---
title: 'The Simple 4-Step Process to Build Your Own AI Trading Assistant with Claude (for Beginners)'
source: 'https://youtube.com/watch?v=45eaVU5NVi8'
video_id: '45eaVU5NVi8'
date: 2026-08-10
duration_sec: 2764
---

# The Simple 4-Step Process to Build Your Own AI Trading Assistant with Claude (for Beginners)

> Source: [The Simple 4-Step Process to Build Your Own AI Trading Assistant with Claude (for Beginners)](https://youtube.com/watch?v=45eaVU5NVi8)

## Summary

In this video, Tim and Garrett from SMB Capital demonstrate a four-step workflow for building a personalized AI trading assistant using Claude Code. They emphasize that the process is more important than the specific dashboard, and that anyone, even those without coding experience, can create custom tools to improve their trading. The video walks through planning, building, personalizing, and automating a trading dashboard that tracks trades, identifies tendencies, and provides coaching notes.

### Key Points

- **Introduction to AI Trading Assistant** [00:02] — Tim introduces his personal AI trading assistant built with Claude Code that tracks trades, flags repeated mistakes, auto-populates a dashboard, and runs a daily ritual. He built it in a single afternoon and promises to show the exact workflow.
- **Claude Code Unlocks AI for Traders** [00:29] — Claude Code enables traders to build custom tools. The video will show a four-step workflow from a blank page to a working trading assistant, including every prompt and guardrail used.
- **Target Audience: Non-Coders** [01:10] — The workflow is designed for traders with zero coding background, like Garrett's dad, who can trade well but doesn't know how to use AI. The goal is to provide a recipe that anyone can follow.
- **Key Takeaway: Workflow Over Dashboard** [02:51] — The main takeaway is not the dashboard itself but the four-step workflow that can be applied to build any tool. The first prompt can be adapted by replacing keywords to suit any trading process.
- **The Four Steps Overview** [04:40] — The four steps are: 1) Plan mode - brainstorm with Claude without making edits, 2) Build mode - give Claude the plan and let it build the skeleton, 3) Personalize - tweak and add elements, 4) Make it repeatable - set up routines for daily input.
- **Step 1: Plan Mode** [05:08] — In Claude Code, plan mode prevents Claude from making any edits. You brainstorm and create a concrete implementation plan, which is saved as an .md file. This saves time debugging later.
- **Step 2: Build Mode** [05:38] — After the plan is ready, switch to build mode, paste the plan, and Claude builds the skeleton of the dashboard. This is where the actual coding happens.
- **Step 3: Personalization** [05:52] — This is the fun step where you customize the dashboard. You can add tabs, change stats, and even add a 'coach layer' that analyzes trade write-ups for tendencies and mistakes.
- **Step 4: Make it Repeatable** [06:21] — Use Claude Code's 'routines' feature to schedule daily questions that populate the dashboard. For example, it asks about trades taken, journal entries, and best opportunities.
- **What is the Dashboard?** [07:18] — The dashboard is a personal HTML webpage, similar to TraderView but customized. It displays stats, trade history, and tendencies, and is updated by Claude Code based on user inputs.
- **Claude Code vs Chat Mode** [09:02] — Claude Code is an agentic tool that builds things, unlike chat mode which just answers questions. Plan mode within Claude Code allows for brainstorming without making changes.
- **First Prompt Example** [10:50] — The first prompt used was: 'I want to create a trading dashboard that I update every day with the trades I took, the best ops of the day, and I want to model it slightly off of a template dashboard doc with three images of a trading dashboard that I attached. But I want to incorporate a lot more into it. So for right now, I wanted use the ask me questions mode for every tab that I want on the dashboard. And since we're in plan mode, let's really take our time. The idea is I want to create my own trader view super unique to the stuff that matters to me, as well as a bunch of potential problem patterns that I want to include and have you aware of. This is going to take a good amount of time. I don't want you to decide on anything yet. So let's just take it slow. And then I reiterate use ask question mode, ask user question mode for every major decision. Batch three to four questions per round. After each batch, update the plan document.'
- **Using Ask User Questions** [13:28] — To be efficient with tokens, use 'ask user questions' mode. This presents multiple-choice questions, which take fewer tokens than typing responses. It feels like brainstorming.
