
Loading summary
A
MIT recently had a study that said 95% of all AI use cases are failing. What does that mean? Is anyone even getting value from all of this AI stuff? Well, we're going to show you how I use these three use cases inspired by others that I've talked to to actually supercharge my morning and to make me much more strategic, to make me much more focused and to drive clarity across the work that I'm doing, to make me a true AI and enabled leader and to be able to 10x the amount of work that I can get done. This is a morning in the life of Kieran Flanagan. Let's get into today's show. All right, so in this episode we're going to go through kind of my early morning and late morning. So. So this was kind of inspired by an MIT report that was launched recently that said 95% of all AI pilots are failing. Now I'm going to come back to that in a future show because there's a reason all these AI pilots are failing that I think I can help companies with. But because of that I thought, well, I want to show how I use AI in my day to day work because AI is being really impactful, is driving value if you know how to use it. So in this episode we're not going to do a complete day in the life because I started to do that and it would take couple of hours. We're going to do in morning in the life. We're going to do me getting on my laptop between 8 and 8:30. People will say, wow, that's pretty late. I have a toddler, so I am up from half five, probably doing parenting things. And so I spend the first four hours typically doing the kind of use cases that I'm going to show you today. I'm going to use AI as my chief of staff. That is going to help me drive clarity and focus. Then I am going to use AI to create any updates I need to create. I might have to create an update for the board. I might have to create updates for the exec team. You typically do have to spend a lot of your time, if you are across a broad range of things, updating folks in your company, how those things are going. Well, we want the AI to do that. Why do I love AI? Because I can spend more time doing the work, practicing the craft and having the. I do a lot of the other things like comms, my kind of chief of staff, those kind of things. And then the third one, we're going to use AI, my kind of late morning. So I've done my morning, I've got clarity and focus. I've got the very early morning comms at 10am I really need to switch on my focus brain. I'm a morning person and I'm going to start using AI as my red team and I'm going to show you how that works, which is basically help me think more critically, more strategic about the things that I want to do. So let's get into these use cases. All right, so what is AI Chief of Staff Now? I will say that I have a more complex tutorial on my substack that goes through this in much more detail if anyone wants to use AI to run a complete operating model. But I'm going to give you one that I use time and time again to run projects. And I think it's just as valuable. And this is one that I love because when you are running large, large teams or large projects, but even if you're not, even if you're just running something for yourself or a couple of folks, you are collecting a lot of information and it becomes hard to keep track of those things. So this Chief of Staff is really set up to do three core things that I think are really important. It's able to kind of surface any blockers you have. Really what you want to do is solve blockers for strategic projects as quick as possible. Blockers are the things that are stopping you from being successful. The other one is trying to identify KPI gaps. And basically what that is is whether you're measuring your KPIs, the things you said you would deliver, the metrics you said you would deliver, the things you told the business you would deliver. Maybe you're measuring those on a weekly basis, on a bi weekly basis, on a monthly basis. But what you want to be able to do is easily see where across those different projects you are missing those deliverables you are not doing the things you said you were going to do. And then the other one I really like is what parts of my project are behind schedule. We said we would deliver something on a certain date and we're tracking behind that date because I want to dive in and start to really figure out how to solve that. So there's three core things there. There's the blockers, there's the KPI gaps, then there's any initiatives that are behind schedule. So what we do here is we create a project and then within that project we have very clear instructions. And now you can set the instructions however you want. We are going to tell it to be clear and concise and insightful. We are going to tell it to ground itself in clear evidence so it always cite its sources. We want to surface blind spots. We want it to challenge my thinking, be respectful AI. We want to bring in external perspective where valuable. You give it access to the web. We want to think long term and systematically. We want to prioritize accountability. We want to always end with clear next steps, questions to consider, et cetera, et cetera. So you can set this up in any way you want. But that's my kind of instructions. Another instruction could be really simple, like act as my chief of staff. These are the things that I want you to replicate. Then we want to make sure that we have the core files and projects to be able to do the things I'm about to show you. And so big important ones are the docs that kind of keep a running update of these projects and show blockers that show metrics, that show core deliverable dates. And so you have to continue to. You upload those into your context. Now the problem with all of these assistants is every time you update a KPI doc, every time you update a blocker doc, whatever