Microsoft 365

Part 3 – How to write Microsoft Copilot Prompts

Posted on by Imran Rashid
Reading Time: 8 minutes



In this blog post, we’ll delve into the art of prompting and how to achieve the best possible responses from Copilot.

What is a prompt?


A prompt is essentially a question or set of instructions you give to Copilot, such as asking it to create, summarise, edit, or transform content. Think of prompting as having a conversation, using plain and clear language and providing context as you would with a human. Simply put, a prompt is how you ask Copilot to do something for you. Being clear and concise when entering your prompt is crucial, as it helps unlock Copilot’s full potential. If you don’t get the desired result with the first prompt, you can keep the conversation going, just like you would when communicating with a human. Following up on your prompts helps you collaborate with Copilot to gain more useful, tailored responses. However, aiming to include what you need in the first prompt will help Copilot provide you with the required response. We will explore a few examples in this post.

Here’s a brief two minute video that explains what a prompt is. I recommend watching it before proceeding.


It’s important to remember that when you enter a prompt, the more specific and detailed your instructions are, the better Copilot can customise its response to meet your desired outcome.

Let’s use a couple of examples to illustrate this. Imagine you walk into a restaurant and simply ask the waiter for food. The waiter would likely ask for more details, such as what type of food you want and whether you prefer to dine in or take away. Similarly, if you ask a builder to construct a house, they would need more information, like the type of house and the number of rooms you want. When working with AI, we need to adopt a similar approach. The quality of Copilot’s response depends on the details you provide. If your instructions are vague, the response will be limited, and you’ll need to provide additional prompts to get the information you need. It’s ok to continue having a conversation with Copilot but it’s possible to limit that conversation with less prompts.

Creating effective prompts when working with Copilot will improve over time. However, at a minimum, aim to have a clear goal in mind when crafting your prompt. If Copilot’s initial response isn’t exactly what you need, you can always refine it by providing additional information through follow-up prompts. For instance, you might get the right information but find the response too lengthy. In that case, you can ask Copilot to shorten it. Think of it as an ongoing conversation with a human. The key is to be clear and concise from the start, and then build on that if necessary.

To receive the best response from Copilot the first time, think about including the following four parts in to your prompt if possible or at a minimum, a goal and context.

Goal
What is it that you want Copilot to do for you, summarise an email, generate a document, creating a bullet point list and so on. Provide it with specific instructions of what you require from Copilot)

Context
Why do you require the information? what is it you’re going to do with the information? for example, is the information needed for an upcoming presentation for school children.

Expectations
What type of tone do you wish for Copilot to respond. Should the response be in a simple and friendly manner so school children can understand it. Should the response be a couple of paragraphs or a page of text.

Source
Do you need Copilot to refer to a particular document, from a website or from an email conversation between you and a colleague where the email includes a list of topics which need to be discussed in the presentation.


It’s not always important to include all four parts into a prompt, a couple if parts may be sufficient, but including the four parts into your prompt provides a more specific set of instructions to Copilot and a good response.

Let’s go through a few examples,

Example 1 – Copilot Prompt

I input the prompt below at copilot.microsoft.com

“Write about water”

Copilot provides a response about water, but I could rephrase the prompt so it is more specific.


Let me try again, but this time I’ll provide a clear goal and make it specific.

Goal: Please write a speech to prepare me
Context: for an event where I will be presenting a topic about the benefits of drinking water to children
Expectations: so make it sound casual and youthful.
Source: Refer to data from the following link Water, drinks and hydration – NHS (www.nhs.uk)

Here is the prompt:

Please write a speech to prepare me for an event where I will be presenting a topic about the benefits of drinking water to children so make it sound casual and youthful. Refer to data from the following link Water, drinks and hydration – NHS (www.nhs.uk)

It’s not always important to all four parts into a prompt, if needed we can always input further prompts. For example, I could input a prompt without the source and let Copilot provide a response based on it’s own sources. Such as,

Please write a speech to prepare me for an event where I will be presenting a topic about the benefits of drinking water to children so make it sound casual and youthful. Refer to data from the following link Water, drinks and hydration – NHS (www.nhs.uk)

and here is the response from Copilot. Copilot has provided a response with a tone that can relate to children, such as the sentence below.

