According to Microsoft’s 2024 Work Trend Index Annual Report, 75% of global knowledge workers are using generative AI to augment productivity in the workplace, and that’s a figure that’s doubled over past six months.
So, gen AI use is growing but how can we make sure we’re getting the most out of our AI tools?
The answer is effective prompt engineering.
In this blog we’ll be sharing prompt engineering best practices, tips and tricks, plus our recommendations for the best, free prompt engineering courses available for you to start your prompt engineering journey, right now.
So, what is prompt engineering and why does it matter?
Prompt engineering is the process of structuring a natural language instruction for a generative AI tool.
You’ll have heard the phrase “garbage in, garbage out” but that doesn’t just apply to data sources, it also applies to the prompts you use.
Good prompts make the difference between time saved and time wasted, with well-crafted prompts helping models to understand the nuances of our language and intent, leading to more accurate and appropriate responses.
The art of effective prompts
According to Microsoft’s Copilot Lab, there are four elements to include when phrasing your prompts:
Goal – what do you want from the gen AI tool?
Context – Why do you want it and who is involved?
Source – What data source do you want it to use?
Expectation – What expectation do you have of the results?
Other considerations for good quality prompts include:
Review and verify responses
Always review and verify the responses you receive from your gen AI tool to ensure they’re accurate. And if in doubt, cross-reference with trusted sources.
Variability in Responses
Reusing the same prompt may yield different results. Large Language Models (LLM) rely on neural networks, introducing randomness. Even with identical input prompts, variations occur.
Responsible Usage
Use Copilot ethically and respectfully. Avoid any actions that could harm yourself or others. Familiarize yourself with responsible AI principles and standards.
Use latest model
There are multiple versions of GPTs out there, including custom GPTs. Always use the most up-to-date GPT available to ensure you’re drawing on the most recent information.
Iterative approach
And don’t forget, prompt engineering is an iterative process. If you don’t get the required outcome first time round, follow up on your prompts by asking additional questions, or adding more detail to get a well-rounded result.
Be a prompt pro
If you want to brush up on your prompt engineering skills, there a tons of online courses out there, both free and paid.
We’ve identified some great free resources to help you get the most out of your AI tools.
Prompt Engineering Guide
This is a basic, introductory course based on GPT-4 usage. Focussing on practicality, accessibility and collaboration, it aims to provide a jargon-free introduction to a non-technical audience, although more technical insights can be found later in the course.
Modules include:
- Applied Prompting: Comprehensive PE process walkthroughs contributed by community members
- Reliability: Enhancing the reliability of LLMs
- Image Prompting: Prompt engineering for text-to-image models, such as DALLE and Stable Diffusion
- Prompt Tuning: Refining prompts using gradient-based techniques
Each module is rated and color-coded according to the level of complexity, from basic through to advanced.
Chat GPT Prompt Engineering for Developers
Deep Learning, in collaboration with OpenAI, has produced this (currently) free course for developers.
Those taking the course should expect to:
- Learn prompt engineering best practices for application development
- Discover new ways to use LLMs, including how to build your own custom chatbot
- Gain hands-on practice writing and iterating on prompts yourself using the OpenAI API
Although this is a beginner course for developers, with a basic knowledge of Python required, it’s also suitable for more advanced machine learning engineers.
Craft effective prompts for Microsoft Copilot for Microsoft 365
This course, created by Microsoft, explores techniques for creating effective prompts in Microsoft Copilot for Microsoft 365.
Learners will discover how to craft clear and context-rich prompts that enhance content creation, simplification, and transformation across various Microsoft 365 applications.
The course includes real-world scenarios and examples that demonstrate Copilot’s capabilities in apps like Word, Excel, PowerPoint, Teams, Outlook, OneNote, and Chat.
The course is divided into four modules, including how to:
- Summarize with Microsoft Copilot for Microsoft 365
- Create and draft with Microsoft Copilot for Microsoft 365
- Edit and transform content with Microsoft Copilot for Microsoft 365
- Ask questions and analyze content with Microsoft Copilot for Microsoft 365
For Copilot users, Microsoft’s Copilot Lab is also a great resource and includes information on data and privacy for Copilot for Windows, Copilot Pro and Copilot for Microsoft 365 users.
***
If your organization is looking to get more out of its AI tools and you need more hands-on help and advice on prompt engineering, Copilot or SharePoint Premium, reach out to the team at Cloudwell and we’ll be happy to help.