doc is keeping track of your projects, your KPIs, your blockers, your deliverable dates, each time one of those is updated, you have to re upload that doc. And so one of my things that I do is I use single docs to keep everything in. So we have a bi weekly sprint doc where we talk about what we're going to get done in the next two weeks and what we got done in the previous two weeks. And that's a running doc. Everything is in there. So I just have to upload one doc. We have a KPI doc where we update every month with our core KPIs. And so I have to just re upload one doc and then we have another doc that really talks about the core roadmaps and deliverables. And we update that every quarter. And because it's a singular doc again, I just have to re upload one doc. So now I have my little chief of staff here. And so in the morning what I really do is I come in here and I basically say, okay, like for my week I have this thing where I don't want any blocker to go unsolved within a given week, right? So everything has to be solved within a given week. So I can basically tell it to summarize, all the blockers from July, tell me if there's a mitigation Plan, because everyone on my team who owns something always have to have mitigation plans. And if there's not, well, the AI should suggest something. You can see here. It will pull out the kind of different work streams within my project. It will pull out the blocker, will pull out the mitigation sided. And so it will say, hey, this is only partial, no resolution yet. And then it would suggest a mitigation plan, and it will go through that for each work stream in my core strategic project. And then it will give me a summary table. And so this is how I plan my week, right? I'm like, okay, these are the hard things that I have to solve. I want to try to solve the hardest problems that can actually drive the most momentum. And so that's what I do. And so this is like use case number one for my AI Chief of staff, I actually will run this and then I will plan my week or my day accordingly. Say, hey, what's the biggest blocker that I need to solve with the team? All of these need to be solved within a given week. Let me show you one other one for the chief of staff, which is the metric gaps, which I really love. So basically what I want to do is look at anything that I missed, right? Any of the core work streams that are missing their metrics, I want to help to get back on track. That is really important. Dive into the work, dive into the weeds, make sure we get those things back on track. And so you can see here, it would pull out the work streams, it would tell me about the mismetric, it would tell me about the reason given, because you always have to give a reason if you have a missed metric. You have to give a mitigation plan if you have a blocker. And then it will go through these things. And basically again at the end, it's going to create this really cool table to show me all of the different work streams in this strategic project that are currently missing their metrics. Now I have a summary of the biggest blockers, and that's one part that can actually help me focus my time. And then I'm going to have a summary of the missed metrics and then I'm going to have consistency of pattern. So that's the other thing I love, which I put in because you can see the product trial, the paid is an emergent pattern, whereas the customer acquisition is. Yes, we've missed this four months in a row. And so this is pretty cool, right? Merging flag twice. This kind of detail is just so useful for me and so that's an example of what my AI Chief of Staff does. I set it up, I give it all of the project details, the project data, the metrics, the deliverables, the roadmaps, and I use it to clarify and focus my time. Real quick. I'm giving you my own AI executive team tech stack. This is what I use in my actual morning routine with my three AI assistants, my chief of staff, my executive communications assistant, and my ruthless strategic advisor. You can steal my prompts, my workflow, and everything in between to transform your own workflow into get it right now, grab it at the link in the description. Now let's get back to the show. Then what we want to do is, well, hey, I've got some updates here. Now I've got a bunch of information. I'm going through all of my strategic projects. Usually what you'd have to do then is do some updates for different teams for the board. And so that takes a bunch of time and so we don't want to do that. So now I switch over to my custom GPT and this is my little custom GPT that helps me do all of my comms. And so let's see what's in the custom GPT. It's very, very simple actually. And you'll be able to craft this for your needs. The instructions are basically, you're my executive assistant. You take notes metrics and you put them into executive level updates. I've only got two options here. In a different one, I have many more options. So you can have whatever are the kind of comms that you usually have to do. Now what you have to do is give it the example output. So I have here there's two options. There's a two page memo or there's an exec slide outline. And you can pick either of those things and then I give it the actual format for each. So this is the memo and then this is the deck. Now the easiest way to do this is actually take an example of one of these you've done in the past and ask ChatGPT to create a template for it. And then use that template in your custom GPT as the output format. And that's the easiest way to replicate comms that have worked really well that your exec team have given you really good feedback on that the board really like. And then that way you can replicate that each and every time so you don't have to redo it. And so that's what we do. We have two options