Hey everyone! I’m super excited to talk to you today about something really important – water! Yep, that simple drink we all know and love. But did you know that water is like a superhero for our bodies? Let me tell you why.

Additional wording has been used such as awesome and playground to catch the attention of the children.

It has also included in the prompt recommended amount of water as per NHS UK guidelines as per my request.


Furthermore, I can continue with the conversation and ask Copilot to adjust the response if needed.

If I am not happy with the response or Copilot provided me with incorrect information, I can use the thumbs up or thumbs down icon to feedback to Microsoft. It’s important to verify the information Copilot provides, as it may sometimes provide a convincing, yet incorrect response. Use the feedback options to report such responses to Microsoft.


Let’s try that prompt again, this time removing the NHS website as a source.

Please write a speech to prepare me for an event where I will be presenting a topic about the benefits of drinking water to children so make it sound casual and youthful. Refer to data from the following link Water, drinks and hydration – NHS (www.nhs.uk)

Before I enter the above prompt without the last part. I click the option for a new chat so Copilot no longer uses the information I have already provided it in my recent conversation.


Prompt and response below.

This time Copilot has used generated me a different response without any reference to the NHS website. I could also reference an internal document if needed, but the point here is that depending on what you ask for is what you’ll receive from Copilot.

Example 2 – Copilot Prompt

Another prompt that I could input into Copilot could be,

Please write me a technical blog post article

Copilot responds requesting for more detail about what technical topic I would like Copilot to write a blog post about. Yes, we can provide that additional information, but if you focus on including that in the initial prompt, you may get what you need from the first prompt.

This prompt may work better. I have not used a source in this prompt, but I could ask Copilot to use a document or a website when putting together its response.

Please write me a blog article about the benefits of Microsoft 365 Copilot. The blog post must be 1,000 characters and simple to understand for someone who is new to the Copilot for 365.

Goal: Please write me a blog article
Context: about the benefits of Microsoft 365 Copilot
Expectations: The blog post must be 1,000 characters and simple to understand for someone who is new to the Copilot for 365.
Source:

Here is the response from Copilot. Much better, however I can always insert another prompt and ask Copilot to shorten or make the blog post longer.


Finally, lets ask Copilot to explain what a good prompt should include.

Create a bullet point list of what I should take into consideration with prompting with Copilot.

Goal: Please write me a blog article
Context: for an event where I will be presenting a topic about the benefits of drinking water to children
Expectations:
Source:

I could add additional prompts if the response required additional information, however, I feel the below response is what I need to summarise what to consider when prompting with Copilot.


Example 3 – Copilot Prompt

Give me 5 bullet points for me to use in a talk from the Copilot email discussion I had with my colleague Darren last week. The response should be easy to understand and keep it short.

Goal: Give me 5 bullet points
Context: for me to use in a talk to my colleagues
Source: from the Copilot email discussion I had with my colleague Darren last week
Expectations: The response should be easy to understand and keep it short

Example 4 – Copilot Prompt

I want a list of five games to play for a team get together at Las Vegas. Provide a list of ideas that are fun and family friendly and include any supplies that may be needed. Please take ideas from the Internet.

Goal: I want a list of five games to play
Context: for a team get together at Las Vegas
Expectations: Provide a list of ideas that are fun and family friendly and include any supplies that may be needed.
Source: Please take ideas from the Internet.

Example 5 – Copilot Prompt

Create me 5-8 bullet points to prepare me for a meeting with Client Contoso Ltd to discuss their new “Phase 10” brand campaign. Focus on email and teams messages since August. Please use simple language so I can get up to speed quickly for the meeting tomorrow.

Goal: Create me 5-8 bullet points
Context: to prepare me for a meeting with Client Contoso Ltd to discuss their new “Phase 10” brand campaign.
Source: Focus on email and teams messages since August.
Expectations: Please use simple language so I can get up to speed quickly for the meeting tomorrow.

And remember, keep the conversation going. Following up on your prompts help you collaborate with Copilot to gain more useful, tailored responses.

Looking for additional prompts?

Check out Copilot Lab, a free resource which provides hundreds of example prompts you could use.

