How to Write Microsoft Copilot & AI Prompts That Actually Work!

Summary

Highlights

• Using the GCSC framework: Goal, Context, Sources, Expectations for AI prompts! 🎯
• Improved AI email drafting by clearly specifying goals and sources! 📧
• Enhance meeting summaries in AI by providing precise goals and context. 📝
• Encourage specific tone and language settings for more accurate AI outputs. 🌐
• Treat AI as an intern; provide detailed instructions to enhance results! 🧑‍💻

Key Takeaways

• Optimize your AI prompts using the GCSC framework for better results! 🎯
• Be specific in your prompts to prevent AI from filling in gaps inaccurately. ✍️
• Always set a clear goal, context, information sources, and expectations in your AI instructions. 🎯
• Interact with AI like you would with an intern, providing complete information for the best outcomes. 💪
• Use AI’s feedback positively to improve continual interaction and results. 🔄

Overview

Chapters

• 00:00 - 00:30: Introduction to AI Prompt Challenges This chapter introduces the concept of AI Prompt Challenges by highlighting common frustrations experienced when using AI tools like Microsoft co-pilot or smart speakers. It suggests that these frustrations often stem from issues other than the AI tools themselves.
• 00:30 - 01:00: The Importance of Optimized Prompts The chapter discusses the significance of using optimized prompts when interacting with AI. It highlights that, unlike casual human conversation which can lead to varied outcomes, detailed and optimized prompts can lead to better results when using AI tools.
• 01:00 - 01:30: Framework for Better AI Prompts This chapter introduces a framework designed to help users craft better prompts for AI tools like Microsoft Copilot and ChatGPT. The framework aims to improve the user experience and results when interacting with such tools. The speaker encourages viewers to subscribe for more content, mentioning that a significant portion of the audience are not subscribers.
• 01:30 - 02:00: Understanding Frameworks with STAR Example The chapter introduces the concept of frameworks by inviting the reader to engage with and learn from weekly content. Frameworks are described as basic structures or underlying systems that are essential for understanding core concepts. The chapter suggests considering practical applications of frameworks, such as their use in job interviews, to illustrate their importance and utility.
• 02:00 - 02:30: Introducing the GCSC Framework In this chapter, the focus is on introducing the GCSC framework, which is instrumental in preparing for interviews. It highlights the importance of thoroughly addressing the interviewer's questions, which often aim to uncover specific competencies or behaviors crucial for securing job offers. A common mistake candidates make is to jump straight to the results when describing a situation or area where they exhibited great behaviors. Instead, it is essential to detail the entire scenario, including the actions taken, to provide a comprehensive response.
• 02:30 - 03:00: Applying GCSC to Co-pilot Prompts The chapter discusses the use of the GCSC (Global Catalog of Social Competencies) framework in formulating effective prompts for co-pilots, particularly during interviews. The STAR method (Situation, Task, Action, Result) is highlighted as a tool to assess whether a person provides adequate information to show their behaviors or competencies relevant to the role. This approach is advocated for its effectiveness in evaluating candidates by structuring their responses methodically.
• 03:00 - 03:30: Example Prompt without Framework The chapter discusses the responsibilities and actions taken to ensure a positive outcome for a customer experience. It emphasizes the use of the STAR (Situation, Task, Action, Result) framework in interviews to structure responses effectively, demonstrate desired behaviors and competencies, and improve the chances of securing a job role.
• 03:30 - 04:00: Improving Prompts with GCSC Framework The chapter introduces the concept of improving prompts using a specific framework called GCSC. The speaker acknowledges that although the tutorial is centered around enhancing prompts with AI and co-pilot, it is important to understand the GCSC framework. This framework is not an academic qualification but a method defined by the speaker to create better prompts.
• 04:00 - 05:00: Summary of Better Prompt Results The chapter discusses a framework for improving prompts and introduces the concept of 'gcsc'. It breaks down into four key areas: (1) identifying the goal of the prompt to be sent to co-pilot, (2) considering the context around the scenario being worked on, (3) evaluating information sources that can be used to provide co-pilot with relevant information.
• 05:00 - 06:00: Tips for Effective AI Tool Use The chapter discusses tips for effectively using AI tools, focusing on setting clear expectations for responses. Suggestions include defining the form of the desired output, which could be tables, lists, paragraphs, or emails, and being precise about this when interacting with tools like co-pilot.
• 06:00 - 06:30: Conclusion and Encouragement to Subscribe This chapter discusses the GCFC framework, a method for creating better prompts in Microsoft Co-pilot. The author addresses the audience with a challenge to try the framework and see the improvements in their results with Co-pilot. The emphasis is on practicality, as they plan to demonstrate the framework's effectiveness through examples. The chapter ends with encouragement to use this technique in all future prompts for enhanced outcomes.