- **Using Templates** [14:32] — Tim attached screenshots of his TraderView pages as templates to give Claude a visual reference for the desired end goal. This helps Claude understand the target design.
- **Plan Mode Questions** [15:42] — Claude asks questions about platform (local HTML), data storage, how data gets in (manual entry, API, or routine), and the purpose of the dashboard (review, coach, etc.).
- **Plan Document Creation** [18:21] — After answering all questions, Claude creates an .md implementation plan. You save this file and use it in build mode.
- **Transition to Build Mode** [19:57] — Once the plan is complete, start a new session, switch to build mode (accept edits), paste the plan, and say 'time to build'. Claude builds the skeleton.
- **Daily Routine Example** [20:52] — Tim set up a routine that pops up at 4:15 PM daily, asking about trades taken. This is part of his daily close ritual.
- **Step 3: Personalization in Action** [23:21] — In build mode, Tim made small changes like adding exact accuracy percentages and cumulative P&L to charts. He also added his playbook and journal entries.
- **Step 4: Making it Repeatable** [24:40] — After building and personalizing, the final step is to set up routines so the dashboard is updated consistently. This involves scheduling questions at specific times.
- **Dashboard Features** [28:22] — The dashboard includes today's entry, trade stats, a calendar view, trades by grade, and a 'tendencies' page that identifies patterns from trade write-ups.
- **Tendencies Page** [29:07] — The tendencies page highlights patterns like 'no man's land sizing' that appear in multiple write-ups. Claude identifies these automatically, helping traders focus on recurring mistakes.
- **Coach's Note** [31:07] — After inputting trades, the dashboard generates a coach's note that reviews stats and write-ups, reminding the trader of known tendencies and mistakes.
- **Customization vs TraderView** [33:12] — The dashboard is more personalized than TraderView, allowing for features like grading accuracy and custom playbook checks. It's a one-stop shop for trading data.
- **Garrett's Idea: Dynamic Resource Page** [35:07] — Garrett considers building a resource page that tracks catalysts, setups, and grading criteria. He wants AI to understand his business and provide feedback on setups.
- **Simplicity in Trading** [38:46] — Garrett cautions against over-complicating trading with complex systems. The goal is to improve decision-making in simple moments, not to build flashy dashboards.
- **Focus on Key Awareness** [40:33] — The dashboard's main purpose is to keep the trader aware of key tendencies and stats, helping them trade better in discretionary moments.
- **Intraday Use Case** [41:33] — Garrett considers using the AI intraday to check setups against his process and tendencies, potentially asking for stats on similar past trades.
- **Example: Breakout Trade Analysis** [43:10] — Tim gives an example: if a breakout trade closes below the breakout level, Claude can pull stats on how often such trades worked in the past, helping the trader decide.
- **Quality of Inputs** [44:01] — The power of the AI depends on the quality of inputs. Good reviews, playbooks, and data allow Claude to provide better insights. It's not about the AI's processing alone.
- **No Decision-Making by AI** [44:56] — Garrett emphasizes he doesn't want the AI making decisions, but rather providing data and analysis to inform his own decisions.

### Conclusion

The video provides a clear, actionable four-step workflow for building a personalized AI trading assistant with Claude Code. The emphasis is on the process—plan, build, personalize, and automate—which can be applied to any trading tool, empowering traders to create custom solutions that enhance their decision-making.

## Transcript

of HTML. Today, I have a personal AI trading assistant that tracks every trade I've made this year. It flags the moment I start repeating mistakes I've already paid for, auto-populates my dashboard while I just talk to it about
dashboard while I just talk to it about my day, and runs a full ritual at 4:15 finger. I built the entire thing in a single afternoon. And by the end of this video, so will you.
Claude Code unlocked this for anyone, and the traders already using it are talk in theory. I'm not going to tell you AI is powerful. You know that. I'm going to show you the exact
I'm going to show you the exact four-step workflow I use, and take you from a blank page to a working trading assistant today. Every prompt I typed, the exact order, the shortcuts, the guardrails that kept me out of trouble.
You'll have them all. And you can pause this video, copy them, and paste them into your own session as I go along. That's the point. You're not copying my dashboard. You're learning the workflow that lets you build yours.