I hope you found this post useful

In the next posts, I explore and demo Copilot in 365 apps such as Word, Excel, Powerpoint, Outlook and Teams. Stay tuned for more prompting. If you’ve not already done so, please subscribe to stay informed of new posts.



Part 2 – Microsoft 365 Copilot under the hood

Posted on by Imran Rashid
Reading Time: 7 minutes


In part one of this blog series, I provided a short introduction to Microsoft 365 Copilot. Here is the link to that post if you missed it, Save time and be more productive at work with Microsoft 365 Copilot – Part 1

In part 2 of this blog series I explore how Microsoft 365 Copilot works under the hood.

After being assigned a Microsoft 365 Copilot license, the Copilot icon will appear in the different Microsoft 365 Apps. We will showcase and demonstrate Copilot in several of these Apps in a future post.

Step 1

Andrew Doe, a Manager, returns from holiday to find a lengthy email discussion including a few attachments about a new office location project. Upon opening the email, he asks Copilot to summarise the conversation and identify any actions assigned to him which he needs to be aware of. As a user when we ask Copilot to do something, such as summarise an email or drafting an email, this is known as a prompt. More on prompts later.

The Summarise button will summarise the email conversation with the email thread.

However, if you wish to ask Copilot to check for any outstanding tasks in the last couple of weeks, there is a Copilot button which works across Outlook as a whole instead of focusing on one email thread. See image below.

Step 2

The Copilot orchestration engine receives the prompt from Andrew Doe’s Outlook application.

Step 3

The Copilot orchestration engine undergoes a task known as post-processing or grounding, during which it accesses Microsoft Graph and Semantic search. Microsoft Graph is basically your Microsoft 365 data, such as your calendar, SharePoint, OneDrive files, meetings, chats, and more. Additionally, Copilot can search other services using plugins and connectors, such as a Bing search plugin that allows access to internet content or third party applications such as ServiceNow. This grounding/post-processing step enhances the quality of the prompt, ensuring you receive relevant answers.



What is Semantic search?
The semantic index is a new feature of Microsoft 365 search that uses the Microsoft Graph to better interact with your personal and organisational data. Relevant information is obtained in the Microsoft Graph and semantic index to provide the Large Language Model (LLM) with more information to reason over. As an example, suppose you want Microsoft Copilot to locate an email where a colleague praised the design work of a vendor. Semantic index includes nearby words (for example, elated, excited, amazed) into the search to broaden the search area and give the best result. All of this work takes place behind the scenes to add relevance to results that you search for with Microsoft Copilot. Another example of Semantic search, it’s like a librarian who not only knows every book in the library but also understands the story behind your question. Traditional search is like looking for books with a specific title, while semantic search finds books by understanding the story you’re really interested in, even if the title is slightly off.


Step 4

The Copilot orchestration engine combines the user data retrieved from graph and Semantic search and sends the modified prompt to the Large Language Model (LLM).



What is a Large Language Model (LLM)?

There is a lot more to LLMs but to simplify, so this post is in a no way a deep dive into LLMs. Here are a few points about LLMs.

Large language models (LLMs) represent a class of artificial intelligence models that specialise in understanding and generating human like text. In the context of Microsoft 365 Copilot, LLMs are the engine that drives Microsoft 365 Copilot capabilities. You may have heard/read about the company OpenAI who developed the popular ChatGPT service. Microsoft have invested billions of dollars into OpenAI and the LLMs they develop. The ChatGPT models are utilised by Microsoft, however, Microsoft privately hosts these models on the Microsoft’s Azure OpenAI Service, so your company data is not shared with OpenAI. Microsoft aims to push the boundaries of AI research and development. By partnering with OpenAI, they can leverage cutting edge AI technologies and innovations. This collaboration is seen as a way to accelerate AI breakthroughs and ensure these benefits are broadly shared with the world.

A few points about LLMs below.

1. LLMs are used to understand user inputs and generate relevant responses.

2. LLMs allow computers to understand and generate language.

3. LLMs specialise in understanding and generating human like text.

4. Operate as generative AI, producing new content and can have a real conversation mimicking human behaviour. It can be difficult to tell whether you’re having a conversation with a human or a machine.

