How I Finally Stopped Losing Track When Managing Large AI Projects | Ep 295

Please note: This transcript was generated using automated transcription tools and may contain typographical errors or inaccurately captured words or phrases.

Dr Steve Day: Keeping track of all the projects that we\'re currently working on in the business. Knowing where each one is up to, what\'s been done, what\'s coming. And making sure that we\'re actually delivering value as we work through these and not just waiting till the end. And hoping that someday we\'re gonna get value from these projects. That\'s something that I\'ve been keen to get right in my business.

I\'m, got ADHD, and so therefore, I basically forget anything that I\'m not currently focusing on. And so picking up on projects I\'ve worked on in the past used to be a massive challenge. We created what I call the Dev Project Framework for doing this. Which allowed me to track projects in terms of things called Journeys and In Stages. To see exactly what\'s being done, where we\'re up to, break work up into logical steps. To create milestones that delivered real value.

And that was good. But it was often missed because it\'s quite time-consuming to figure out and set up. The big breakthrough, which I want to talk about in this episode, came when I realized that I was losing track of my AI projects. I would start multiple things simultaneously, move on to something else, just typical ADHD distraction.

But the challenge was going back to those was, well, challenging. Because I\'d have to remember where we\'re up to, what\'s been done. And I\'ve created a number of things to help with that. However, the massive breakthrough is when I realized that I could combine that Dev Project human-centric approach. To managing my more complex, longer projects in AI.

That\'s what I want to share with you today. Firstly, how we manage our projects in the business. Secondly, how we applied AI or used it to manage the AI projects, I should say. And thirdly, the knock-on benefit of how now the AI understands how those projects are managed, it can help create and manage the project or the Dev Projects, as I call them, going forward.

This means our team is now using the proper framework for project management, which I\'ve always wanted them to. Because it\'s far less time-consuming for them to figure out, how to break projects up into stages, et cetera. Because the AI does it all for them. This episode is all about project management, but using AI as the tool to make this happen. And using that same project management system to manage your AI projects. Enjoy.

When I started my business and started running multiple projects for different things, whether it was setting up a new piece of software, designing a new course, building something cool using AI in more later years, there are so many steps involved in these. And so many different possible things we could do that very quickly many of these projects became overwhelming. They were half-finished without any clear deliverables. And so a lot of time got wasted on working on things that never really produced any real tangible value for the business.

I spent many years on building a framework that I call my Dev Project Framework. And that\'s how we manage internal projects. To make sense of what the entire scope of the best solution could possibly be. But then break it down into phases where we get the minimal viable product. The least amount of functionality that will give us some benefit.

And then what would the next best thing be. Or the next best thing going up and up until finally we might complete something that is absolutely nailed. And only needs maintenance going forward. And an example of this could be setting up a new CRM, for example.

There are so many things you can do with most CRM software. That if you start it and said, \"I\'m just gonna basically set up this and just keep going until it\'s done,\" you would never actually have any tangible milestones that you could fix on a specific deliverable date. And work towards actually getting that thing done.

So we, for example, when I roll out my latest CRM, we looked at communication first. Then we looked at building landing pages, then we looked at the social media. And we incrementally added more to it to make it into what it is today.

So I had actually developed a pretty good framework for this. But using that had always been a challenge with my staff. Even understanding how to construct these Dev Projects in a standardized way was quite a challenge. It took time to do it and also meant that people had to be properly trained up on what all the different components meant and where they should sit. And how they should actually use them.

And so often, people would just crack on and get started because they either didn\'t know how to do it or they just felt it was a big blocker to them doing the work. Or they would do it and it wouldn\'t really fit the way it was meant to be used. And so it wouldn\'t, give the benefit that it should do.

Recently, I\'ve started building a lot of stuff with AI. And I\'ve got multiple projects on the go at any one time. Just the type of character I am. I get distracted or I get excited about something, and I\'ll kick it off. And then realize that I haven\'t got time to finish it as much as I would like to, and I need to move on to something else.

And so I needed a way to keep track of these projects. I initially started doing this using the, sort of inbuilt structures within AI. Reading up some cool stuff from great teachers of AI. I\'ve got an AI coach that helped me with some of this stuff. And creating handover documents for my projects.

This is like a memory for a project in AI. Which allows you to keep track of what you\'ve done, what the decisions were made, where we\'re going, the goals, the milestones. And also, what the next steps would be. That was a massive, massive benefit. But it was all sort of hidden inside text files. And I would often have to ask the AI where I was up to and remind me of what we\'re actually working on.