How to Write Microsoft Copilot & AI Prompts That Actually Work! Transcription

• 00:00 - 00:30 have you ever used Microsoft co-pilot or other AI tools and become frustrated with the results maybe you have spoken to your AI smart speaker only to be confused with the relevant result or even drafted an email with the help of Microsoft 365 co-pilot only they're needing to rewrite the email immediately because it didn't meet your requirements or Worse simply delete the result but here's the thing often it's not the AI tools that cause those problems it comes
• 00:30 - 01:00 from the way we communicate with AI itself we can communicate with AI using something called prompts which is the ability to give AI an instruction but as humans we often get involved in casual conversation and that conversation can elicit many different requirements leading to really good outcomes with AI tools it's much better to give them a very detailed and optimized prompt to get started with to get better results first time
• 01:00 - 01:30 and in today's tutorial I'm going to share with you a framework to use to improve your prompts and get better results from AI tools whether that's Microsoft copilot chat GPT or other AI tools out there on the market I'm sure you'll be able to take this framework and improve the way that you use AI tools and before we dive in and check out this framework if you haven't already do hit that subscribe button to find more great content just like this because did you know over 90% of people that that watch our channel here are not
• 01:30 - 02:00 subscribed to the channel so why not hit the Subscribe button and find great content like this every single week and also continue to help our Channel grow so let's go and check out what this framework can do for us so what are Frameworks we can consider these as a basic structure for an underlying core concept or a system and if we also consider where Frameworks are used how about we consider going to a job interview when you sit down in
• 02:00 - 02:30 front of your interviewer they're going to give you many questions those questions are often going to cover competencies or behaviors that you should show to how to get that amazing job offer but the problem is that when many people get interviewed and as to also describe a situation or an area that they also have shown some great behaviors in they'll often jump straight to the result we forget to tell the interviewer about what happened in that scenario or the actions they took to
• 02:30 - 03:00 come to that great conclusion that then means that person has not given the right information to the interviewer and therefore maybe he's not showing the behaviors or competencies required for the role but using a framework such as star it gives us the ability straight in the interview to think about that question and apply This Acronym which means situation task action and result we take the situation that led to this amazing experience the tasks and the
• 03:00 - 03:30 responsibilities that we took to ensure that great outcome and what were the actions that we took personally then led to the result being this amazing experience for one of our customers using the star framework we can immediately see that when we get a question from an interviewer we can structure it in a way that gives the best possible response meaning we have an idal opportunity to also show the right behaviors and competencies and ultimately get that awesome job roll
• 03:30 - 04:00 but I know what you're thinking this tutorial Scott is not built around you getting a new job and they absolutely agree we're certainly here to talk about better prompts with AI and co-pilot but understanding that framework we can now consider the framework we can use for better prompts and that's something that I've defined is gcsc now if you're in the UK and watching this video you may think that's an academic qualification but in this scenario we're certainly be
• 04:00 - 04:30 focusing on it as a framework to improve prompts but what does gcsc mean we can break it down into four areas the first is what is the goal of the prompt that we're also going to send to co-pilot the next will be the context what's the context of what I'm also going to send into co-pilot around the scenario that I'm working with the next will also consider the information sources I could use being the sources of information that we could give co-pilot
• 04:30 - 05:00 to generate a better response maybe a document or a website and finally e is for expectations with the response that comes back from co-pilot what are you expecting now at this point we're not going to give co-pilot the answer instead what's the form of that response maybe it's a table of suppliers you're working with or maybe it's a bullet point list or a paragraph or an email being clear with expectations or s ures
