Introduction / Using GPT-5 for Strategy

00:00 Now that we have this GPT5 model available in Microsoft 365 C-pilot which can do this chain of thought reasoning, the most frequent and frankly quite reasonable question I've seen is why would we ever use the researcher agent? What is the difference between these two things? So I am going to take you through it with a reall life practical example. I'm going to start here in the

00:23 GPT5 experience and then we'll switch over to the researcher and show the difference. So this is my chat. What I need to do here to use that new model is click here. If you're not across the GPT5 model, check out this other video. It's only 6 minutes long. Come back here and and and see where we're at. So let's say I need to do something. I need to put something in here to trigger the

00:46 chain of thought reasoning that requires some kind of complexity or thought process. Let's try this prompt. Help me prepare a strategy to present to our senior leadership team on how we can plan for a project. You'll see what I'm doing here is including in the prompt a few things that give it a little bit of structure that is helpful to the chain of thought reasoning in GPT5 because it is very good at kind of following that

01:09 type of plan and thinking it through. So you'll see I'm winning. I've put in a prompt here that does require that level of complexity. So it goes through this thought process planning an internal strategy creating a detailed plan. This is using the work tab. I've actually got the Microsoft Copilot Studio implementation deck in my SharePoint in my demo environment. So that's something that hopefully it will figure out that

01:33 it can refer to as well as public things. So it's searching for various things, assessing business case details, planning research. It's picking up some recent other things out there. So you'll see it's kind of taking itself through a thought process of actually quite an impressive amount of stuff. I thought it would just refer to that document, but it's going through

01:56 quite a lot of things. All right, so here's a leadership ready strategy for me to take to the SLT. Here's my executive narrative. What we're proposing and why, why now? I always like to use examples that I know what the correct answer should be because then I can sense check it. And if you're not a co-pilot studio expert, uh trust me, these are reasonable. It's a reasonable starting point. So, here we

02:19 go. 90 days to pilot. So we've got foundations design. Okay, so you'll see the level of this risk and register mitigations. Okay, it goes on and on. Deliverables, a few notes, quick next steps, and there we go. It has referenced that wonderful document that we had in there. Now what I can do here

02:43 is take it to that next level. I was actually going to say I just want an executive summary of this thing. So, this is using chain of thought reasoning in a way that is still keeping it at a very uh practical kind of productivity type of level. If I ask a follow-up question, importantly, draft a one-page executive summary. I could have typed in anything I liked, but I just clicked one of those those prompts. It's able to

03:06 keep the context in this multi-turn conversation. It picks up that chain of thought reasoning again, but you'll see it doesn't start from the start again. You can might see where I'm going to go with the research agent in a minute. But it actually just kind of brings out those next things that it needs to be able to do. And um it is putting in my different things in here. There we go.