And this became a bit of a chore, and I found it quite difficult to visualize what the projects were about and where we\'re up to, and just lost track of a lot of things. Because working on things simultaneously, I just wouldn\'t be able to keep it all in my brain. And I\'m a visual person. I like to see things in front of me and know what\'s going on.

That\'s the thing I loved about the Dev Project setup I had with my staff. It was a project set up in our project management tool. We use Asana, but you can do this in pretty much any project management tool. And it defines really clearly the journey steps in a project, and I\'ll explain what that is in a minute. And also the phases or the milestones or the goals that project\'s gonna work through. To actually deliver incremental benefits to the business.

The real benefit came was when I realized I could use the same exact process in my project management tool to manage the projects I was working on with AI. That meant that AI would build these Dev Projects for me in Asana. It would populate them with all of the tasks that were currently scheduled and agreed. And put it into logical phases.

So I could sit down and have a decent amount of time to work on the next phase. And then pause, go away, do some other stuff. And then come back a day, a week, a month later, pick up where I left off and be able to see exactly where we\'re up to. Without going back and forth and asking the AI what we talked about last time.

The added benefit to this, because it\'s actually human-centric, it\'s in my task management app. There are things I can pick up and drag. I can add tasks to this, just like I would do any project in my, in Asana. I can actually go away and think about this stuff and maybe go for a run. I\'ll come up with some great ideas.

And rather than coming back and then going back into the AI, I use Claude Code for this. I would then tell Claude, \"Hey, we could do this. And we could do this.\" Then trying to then schedule, like, well, does that new idea come before the last idea. Or does it become in the middle or does it get pushed to the end?

And these conversations are often quite convoluted. Because the AI would have its own opinions about the way, the order things should be done. I would have my opinions. And then I\'m trying to sort of like compromise and actually put it into a working order.

AI also forgets a lot of stuff. Like you\'ve got to be careful it doesn\'t overwrite its own memories. Because it\'s picked up on something you said and it\'s misinterpreted it. And so having something more concrete and visual, I was able to then easily sense check what it is that the AI has interpreted as what the way this project is running.

It also meant, when I got back from my run or whatever and I had those inspirations, I could literally open up the project in Asana, which is where I\'m used to working anyway with my team. And just like I would if I was working with a teammate, I could then add new tasks onto the various phases of that project. Or onto what\'s called the backlog or the new idea space.

So they are there captured and scheduled to be either, executed or done at a certain time. Or for review later. So if it\'s a new idea, for example, I might wanna just sense check before I actually slot it into any given phase. The idea here or the win here was actually using the exact same project management system or framework for my staff and the AI projects. And to have the AI update the project in real time. And also review the project each time we started a fresh conversation about it.

This means that I could do the human thing and put tasks onto the board. Move things around, like correct where it\'s made mistakes, et cetera. I could then fire up AI and say, \"Hey, check the updated board.\" It would then go away, look at all the tasks on that board. And then go and update its own plan based on the new structure.

That meant that we were both in sync. But I wasn\'t working or being forced to work in a new way. And the AI was doing all the hard sort of syncing of those two things together. This may, for some people, seem like a bit of duplication in that I\'ve got my way I deal with or manage my projects in the AI. And now I\'m forcing it to then use up tokens and use up time to actually build and maintain and sync this project to Asana.

But for me, the benefits massively outweigh the costs. Because the simple fact I have a visual representation of each project I\'m working on means that it allows me to feel calm and confident that we are going in the right direction. We\'ve both actually understood what the project is about. And I\'m actually able to interact with it in a much more humanized way than just chatting away in a text box. Which is basically how most people interact with AI.

So the next few minutes I want to just run through how we do this in a practical sense. The setup, the projects or how the projects are actually set up itself. And how we actually then use those projects to manage both the AI stuff but also the humanized stuff.

And the last thing I\'ll discuss is how much this has benefited my team and the work they\'re doing as a result of me working on it in AI. It\'s now fed back in to massively improve the adherence to using this concept, this project management concept with the team\'s own projects. Because the AI\'s able to help them out and actually get it all set up and do it properly.

So the first step to setting something up like this in your business is, first of all, you need a system for managing your own internal projects. If you haven\'t got one, then check out some of the previous episodes when I go into detail about that. And I\'ll probably record a future episode on that again.