for this one. I give it a Bunch of data that I got from here, right? I've grabbed a bunch of updates, I've put them into this kind of format and and then I just asked it to create an exact slide outline. And that exact slide outline is pretty basic, right? I have versions of this that are more complex. But just to show you how I use AI to do my work, I would basically go get all those updates, I would grab some of that information and I would put it into my custom GPT and I would craft a deck that I need to update the exact team with the board team with whoever it may be, and then it will give me this really cool outline. Now I've got my outline again, I have not really had to do much, right? I've just had to guide the AI to do it. And the cool thing is because of genspark, I can get my slide deck done pretty quickly, right? This is genspark. I just put that into genspark, I ask it to create the deck. Now you're going to say this is really basic. Again, this is just for example purposes. I have a much more complex one that I use for exact updates that are based upon our format. So you can change the format to however suits you and genspark can really just replicate that. Then I can just download this as a PDF and send it. And now I've got my AI chief of staff have done all of my focus time and now I've got my exact UpDater, my custom GPT to take a lot of that information and to craft it into a deck or a two page memo and then I quickly use genspark to create that internal comms and then I send it to my team. So now that's been my morning, right? I've come in, I've worked with my chief of staff to get the core work streams that are missing their metrics. I look at the patterns and I say, okay, I'm going to solve this blocker today. I'm going to help these teams solve this miss metric today. And then I go and say, okay, well I need to provide some updates to tell people how these different strategic projects are going. Grab the updates from the chief of staff, I give them to my custom gbt, I create the deck into genspark and then genspark creates the content for me. And there you go. That's a busy morning, you know, an hour and hour and a half. I'm pretty focused, I'm feeling good about what I need to do, but now I need to jump into actually solving problems and so because you know, I've got my blockers, I've got my core work streams and I miss some metrics. So then I jump into my red team prompt. So my red team prompt is really going to help me with my strategic thinking. So I'm going to give it a brief and then I'm going to ask it in this case to give me the top five counter arguments with evidence, any kind of hidden assumptions and how I test them and a pre mortem what could go wrong and early warning signals. What I'm showing here is AI is a really great thought partner. So let's put in a strategic brief. This is again an example. You would have taken this from the blocker, right? I would have just taken a blocker, I would have just taken a work stream that's missing this metric. I would have created a little brief which I've done here. And so this brief is basically we're planning to shift our paid acquisition budget heavily into TikTok ads. We're aiming to double signups from Gen C founders in the next 12 months. We are going to reduce Google Ads and we're going to spend by 40% to fund this. We've got a goal here to double signups, maintain our cac, increase brand awareness. We have content constraints and assumptions, right? We have a fixed budget. We only have one in house creative team. Our assumption is TikTok will remain a fast growth platform in 2025 for Gen Z the market context, right? The ICP Gen C founders building SaaS are consumer tech. Competitor a scaled TikTok in 2024 but struggle with ad fatigue. Competitor B still focus on LinkedIn. Higher CAC but stronger B2B trust. Known risks and questions will reduce in Google Ads hurt bottom of the funnel conversions. Can our small creative team sustain TikTok content demands? Will investors perceive TikTok heavy strategy as too risky for B2B? No. Look at Cluey who have about 100 billion different creators created on TikTok. Okay, so we have a brief. So again I'm just trying to really show you how this integrates into someone's day because I don't think anyone really talks about that. They just talk about a lot of hype. I want to paint the picture again that I've gone from my AI chief of staff. I've got my focus in the blockers and the metric gaps. I've done my updates that I need to do for the different teams and then I've taken one of those pieces of focused work that I want to do. It's a blocker problem. It's a work stream KPI gap. There is another one that this speaks to which is the opportunities that we have, we always keep track of like what are the bigger opportunities that we can go after we create a little brief and then we want to use chat GPT to really start helping us with our critical thinking. Now a very important thing here is you do not want to outsource your thinking. Never outsource the core skills. AI is an assistant. So when I get this, what I'm going to do is take it away, paste it into a different doc and work in a different doc with it, right? Because I want to remove myself from AI for some amount of time to apply my own critical thinking. What I'm seeing a lot more of and it does worry me is people creating things verbatim from AI and they've done none of their own critical thinking. I read through it and I'm like do you actually really believe any of this or are you just pasting what AI told you was a good idea or a bad idea? That is a terrible use of AI. There has been a study released very recently that AI can make people much much dumber. And the reason it makes people much dumber is because you don't outsource yourself your core skills to AI. You actually use