5. Provides the engine that drives Copilot capabilities. The LLM is what provides a response to our prompts/instructions we send it.

6. Instead of merely predicting or classifying, generative AI, like LLMs, can produce entirely new content.

LLM’s are trained using a large amount of data sourced from the Internet, books, conversations, movies and a lot more. An LLM can be used for all sorts of tasks including chat, translation, summarisation, brain storming, writing poems, code generation, writing a book, troubleshooting, writing a FAQ, image creation/detection and a lot more.

That’s where the name Large Language Model comes from. A large amount of work goes into training the LLM. In simple terms, a LLM is a super intelligent auto complete so if we input Roses are _______. The LLM will respond with the next word of Red. At the time of training these models, the LLM will make errors and is then trained/corrected. For example, if an LLM responded with Roses are Green, the team of data analysts would retrain the LLM with the correct answer and this process continues as the LLM fine tunes itself and gets better.

We can compare an LLM to how neurons/brain cells work in the human brain. In the human brain there are some 80 – 100 billion neurons with 100 trillion connections to each other. The brain is structured so that each neuron is connected to thousands of other cells. Human brain cells form a very complex and highly interconnected network which send electrical signals to each other to allow us humans to process information.


Let’s take an example of a toddler/baby who is shown a picture of a dog. At first the baby will make mistakes when learning to identify the differences between animals. When a baby incorrectly identifies a dog as a cat, a parent or teacher may correct the toddler and the more practice the baby gets overtime by viewing pictures of different animals or seeing animals in the real world, the neurons in the brain adjust, allowing the toddler to get better at identifying animals correctly.

Data scientists created LLM’s in a similar way to how the brain works. An LLM is like a human brain made up of a neural network, each neuron is connected to the others. As mentioned earlier, the LLM is pre-trained on a large amount of data. For example, an LLM can be provided with pictures of thousands or even millions of pictures of a dog and then is trained on how to identify the correct one. When an LLM makes a mistake in identifying an animal, it is corrected and the neural networks start to adjust and this process continues as the LLM learns. Similar to the way we learn as humans.

In the diagram below each circle below represents a neuron. When we provide an input we expect to receive an output. Under the hood we have the hidden layer where all the processing takes place before we are provide with the result, known as the output. Simply put, as we make mistakes and learn, neurons are activated/deactivated.


Scientists discovered that the neural network within a large language model (LLM) can be structured to allow neurons to loop back into previous layers, enabling two way communication similar to human neurons. This breakthrough led to more complex behaviors in LLMs, culminating in the development of ChatGPT by the company OpenAI, which can engage in human like conversations. Microsoft invested in OpenAI and use the LLMs in their products. As you’ll appreciate, there is a lot more to this topic and the information I have provided is basic, but I hope this provides you with a simple overview.

Step 5

The LLM (Large Language Model) retrieves the prompt from the Copilot orchestration engine and generates a response. It then returns the response to the Copilot orchestration engine.

Step 6

Copilot takes the response from the LLM and post-processes it. The post-processing involves aditional grounding calls to graph, security, compliance, privacy and responsible AI checks. This is a final check before it is safe to forward the generated response from the LLM to the user Andrew.

Final Diagram


Stay tuned for the next post where I will explore Copilot in several 365 Apps

Part 1 – Save time and be more productive at work with Microsoft 365 Copilot

Posted on by Imran Rashid
Reading Time: 4 minutes


In this blog series, I’ll delve into the advantages of using Microsoft 365 Copilot. In this first post, I’ll discuss the common challenges we face in our daily tasks at work and how Copilot can help alleviate these burdens, save time and money, and boost our productivity.

We all feel the pressure of work. Information, deadlines, and constant communication can often overwhelm us. AI can help, not just by making work easier or faster, but by making it more fulfilling. When we don’t have to spend as much mental energy figuring out what happened in that meeting, catching up on emails, or finding that document from last week’s chat, we can focus more on the core of our work and the purpose behind it.

In recent years, the pace and volume of work have continued to increase. Data from searches across Microsoft 365 services reveals that on any given workday, Microsoft’s most active Microsoft 365 users:

  • Conduct 18 searches for what they need.
  • Receive over 250 emails.
  • Send or read nearly 150 chats.