But just to, for the sake of brevity, I\'m gonna assume you have a method that you use your project, in your project management tool for managing your longer projects, your bigger projects. That may happen in cycles or over an extended period of time. And hopefully this is working with your existing team.

If it is, you\'re totally set. And you have everything you need to then actually get AI to, A, use the same principle for managing its projects. And B, help your team then to create these projects in the future, these project planning projects. And then maintain them going forward. So as I said, the first step is to actually have a system that you currently use and are happy with.

Once you\'ve got that, you can tell your Claude Code to access the project and read how the projects are written or used. So give it access to a few of your existing projects you\'ve run, or new projects you\'ve got. Any guidelines you\'ve already created or policies around this. Give it basically all the information about how your current projects are run in your business. And ask it to analyze this so it understands how you structure this within your task management app. And how your team actually run these projects.

So for example, if you do meetings of these projects or you have reviews. Whether it\'s a end of phase review or goal review. Or however you manage to put like all that stuff goes in. Even if you\'ve got recording transcripts of meetings you had regarding the project management, chuck those in as well. Just give it all the context that it can possibly get information from. And ask it to digest that and then to come back with its understanding.

I recommend doing this in the brainstorming mode if you\'re building out something like this. Because you wanna be going back and forth to make sure your understanding is clear before you start planning anything going forward. And especially before you start building. So if you haven\'t yet done this, like in Claude Code, you can install a plugin called, I think it\'s called Superpowers. It\'s one of the Anthropic\'s own plugins, so you can find it pretty easy, and Brainstorming is one of the Superpowers.

And it allows you to have quite long-winded conversations going back and forth that don\'t eat up lots of tokens. And also allow you to get really lined up, if you like. Get on the same page with the AI. To make sure that you\'ve got all the information, or it\'s got all the information it needs. And understands what the thing is that it\'s building. So in Brainstorming Mode, we\'re gonna give it all this stuff. And then we\'re going back and forth until we are confident that it understands the way that you run your projects in your own business.

Once you\'ve got that agreed, it should then flick into Plan Mode. So we\'re not in the going into building anything yet. We\'re then gonna go into Plan Mode once we\'ve agreed the framework of what we\'re gonna do. The Brainstorming should automatically switch into Plan Mode next. It shouldn\'t jump straight into Build Mode.

In Plan Mode, it\'s then gonna go away and create a detailed step-by-step plan about what it\'s gonna do. In this case, you want it to create a skill which will- a couple of different functions. Firstly, the skill needs to build the project management structure within a new project for each of the projects it\'s running within your Claude Code.

This, just as a caveat, this wouldn\'t be done for quick one-time tasks. This is something that\'s ongoing that you\'re gonna go back to time and time again and build over time. Or it\'s, hours or hours of work rather than doing 10 minutes or an hour. Something you\'re gonna continue in multiple sessions.

So that\'s the first skill. It\'s gonna actually be able to build a project using your structure. So for us, we have the journey steps across the top of the sections across the top. And then we have what are called the phases down the left-hand side.

The journey steps, for example, if you were working on a project for building a new app, it may be that the, you have the login functionality. You have the profile review functionality, you have the booking functionality, the payment functionality, the delivery confirmation functionality, and the follow-up functionality.

And each of those different functions you could have phases for. So in phase one, for example, the login could simply be, you just log in with a username and a password that you give to the client. In phase two, it could be that you integrate Google and Microsoft and those different third-party logins.

Phase four or phase three could be that you integrate a password reset or two-factor authentication. Phase four could be adding pass keys in, for example. So you can see we\'ve got the same thing, the login. But we\'re building up the complexity as we go through the different phases. This means we can ship really early in terms of getting something out the door. Which is the least effort to create.

And then as we\'ve got more time, we can build up more complexity, add more features and more bells and whistles, so that the thing gets better over time. But we are only adding to something that\'s already live in the business. Rather than saying, \"Hey, we\'re gonna build the best version of this we could possibly can from day one.\" Because that means it\'s delayed in getting launched. And things often go wrong or you just get distracted and you never get it finished. So this incremental phase-driven project management is definitely how I would agree to do it.

So once you\'ve got that established, you\'ve got your Claude can now actually set up or, I say Claude, but it could be any AI here. I use Claude. You\'ve got the ability for your Claude to set the projects up in a logical way.

The next time you\'re working on a project, maybe within Claude, maybe you go through the brainstorming, and then you do the planning. And then before it actually then goes and starts building something, you can say, \"Great, can you go and populate that board with the plan?\" You may decide to just crack on. And then maybe at the end of your session when you might, wrap up your session and save your progress to your memory for that project, creating your handoff document.