You have everything you need to close it right now, today. All right, let's get into it. Garrett, how's it going? It's good, man. So, I'm excited about this one because as we talked about it for a few minutes, briefly, and you kind of
about today, I love it because this is like this is just let's lay out the actionable stuff, right? And I I I thought about myself in the situation, but I'm somebody who knows how to code,
how to code and and do a lot of this stuff on our own before AI came in. But even then, like once Claude Code came about and co-work and all you know the agentic AI stuff, like like I was even even kind of coming from
where do I start, right? Like I was asking you all kinds of questions like, what are you even doing? How do you how do you kick off this project? So, I'm thinking about you know, people who people I know, traders I know, who have
zero coding background, not much AI experience, right? And I've mentioned my dad before cuz he's like he he can trade. He knows he's doing. Um but he doesn't necessarily know what he's doing with the AI stuff,
and it's like for him, it's like, okay, can I just give you like a recipe for can I just give you like a recipe for how to build your own AI assistant that you could follow and do the same kind of stuff that Tim is doing? And that's what
I'm excited. So, why don't we just start? What's Where do Where do we begin if you're just getting into this and you want to create your own AI trading assistant? And the big like takeaway I want to have
today is that it's not about the dashboard that we're building. The idea is you're going to have the steps, four steps, that you can build anything. And what I've found, like what makes Claude awesome,
is it's not like you're doing crazy complicated prompts. It's not like you need to really know specifically what to ask. You have to just be curious, and like the possibilities are endless. And if you follow like four steps, which we
will go over, like you can build anything. So, let's just jump into it, right? >> what's the take If the takeaway is not to not the cuz you're going to show the tool that you built, right? The the
But you're saying the takeaway is not the dashboard. That's going to be like a a fun accident that people can people can certainly copy that and take away takeaway to you? The takeaway is going to just be the
workflow, you're going to be able to build any tool you possibly want. Like really any project within your trading process that would help you make you you make decisions,
discretionary trade, quantitative trade. You can really build anything, and these You can really build anything, and these four steps, all you would replace is pretty much keywords in that first prompt, which we will show you uh today.
the blueprint, and then we can cater it to anything we can dream up that is great cuz we're all different. We all have different processes and strategies and stuff like that, too. So, what works for you might be different for somebody
else. All right, let's go. And I'll uh at some point take over the screen and just use my own Claude. So, to start, Garrett, do you want to pull up image number one? Would you like to gain the biggest edge a retail trader
can get? All of our daily and weekly in-house trader meetings are now available to you. Just head over to smbtradingfloor.com to find out more. I just kind of want to flash this to you cuz this is the
takeaway. This is the main point. So, these are my four steps. You could have more steps. All you need is these four. And this is exactly how I built the trading dashboard on just a Sunday afternoon. And it
starts with plan mode. So, I go into Claude Code, and I'll show on my screen, but you go into plan mode, where Claude doesn't accept any edits. It's not making any changes to scripts or files. You're just brainstorming with the goal
of coming up with a concrete implement implementation plan. Step two is you have that plan. It's going to be just an .md file, and then you're going to go into build mode, where you're going to
give Claude Code that plan, accept edits, and it's going to build the skeleton of the dashboard. Then the third one, the third step, is where it gets fun. This is where you personalize it. You can say, all right,
I actually want this stat in this tab. I actually missed a tab, so can you add a tendencies tab? And you can really add any element that you want. And one cool thing um that I do want to show you is you can add this
coach layer, a performance coach, where it looks at For me, it looks at different trade write-ups, and looks for tendencies and mistakes, and is able to do analysis that you might not be able to see yourself.
thing, how do you implement it in your daily workflow, into your routine? And Claude Code got this cool thing called routines, where you just schedule it, questions to help fill out your dashboard every day. So, it asked me
like, what trade did you take today? Do you want to add anything to the journal, best ops, etc.? And those four steps, emphasis on plan first before you build, then you
build, then you personalize, then you make it repeatable. go into each of these and how exactly you do each of these steps. Um but when just to clear this up, so when you say dashboard, you're like, okay, we're
going to build, you know, in step two, build mode, and that's where Claude's dashboard. Like, what are you talking about? Like, pretend that I know nothing about what Claude Code even is. Like, what are you doing? Dashboard where?