• 05:00 - 05:30 better results so it's a simple framework gcsc to build out your better prompts but if you don't believe me how about we go into Microsoft co-pilot and try this out I'll show you a couple of scenarios using prompts I've seen many people use that get okay results but applying the gcfc framework will'll see much better results from co-pilot first time that you could use in all of your prompts moving forward so let's head into Co pilot and see how this framework can
• 05:30 - 06:00 work for us so we're now in Microsoft co-pilot and I'm going to use the free version for work meaning it doesn't have any access to our business data but it does come with Enterprise data protections what we can now do is share a prompts with our co-pilot this will be built not using the framework I've shared with you today we use a prompt I've seen many other people used before and not generate some great outcomes so let's go ahead and send this into co-pilot
• 06:00 - 06:30 this prompt now focuses on drafting an email to John Smith outlining that we would deliver the best Microsoft 365 it consultancy and training for businesses around the world and they should get in touch to find out more now of course if we send this into the co-pilot it will go ahead and generate us an email it won't come back and ask any clarification questions and in the email it's now generated we can see here it's simply okay it includes information that I also gave it inside of the email but
• 06:30 - 07:00 it hasn't taken any information on who we are as a company or also the goal I'm trying to achieve and also there are no expectations around the type of email that's going to be generated what we're going to be looking for when the person finishes reading this email so changing the prompt to something that follows our gcsc framework we can get a much better outcome let's now share another prompt with cpilot using a better prompt so in our new optimized prompt we can see here that we're going to draft an email John
• 07:00 - 07:30 Smith once again for me introducing our company your 365 coach we're going to highlight our it consultancy and Training Services but the goal is to secure a meeting with him to discuss how we can help his business I'm also outlined to co-pilot that we want to set up of call with John to talk about our services in addition for the context I want to mention that we deliver topnotch Microsoft 365 Solutions and to use our website as information source to take outline our services and unique
• 07:30 - 08:00 offerings that means we use specifics from the web to build out the email for expectations like to encourage John to schedule a call with us via the form and our website and to ensure this email we're going to create is also professional given is to co-pilot you'll see immediately a different result begins to search the web and find relevant information taken from our website we can now see the email itself is very different yes it includes John's name but now also in addition to
• 08:00 - 08:30 mentioning our company it's taking information from the web covering the training the consultancy and our master classes we offer to businesses and also a link to our website alongside that even has the expectation or the goal included I'd love to schedule a call with you to discuss how we can support your business in maximizing the benefits of Microsoft 365 just use the form on our website so you can see this email is very different the only different being using a gcsc framework the only
• 08:30 - 09:00 difference being with copil it we're very clear in the goal the context also the information sources and the expectation to car an overall better email that we can add paste into an email of your choice and send over to John and taking this another step further let's consider Microsoft 365 co-pilot has access to your business data we could use a gcsc framework to also get better results to catch up on meetings let's go ahead and share our prompt of our Microsoft 365 co-pilot to
• 09:00 - 09:30 summarize a meeting that I also missed earlier on this week and to get a summary from co-pilot and by adding this prompt that I need a summary of this meeting and simply selecting the meeting from my co-pilot experience we can now have a summary generated by co-pilot but the prompt is very open-ended it's a very short sentence just to let the co-pilot know I wasn't able to attend to summarize the key points from the meetings and while this response we've got back from when co-pilot is helpful
• 09:30 - 10:00 it could be better for example there isn't a goal I've set for co-pilot what am I trying to achieve by getting a summary what's the expectations or the context I'm setting where's the information coming from we could also take much more information to Microsoft 365 co-pilot from emails or teams chats so to optimize this prompt with our gcsc framework and see that we can get a much better result from co-pilot with our new prompt now added into co-pilot you'll see it's very different once again