At that point, that\'s when I typically will start a project in my project management app. So it\'s at that time when I wanna capture the status of where we\'re up to. Capture what has been done and provide me with a visual representation of that. That\'s when you\'ve now got a project management system, a visual project management system that you can interact with.

As I said, you can go into that project, for me it\'s in Asana. Go into that Asana project, look at what\'s been done, look at what\'s in the next phase. You can add things in there, move them right, drag and drop them just like you would anyway, if you\'re working with a human. Add notes on them, add the what, why are we doing this task, what the deliverables are, et cetera. Do all that stuff. But do it in a really human way.

And the next time you start the session, you can ask AI to then sync with that and to then match its plan within the context of that chat that you\'re in. With the updated version that you\'ve done manually within your Asana or your project management software.

Once you get that bit working, you\'ve now got a really useful, functioning visual way of managing your projects in your AI-driven projects. For me, that\'s been an absolute game changer. I can have multiple projects. They all have their own place in my task management app.

I can see which ones I\'m working on. I can, like that\'s one of the challenges I\'ve had is forgetting what I\'m even working on and remembering where the hell in my Claude setup did I start that project? Which folder was it in? Even trying to find the conversation that I was previous working on used to be a nightmare.

Now they\'ve all got a place. So anything I\'m going back to again and again has a project in Asana. That\'s what I\'m used to using. And it\'s also, so it allows me to actually see exactly what I\'ve started. And I can then prioritize, which ones I\'m gonna work on next and agree, where I\'m gonna, which phase I\'m gonna take that project up to before I then say, \"Pause. Okay, that can sit there for a while. I\'m gonna move on to something else. And I\'ll come back to that later.\"

Once that\'s all working, you\'ve now got a Claude or a AI system that understands your project management system completely. You\'ve tested it, you\'ve run a few projects through it, it\'s all swimming, all singing and dancing. Now you can start getting it to help your team.

So instead of it just building projects for itself, we can then get it to build these Dev Projects for your team. So I have something called the Project Vision Board. And this is a framework that I created for sense-checking projects before we start. I\'ve again talked about this on my, on this podcast in previous episodes, so go and have a little look for the Project Vision Board episode.

On there, in that Project Vision Board, we do things like sense-check, is it aligned with your company mission, your values? Is it in line with actually solving real problems that need to be solved now? Does it have real benefits? Do we have the capacity to do this? Do we have the resources to do it, the knowledge to do it?

It just makes- it\'s called a 20/20 check. And if it scores a five out of five for each of the four components, then you\'ve got a 20/20 project and you\'re ready to go forward. In that Vision Board, we are defining why we\'re doing this project? What\'s the vision of the project? And also brain dumping what our current ideas are for that as well.

Like putting down all the stuff that you as the business owner or the manager, everything you envisage this project to be right now. Obviously, that may change as you start developing it or learning more about it. But we\'re trying to just basically brain dump.

So anyone who\'s coming into this project in the future can look at that and be totally clear on understanding of why we\'re doing it, what the benefits are, what the actual expectations of the finished product will be. And also then, be able to then, or so they can then jump in and actually understand it. Without that having to be assumed or explained or otherwise.

So the idea here is if we create that Vision Board, then we can give that to the AI to build out the first iteration of our Dev Project, of our project planning project. So we can take this entire choresome. Choresome? That\'s not a real word. This cumbersome, time-consuming process of converting a Vision Board or vision idea for a project into a properly phased plan. And get the AI to do all that for you. Because it knows how to do that. Because you\'ve been using it to manage its own projects in that way.

The next thing I created, just as a caveat or aside, is I created the Vision Board Creator skill. So I can now just talk to the AI. I brain dump in literally, I voice chat it in, brain dump what my ideas for a project is. It knows my values, it knows my vision, it knows my current mission and my five-year targets, et cetera. And it can then actually do all the analysis for me and actually create the Vision Board as a draft. So I can then review it, update it.

And so that entire process end to end now of our project, deciding which projects we\'re gonna work on, the Project Vision Board Creation, and the Project Management Project Creation can now all be driven by the AI.

I would start personally- actually, it doesn\'t really matter where you start. If you start with a Vision Board, if that\'s your challenge, start there. If it\'s the project management, start there. Start with the one that you feel is the biggest chore right now that you never do or you skip. And you know that if you just did it, projects would go better. I would say get AI to actually help you build that stuff out so it becomes effortless to do. And then work on, move on to the next thing.