Where is this dashboard? 100%. Yeah, so we've used TraderView in the past. Um I know that is common just in general. The idea was I'm going to build my own TraderView
trading, and include stats that might not be included in TraderView. So, TraderView being like a performance tracker that gives you statistics on your different setups, and you lets you tag trades, and
gives you your P&L breakdowns, and win-loss ratio, and all that stuff, right? Exactly. And when you say a dashboard, like when we go to TraderView, it's like a website, and we like log our trades, and we can look at
building like a Tim website? Yeah, it's an HTML. Okay, so it's So, it's an HTML that's now going to show you that's personal to you, no one else can get in there, right? Yep. And and
it's personal to you, and that it's going to display that is like going to be constantly updated. So, when you say dashboard, internet webpage, okay. And basically then you will input
trades, input any information you want to update the dashboard just in Claude Code, and then it will update it for me, and then resave that HTML website. Yeah.
Great. All right. I think, Garrett, let's just let's just go into the first prompt. So, So, we're go we're going to go in into plan mode, step one, right? >> Yep. We're in Claude Code. Um just to go
over it like some basics, this is the desktop app of Claude. Um you have chat here. You can then go to Claude Code. I'm starting a new session. And this is where you can go down to
accept edits. So, that's build mode right now. You're allowing Claude to edit anything, change anything. And if you click this plan mode, now Okay, and let's back up a second, and I don't want to belabor this stuff, but I
know that there like I mean, there are a lot of people that don't get this because I didn't at one point. Chat mode is one thing. Claude Code is a different thing. So, chat mode is like what we all think of like chat GPT,
where we can just talk to it, and it can give us answers. Claude Code, the difference, the agentic part, is that it is an agent for you, so it's building things. It's coding for you, and building these
these websites and consoles, and it can even manipulate your computer through you're There is a Claude chat, but you're going into Claude code. So now you're in the agentic part, but you're choosing a
toggle that allows you to plan with Claude code. So it's not actually going to do anything yet. >> Exactly. Exactly. And this is
the reason this is the first step is before you jump into any project, you really want to come up with a plan. And you'll have Claude write the plan, and you can revise the plan. And
it will just save so much time down the road of debugging, trying to fix any problems, anything you don't like with the plan before it starts changing things. This is the first prompt
that we want to use. And I'll put it in there and I'll also read it. The first prompt I wrote, and this is the exact prompt that I used that Sunday, is I want to create a trading dashboard that I update every day with
the trades I took, the best ops of the day, and I want to model it slightly off of a template dashboard doc with three images of a trading dashboard that I attached. But I want to incorporate a lot more into it. So for right now, I
wanted use the ask me questions mode for every tab that I want on the dashboard. And since we're in plan mode, let's really take our time. The idea is I want to create my own trader view super unique to the stuff that matters to me,
as well as a bunch of potential problem patterns that I want to include and have you aware of this is going to take a good amount of time. I don't want you to decide on anything yet. So let's just take it slow. And
then I reiterate use ask question mode, ask user question mode for every major decision. Batch three to four questions per round. After each batch, update the plan document. It's great. Okay, so that that first
part was like one sentence. So bravo for that. I would have been kicked out of my English classes in college for that one. But that's great. Um So you're you're basically just giving
it everything up front so that you're kind of preparing it to to get on the same page as you. Do you want to sit in on all of our daily and weekly in-house trader meetings? There's no bigger edge retail
traders can get. Visit smbtradingfloor.com to learn more. And anyone in the comment section, like the next area of Claude that I really am trying to focus on is how to use less tokens, how to be more
efficient. And even rereading that out loud, you know, you can prompt more efficiently than that, more concise. What I do What I really wanted to give was the exact thing I did to build this because it's that simple and everyone
on token usage, efficiency, It's It's as cluttered as my apartment is right now, as you can see. Maybe you can see it. renovating right now. So there's stuff There's stuff everywhere. And I
is all over the place, and that's what that sentence reminds me of. Um All right. So I'm going to run it, and this might take a sec, and we will and this might take a sec, and we will see.
