyse the issues in that summary, and finally to draft a report or email from the analysis. Checking each result before passing it on keeps every prompt short and clear, and lets you fix only the step that went wrong.

Who is the Prompt Engineering for Business course for, and what should I know beforehand?

It suits office staff who write, summarise or analyse information regularly, engineers and supervisors who prepare reports, AI champions in their units and people already using AI who still get off-target results. You should have completed AI Literacy for Everyone or have some experience with generative AI tools, and be comfortable using a computer and typing in Thai and English. No programming knowledge is needed.

What will I be able to do after the Prompt Engineering for Business course?

You will be able to write structured prompts with Role, Context, Task, Format, Constraints and CO-STAR, use few-shot and step-by-step techniques for consistent results, design prompt chains and templates for recurring work, and check AI answers before using them, leaving with a tested prompt library. Participants who complete the training receive a Certificate of Completion with its own number, which can be verified online.

What do I need to prepare for the Prompt Engineering for Business course?

Bring a Microsoft 365 account that can use Copilot Chat under your organisation's policy, as the hands-on practice uses Copilot Chat as the main tool for data security, while ChatGPT, Gemini and Claude are shown in comparison demonstrations. You can also bring examples of your own repetitive work to the workshop, with confidential data removed first.