So just imagine that every project you or your team or your AI run in the future now has a place to be managed properly for you and anyone else involved. To see what\'s going on with or without knowledge of AI. To be able to interact with it in a humanized way. It makes it far more accessible for people within your organization to actually start utilizing AI to run projects.

And if you\'re thinking, \"Look, I\'m not a technical person. This sounds really far too complicated for me. I\'m just getting started with AI.\" I mean, that\'s kind of where I was when I did a lot of this stuff. Is that once you start using Claude Code- Claude Code can access things like your Asana through its API. And then that sound again may sound technical. All you have to do is ask Claude how to do it. How do I connect with Asana? How do I build projects in Asana?

If you wanna do very simple projects, you can just use the native connectors. If you wanna do more complex stuff, then you can ask it, \"How do I set up an API?\" It will walk you through the setup. It tells you, like it did me, how to set up an API connection. So it can build complex projects in Asana or whatever app you happen to be using.

Similarly, when you want it to then manage these projects, all I\'m doing is actually asking it in human English language, \"How do I do this? This is an idea. How do we run with this?\" Go into that Brainstorming Mode I talked about. Then flick into Plan Mode. And then go away and get it to actually build out stuff for you.

You can even get the transcript from this episode. As I\'ve said on previous episodes about building stuff in AI. Get the transcript from this episode. Which you can get on the episode page if you go to systemizeyoursuccess.com and find this episode. On there, you\'ll find the transcript we already print out for you. Copy and paste that and drop that into a fresh chat in Claude Code.

This will definitely work in Claude Code using the CLIs. So that\'s using VS Code on your computer or something. It may work sufficiently in a Claude Code desktop. You can give it a go and see if you hit any challenges. If you do, that may prompt you to then learn a bit more about how to use a CLI version, the command-line interface version.

And if that again, gets more technical, then just have a chat with Claude and ask it, \"How do I set that up?\" And take it from there. That\'s how I\'ve learnt all about AI, how to build cool apps with AI, and how to do complex stuff. Because I\'ve just worked through it systematically. To understand and learn from Claude how to actually do stuff in Claude.

So that is it for today. I hope you found this useful. I hope you found it insightful to how you can actually do some pretty cool practical stuff using AI that you can interact with in a human way. That your team can use in a human way. Without them having to understand how the whole AI thing works.

AI is changing the way we work. And the sooner you get on this and actually start utilizing it for making the stuff you already do more efficient, you\'ll find there\'s so many benefits. So many opportunities to improve your efficiency and your productivity within your workforce, your existing team. And going and expanding things out to, you didn\'t even know were possible before.

So have a go. Let me know how you get on. Send me a message or just subscribe to this podcast and catch up with future episodes if you do nothing else at all after this. If you want to live with more presence, purpose, and peace, that is what my mission is, to help you to do that. So please do subscribe and please do share this with anyone you know who runs a small business. Who may find it useful to use AI or just to use project management system more effectively in your business.

And final thing, as I said, if you\'re totally new to the idea of Dev Projects and phases and journey maps and all that, check out previous episodes where I\'ve talked about that. And also the Vision Board the same thing. The previous episodes which talk about the project Vision Board.

Thank you very much, and I\'ll see you next time. Bye.

How I Finally Stopped Losing Track When Managing Large AI Projects | Ep 295

KEY TAKEAWAYS: How to Bring Clarity and Structure to AI Project Management

BEST MOMENTS: What Better AI Project Management Looks Like

TIMESTAMPED OVERVIEW

Episode Transcript

VALUABLE RESOURCES

LINKS TO CONNECT WITH THE HOST

ABOUT THE HOST

For more articles related to project management, you may also like:

SYS 195: How We Improve Remote Team Loyalty with Health Insurance Incentives

SYS 137: Everything I’ve learned from hiring over 160 Filipino Virtual Assistants (PART 2)

SYS 151: The Strategic Delegation Playbook: Crystal Clear Policies, Structured Meetings, and Collaborative Frameworks, The Ultimate FD Interview, with Josh Keegan

Discover How Small Business Owners Leverage The SixR System To Achieve Time Freedom In Under 90 Days

How Small Business Owners
Leverage The SixR System
To Achieve Time Freedom
In Under 90 Days

Free On-Demand Masterclass