Globally, Microsoft Teams users are now in three times as many meetings each week compared to 2020. Additionally, some people use 11 different apps on Windows in a single day to complete their tasks.

AI helps lighten the workload by boosting human abilities and speeding up natural creativity. When leaders learn to use AI effectively, they can enable their teams to embrace this new era of AI-powered productivity, bringing great benefits to their organisations.

Before diving into Microsoft 365 Copilot, let’s compile a list of tasks we typically handle each day at work. On an average workday, you might find yourself:

  1. Catching up on email threads among colleagues.
  2. Engaging in numerous Microsoft Teams chat conversations.
  3. Reviewing recordings of Teams meetings you missed due to other commitments.
  4. Sending various emails to colleagues and external partners.
  5. Creating PowerPoint presentations
  6. Analysing data in Excel, budgeting, transforming it into tables, graphs, or pie charts
  7. Locating emails and Teams messages where you’ve been directly mentioned with the @ symbol and tasked with specific actions to complete.
  8. Reviewing outstanding tasks from this or last week, including important actions assigned by your manager.
  9. Checking emails or Teams channels to ensure you haven’t missed any company announcements.
  10. Planning for the upcoming week’s tasks and meetings.
  11. Organising the next team get together and ensuring fun activities are arranged.
  12. Reading through a lengthy 100 page document in preparation for a meeting the next morning.
  13. Recalling the last time you had a meeting with a specific colleague.

I could list additional daily tasks, but you get the idea.

How can Microsoft 365 Copilot help?

Microsoft 365 Copilot isn’t just another feature introduced by Microsoft. It’s more than that, it’s your AI (Artificial Intelligence) powered Copilot that accompanies you, the Pilot, throughout your day to day interaction with Microsoft 365 apps such as Outlook, PowerPoint, Word, Excel, Teams, Loop and Whiteboard. Copilot was developed to save time and make you more productive by being able to generate new content mimicking human behavior. This is known as Generative AI where machines are able to generate new unique content and respond like your interacting with a real human being.

Going back to the list of daily tasks I created at the beginning of this post. Well, Copilot can assist with addressing those challenges and more. Those challenges we are all aware of at the workplace where the pace of work is overtaking our ability to keep up with our daily tasks. Microsoft 365 Copilot is designed to assist and reduce that burden, such as being able to generate new emails, summarise email threads, summarise a large word document, summarise team meetings you were not able to attend, create you a PowerPoint deck, generate a business proposal, generate a job advertisement, locate email and teams conversations where you were @ mentioned, list your outstanding tasks for the week and more!

Remember, Microsoft 365 Copilot is not replacing you, it’s your Copilot and you’re the Pilot.

Image generated by Microsoft Copilot in Bing

Is Copilot the same as a search engine like Bing or Google?

Is Copilot the same as a search engine like Bing or Google? Not exactly. Copilot is more advanced than a search engine. When you use a search engine, you may ask a question like, “How do I fix this plumbing issue?” The search engine will then scour its index of relevant content and present a list of website links for you to explore. You then have to sift through these links to find the information you need or perform another search.

Copilot, on the other hand, uses a pre trained Large Language Model (LLM) to perform a similar task but with a twist. It doesn’t just find relevant content; it generates new content, providing a direct answer to your question, such as how to resolve the plumbing issue. This process is known as Generative AI. I’ll cover Large Language Models (LLMs) later in this blog series.

Here is a short video from Microsoft which summarises and provides an insight into Copilot.


Is Microsoft 365 Copilot free?
No, this particular service requires a license for each user who will be using Copilot in your organisation. Microsoft 365 Copilot is available as an add-on plan with one of the following licensing prerequisites listed at the at the following Microsoft Learn page, Microsoft Copilot for Microsoft 365 requirements.

Before exploring Microsoft 365 Copilot within the various Microsoft 365 Apps, such as Word, Excel, PowerPoint, Teams and Outlook, I explore how this AI Powered Copilot functions under the hood and provide a high level architecture overview through a number of diagrams.

Click the link below to progress to the next post. See you there 🙂

Copilot for Microsoft 365 under the hood