So first a couple quick points. first a couple quick points. Once it ran, it saw the prompt, and what ask user question is is one way of trying to be more
efficient with Claude code, where when you're in plan mode, rather than using it like a LLM chat GPT interface where you're asking a question, it responds, you type another response, if you say use ask user questions, it's
going to be building the specs going to do it in a way of just drop-down questions, where you have a question and then you choose your response. And basically clicking
>> It's Exactly. It's like a multiple choice, and just clicking one of those responses takes up way less tokens. It might even be zero tokens, but way less might even be zero tokens, but way less than typing responses out. Okay.
in the prompt, you probably heard that I said I had dashboard examples, templates. And I basically just went into my own trader view, took a screenshot of a couple of the pages,
and I'm giving it to Claude. Basically trying to get an idea of like this is the end goal. Similar to trader view. And it just for me, I like using templates. I do it when we do our scripting. I'll give it a
template script. So it doesn't just start from a blank slate. Um >> giving it three examples of what you might want it to look like using trading view
Um service or >> one of them One of them I just found on looks pretty good." Okay. going to say. >> even you. You're You're not Nick.
>> [laughter] >> Okay. But I was like, "Oh, I like how that's set up. I like the tabs on the left." good stats. Okay. And it just has like a better idea
of, you know, what we're trying to accomplish. Yeah. And so this is where it gets fun. So confirming the platform, local HTML browser. So it's asking how I want this to exist.
So I want just an HTML file that I'm going to get in my file explorer, and I'll be able to easily get it. So I just click that, and now we move on to the next question. Where should the dashboard data live?
Where should the dashboard data live? Daily entries, notes, tags. And they tell you sometimes like I recommend this.
Okay. How does data get into the dashboard How does data get into the dashboard each day? automatically pulls your trades. You can do a manual entry
where you just type in every trade. You can even say not sure yet. And we'll defer until the end. All right. choose? So on this one, I ended up doing my own
where I was saying, "I want to create a routine on the last step on the last step that asks me certain questions."
What I found for my own trading is I like I'll give it the executions. I will then also tell the P&L what pattern it is for my playbook, and then I'll have a short write-up. So you can then just type in your own
and other, hit next. What is this dashboard for? Pick everything that applies. And just to keep it pretty simple, And just to keep it pretty simple, you can do as many as you want.
I'll do review what I did, so the historical review, and the coach accountability view. So you have that coach layer.
just for time. Yeah, great. then it will update the document, our plan.
of questions. And what's fun is it's feels like brainstorming, you know?
saw and didn't take, setups that I took, market wide.
plus detected patterns so I could group all my trades together. The end-of-day routine, what should it ask? Trade write-ups, best ops.
Trade write-ups, best ops. We'll do
Again, I'll do all of them. Just getting all the stats. questions in these batches. You'll go through all the questions, and
then it'll say, "All right, I have all the information we need, and your .md implementation plan document is finished. document is finished. And all you'll do is save that .md .md
file, then go to new session, accept edits, and just paste in your plan and say, "All right, time to build." And it'll build the skeleton.
Okay, so how does So, it decides when it is ready when it's complete, when the plan is like if the if it if the plan's tell it that we're we're good here?
It will ask a question basically saying is there anything else you want to go over before, but it'll say, "I have all the information I think I need to build the skeleton." Okay. So, at that point I say sweet because you
can always personalize it after. Right, got you. So, that gives you a great jumping off point because now it it kind of understands this architecture ahead of time before it starts coding and and creating this thing. Um
Okay. Exactly. And just lastly, before I stopped the screen share, what ends up happening is every day on the close, this is my daily
close ritual, this was 34 minutes ago. I get a pop-up on my desktop. And I even moved it back a little, but then we had the podcast today. But, it'll
say, "All right, I'm here for the daily close. What trades?" It'll start with take today?" And then I told it, you know, "I'll come back later just this one time." So, all right, so you're you're finished
with the plan mode. You have it gives you a that file that the plan is on. And you're then going into the execution and giving it that plan. And what you're showing us here is the
end result. Like, this is what This is what it what it asked you today. Do you have to do anything now? Like, in step two? Like, as we enter into step two, which is the What did you call it? Um
execution mode? Build mode. Build mode. Right. So, what what are you doing in build mode other than giving it the plan that you other than giving it the plan that you just created in step one?
That's all you're doing. That's Basically, it's literally you just you have this now, so you can start a new session and you'll then copy and paste the plan in. So,
here's just a different plan, but you can copy and place one in where you just go there, add a file,
and basically, what you saw on the routine was it looked for this daily ritual file. So,
but you would post paste it in and say, "Here's my implementation plan for a trading dashboard that I'm trying to build. I want you to build the skeleton of this." And once that is built, the foundation's there, I'll most likely
personalize and change up the layout, the visuals, and the stats included. Got you. So, that's step two. So, step two, build mode, it's going to see that right, of that architecture. Now, you're going to be able to see it, right? It'll
give you like a link or something to go to that that website, right? And you'll be able to see it, and then then you can talk to it in build mode to tweak what it has built already. Correct?