AI as an assistant. So let's run this prompt and because this is a thinking task, I've got GPT5 thinking I would use GPT5Pro but I would have had to prepare that in advance because it could take some amount of time. All right, here's a tough love red team on TikTok heavy plan. So the top five counter arguments with likely evidence. So it goes through this. It talks about the platform regulatory risk is non trivial in the US which is true. TikTok's creative treadmill can outrun a small in house team goes through a bunch of data it's collected about that Cut in Google 40% risk kneecap and bottom of the funnel demand capture. Like it has some really good points points here right? Search is shifting into TikTok which helps but it's still early and volatile. What are some of the hidden assumptions? Gen C founders discovered B2B tools on TikTok and it has tests that you can prove or disprove that. Look how good these are. I can actually prove or disprove this hidden assumption. We can keep CAC to below $200 on a TikTok at scale and again it has an actual test for me to Figure out if that's a true assumption or not. This is so good one in house team can sustain TikTok's content cadence. Again, it has a really good test. So it goes through all of the hidden assumptions in that brief and then tells me how I can actually test them. And I love this one. Amazon do a really famous example of this where they do pre mortems how this fails and early warning signals. So failure mode a creative burnout hack spikes. So what goes wrong? Frequency climbs, call to action and watch time fall. Performance stalls because fresh ads aren't shipping. So this is all about the freshers you need for TikTok. You have some early signals, you have mitigations failure mode B demand capture gap after cutting Google, which basically talks about the fact that as you cut off Google you see all of these bad things happen. Audience message mismatch, views galore. But low conversions like TikTok would drive a bunch of vanity metrics and it goes through a bunch of the failure modes and then actually gives you practical guardrails. So actually gives you some really good thinking, really good thoughts about how to run this in a practical way. So this is a great starting point, right for any of those blockers or metric gaps. Then I can actually red team them with ChatGPT and specifically GPT5 thinking. And then I can use that to actually kickstart my approach here and how I want to add my own craft and skill sets to actually start to think about how I solve this. And so that is how you use ChatGPT5 really as your chief of staff, as your really strategic colleague, to supercharge your mornings, to create a ton of focus, to do your comms with your exec team and then to actually be able to red team all of these different things you do to kickstart your strategic thinking so you can apply your own skill sets and craft to make that much, much better. I hope this was useful. Three core use cases for ChatGPT and hopefully you can start applying them today in your own work.
Episode: How the Top 1% Use AI to 10x Their Productivity (Before Lunch)
Date: August 28, 2025
Hosts: Kipp Bodnar & Kieran Flanagan (Guest: Kieran Flanagan, SVP of Marketing, HubSpot)
In this episode, Kieran Flanagan unveils his unique, tactical approach to leveraging AI in his daily workflow to dramatically increase productivity and strategic clarity before lunchtime. Inspired by an MIT study claiming that 95% of AI pilots fail, Kieran walks listeners through how he integrates AI into his morning as a “chief of staff,” a communications assistant, and a ruthless strategic advisor (or “red team”), providing high-impact use cases for marketing leaders and knowledge workers.
"AI is being really impactful, is driving value if you know how to use it."
— Kieran Flanagan (01:59)
Kieran divides his morning AI use into three core use cases—each mapped to a different “assistant”:
(02:53 – 13:40)
“Blockers are the things that are stopping you from being successful… I don’t want any blocker to go unsolved within a given week.”
— Kieran Flanagan (06:08)
(13:42 – 19:30)
"The easiest way…take an example of one of these [decks] you've done in the past and ask ChatGPT to create a template for it. Then use that template..."
— Kieran Flanagan (15:56)
(19:32 – 30:55)
"You do not want to outsource your thinking. Never outsource the core skills. AI is an assistant."
— Kieran Flanagan (22:43)
“What I'm showing here is AI is a really great thought partner.”
— Kieran Flanagan (21:24)
“Amazon do a really famous example of this where they do pre-mortems: how does this fail, and early warning signals.”
— Kieran Flanagan (29:11)
On Failing Fast, Solving Blockers
“Everything has to be solved within a given week. So I can basically tell it to summarize all the blockers… If there's not [a mitigation plan], well, the AI should suggest something.”
— Kieran Flanagan (08:09)
AI as Force Multiplier, Not Substitute
“There’s been a study released very recently that AI can make people much, much dumber… because you don’t outsource your core skills to AI. You actually use AI as an assistant.”
— Kieran Flanagan (23:05)
On Making AI Actionable
"Hopefully you can start applying [these use cases] today in your own work."
— Kieran Flanagan (30:43)
Kieran Flanagan provides a detailed “morning in the life” that demonstrates specific, repeatable workflows for supercharging productivity and strategic focus with AI. He highlights the importance of maintaining core thinking skills and using AI as a leverage tool rather than a replacement. The episode is packed with real examples, clear prompts, and advice any knowledge worker or marketing leader can adapt right away for greater efficiency and impact—and, most importantly, without falling into the trap of overreliance on technology.