• 10:00 - 10:30 following the new framework we've shared at this point I'm of course asking co-pilot to summarize the key decisions and action items from the meeting rather than a broad summary I'm specifically letting co-pilot know what I'm looking for in the meeting summary in addition I've asked co-pilot to review my recent emails and teams chats in relation to project greenspace and include that in any responses the expectation is I need a concise summary that highlights the main outcomes and next steps which is separated into sections around
• 10:30 - 11:00 the meeting itself and anything in my team's chats or my emails sending this into co-pilot will now see that once again it reviews a transcript taken from the meeting but this time it will go further looking across my teams chat messages and emails and looking further at the sponsor and co-pilot we see an overall better outcome because I very clearly stated the goal of what I'm trying to achieve summarizing the meeting points and understanding action items that's very clear from the information copon has returned with in
• 11:00 - 11:30 addition we've also set out those information sources picking up information from my emails and teams chat which are included in the initial response there's no need to ask further questions to find that out and also the expectations were set around what I needed in in what information you can see here that the structure is now set in different sections just like I expected from my copon at prompt so immediately in one single prompt we got a much better result with more
• 11:30 - 12:00 information in a structure that we can also better understand using the gcsc framework so now you can use gcsc Frameworks to improve your prompts what else should we also consider when working with co-pilot or AI tools well first off be very specific if you're vague in your prompts of co-pilot or AI in general it often inserts information to help the prompt that may not actually be true or even worse it may hallucinate
• 12:00 - 12:30 and come up with information that you actually never shared of it but it believes to be correct so where possible always be specific with your co-pilot or AI prompts to ensure the best outcomes and don't forget to tell co-pilot or your AI tool the tone and language to respond in I find often that AI tools like co-pilot are fairly neutral in their tone if you want to draft out a professional email or tell co-pilot exactly that is in a particular language
• 12:30 - 13:00 once again ask it to do exactly that as well what I will do with multiple languages is ask it to generate it in two languages one in English and one in French allows me to ensure that the Tona language used in the English version will also translate well to the French version so once again ensuring you tell your copon AO tool to set the tone and the appropriate language how should we also consider co-pilot all these AI tools often I hear from many people we
• 13:00 - 13:30 should consider these to be interns or assistants in your team they don't know how specifically you're going to work so giving it as much information as possible in the prompt will yield better results just think if it was an internal assistant giving it vague information as I mentioned earlier will also yield pretty poor results to do try and think of your co-pilot assistant as just that and don't forget when co-pilot or your AI tool also responds in a positive way
• 13:30 - 14:00 also let it know it's done exactly that you want to follow on the conversation co-pilot as an example will know it's got the right answer and to build on it whereas if you don't it can also alter the result and change it so making sure you share with your co-pilot or AI it's actually answered it in a correct and positive way is a great way to continue the conversation on and remember of using co-pilot chat GPT or other tools if there's an ability to reset the conversation by a new conversation or
• 14:00 - 14:30 new topic button hit the button when you need to do exactly that thinking of this as humans and casual conversation you can imagine the confusion you would have if I started talking to you about what I was doing on the weekend and immediately stop the conversation to ask for feedback on a contract you may not be able to respond to that question more adequately because I Chang the conversation with no real context ensuring the conversation Fred is relevant to that particular topic is
• 14:30 - 15:00 really important in AI tools so now you know this framework you can use it both in the workplace and at home and get better results first time from your AI tools and that will make you more productive or even more efficient than what you using ai4 previously and if you like this video hit the like button to let me know if you haven't already hit the all important subscribe button to find great new content like this every single week and we don't want you part of the 90% who don't don't subscribe
• 15:00 - 15:30 other than that I look forward to seeing you in the next one