And just to show, we don't need to go through everything. Here's when I did it myself. So, this step three now, we built it.
Step three now looked like me personalizing it. So, it was very like one-off, one-by-one, I want to make this change. Mhm. For example, it had this chart where it was grade accuracy. So, it's each of my grades,
accuracy. So, it's each of my grades, A+, A, A-, B, and did I grade the trade accurately or not? Mhm. So, I'm like, "That's great. Can we actually have the exact accuracy percent as well and the cumulative P&L
on the X axis?" So, it's small changes. Mhm. And then it's like me adding in our playbook. our playbook. Okay, yeah, great.
So, that's enough of probably showing my exact poorly worded prompts. >> [laughter] >> But, step three is honestly where it gets fun, where I'm adding adding journal entries, I'm adding everything.
journal entries, I'm adding everything. And then step four is you have this HTML And then step four is you have this HTML built. It has all the tabs you want. trades, you can backfill your old trades. So, now it has data, it could
And then you're like, "All right, how could I make this so easy to input?" It's basically your DRC. So, how could I make this so repeatable that I can be consistent and I do my DRC every single day. And the easiest way is then telling
it I want to set up a routine where this is scheduled at 4:15 or 4:30 every day. And then you say, "What do I want to input?" And it'll ask you again. That's great. Four steps. So, let's review the steps.
Number one is plan mode where you're coming up with the the thing and it doesn't do anything. Then it's build mode where you give it the And then it actually just builds the skeleton right there. That's the easiest
Yes, that is. >> Step three is personalization. So, now the It gives you a link. You go to that webpage. You're saying, "Oh, cool. All this is my um What did you call it? A console or a
dashboard? This is my This is my dashboard. Now, I want to edit and tweak a few things. I might want to change a few things. So, now I'm going to talk to few things. So, now I'm going to talk to Cloud Code in build mode.
Get it all tweaked the way I want. And of course, that could be a ongoing process, right? As you find new ideas and stuff as as time goes on. is
set up a routine. So, you're basically telling it, "Okay, like periodically, whether that's every hour, every day, maybe after the close, maybe before the maybe there's something that happens over the weekend, you're saying like I
want you to either update at these points or I want you to hit me up at these points and I can I can give you inputs." Here's a mind-bending stat for inputs." Here's a mind-bending stat for you. SMB traders have 20x to 40x higher
traders. Find out why and how to greatly increase smbtradingfloor.com. The reason like we skipped over some of
the personalization and some of the questions is again like that's not the point. Like, if you have the first two steps down, like you can build anything. >> Like, it's all in that first prompt. You're just saying in plain English what
you want it to do. What do you want to build? What tool? And then you go through all of the questions and you answer what you want included, how you want it to look. And step three, build mode,
anyone's thinking, you know, we're hiding stuff, it's just it's building and you're just sitting back doing nothing, hands off the keys. Well, I think the order of operation helps um the learning curve a little bit
and like demystify this whole thing. Realizing that all you're doing is talking to it, but if you can do it in an organized way and prompt it so that and that you know, "Okay, A B C D. Like, here's my Here's
helps a lot. That definitely helps organize my mind versus just this free-for-all like, "Let me open up Cloud Code and just go nuts, right?" This is This is a very thought-out way to do it, which I
enjoy. What Tim, what's the What have you been getting out of this particular tool that you built? Like, what is your favorite part about it? Like, something that you didn't have before that you couldn't do before that you can do now?
Honestly, should I just show it? Yeah, let's Yeah, I'd love to. Yeah, let me let me pull it up. So, here it is where it starts with today's entry. And, you know, it has my trades.
Exactly, you know, all the little stats I put. The fun stuff for me, it it has a ton of tabs, a ton of tabs. What I like, you know, your classic TraderView calendar. You have all your trades based on grade.
Where this is where it's really a dashboard for me. Like, you know, you're the same way. Like, grading is our main focus. Mhm. So, having your trades split by grade helps a ton and you could go back historically.
But, what I really like, and this is my probably favorite part, is I have this page called tendencies. And these are And these are things These are different patterns in
things These are different patterns in my trading that has shown up in multiple trade write-ups. Where a this dashboard and any dashboard is really only as good or as detailed or as personalized as you
make it. So, the way I do it is I'm like giving it a lot of information in my trade write-ups about how I traded it, mistakes, what was really good about how I traded it, how was the allocation grade, accuracy, everything?
And then it was like, all right, I have all of your trades for 2026. I've noticed you mentioned no man's land sizing five times. So that's a main trading tendency that you want to always be aware of.
So did it recognize that as a pattern in your write-ups or did you tell it, "Hey, I I want you to give me a heads-up on no man's land trading if I start to mention Some of these it noticed itself. It picked up itself.
of them cuz it was a big focus. Um some of the ones that I didn't tell was this is end of month B cap where basically I try to limit the amount of B
trades I take and in the past there has been this over-grading when we get closer to the end of the month and I don't want to take as many B trades. Okay. Wrong execution style, that's an example
of one it found on its own. And what's cool is you have this page. you're like, "These are the things that I want to focus on that I messed up in the past." On any day
whenever I input my trades for the day, my entry with the trade write-ups, my journal entry, I have it do a coach's note, basically. Where it's a performance coach looking at all of the stats, looking at the
write-ups, and it knows my mistakes, it knows my tendency, it knows my journal entries, so it knows where I'm at and it's going to tell me, "All right, those were two no man's land sizing today. Like that's
Yeah. >> Which is kind of fun. That's great. That's a great way to use it. You know, you get to focus on your um your traps, right? You're tightening those things up that you need
to to not forget about. Do you have a spot in here for strength? For like things that like like good tendencies that you that you want to tendencies that you that you want to leverage?
strengths. I haven't added that yet. I'm just thinking Dr. Ken is in my head >> With his new book? >> I'm reading his new book and like that I feel like that's something that that he would say based on what I'm reading,
which is great by the way, so far. Yeah, I'm enjoying it as well. No, I that's pretty good. Here's another section that I really trades and it tallies them cuz we're marking
them, gives you your stats, your win rate. But then here's my checks in favor. Here's my rules. Here's how I execute these trades and it's all within this. So it's a one place for everything for me in my trading. So how did it get
those checks in favor? Did you give it your playbook? Exactly. Got it. Okay, great. So you're just creating a um you're just giving it a lot of the the information and and putting it funneling
it all into here so that it sort of understands your business in a way that allows it to help you. Very catered to to the individual, which I think was the main point that you made right off the bat.
right off the bat. Exactly. Where again, you want whatever you're focusing on to be the main point, TraderView didn't have a good way of me grading the accuracy of my trades. It didn't have a great way it has tags,
as we've seen with TraderView and all your different accounts. But just simple win rate P&L by pattern. Mhm. I mean, you can throw in anything. I even have Claude tips in here.
This is great. This is great. Tim, thank you so much for opening up this for everybody because I it's been think you know,
I think was sort of late diving headfirst into this stuff because I've headfirst into this stuff because I've built so many things from actual coding that I wasn't as like enthralled by all this, I think, but the
way a lot of people were initially when I heard about it because I was like, and like I don't know what more I like can do." You know, like I didn't feel like I had a need for it. But then like
and showed me some of the applications and how powerful this stuff is and the anything and this is this is great. It's a it's inspirational
for me to get to kind of start to go go a little nuts here. Hey, I love it. And I it's fun. I like showing it. I got a question for you. So now seeing the four steps, seeing that you can build a dashboard, but you can
build any tool, anything come to mind on something you might want to attack? Yeah, so I mean I think you've nailed the obvious one because like the first of trading assistant, I guess was what we
what we've been calling it. Um but I think that my think that my emphasis would be emphasis would be different. So like rather than
performance-based, like yours is very geared on like DRC, which I think is great and and I would certainly adopt that. like you know all the stuff we're doing like in our notion and like I would try
like give it that, right? Where it's keeping track of catalysts, it's keeping track of um setups that are working, setups that are not working. Um I want it to understand how I grade
setups so that I can give it information and maybe get some feedback on like what kind of grade the setup should
is. Like I I kind of have my kind of have my doubts as to whether that would be better than than than what I could do alone, but it's interesting to me. So I think that like something some sort of
think that like something some sort of like dynamic um resource page, I guess is the way that I put it. Like the like a resource page for the stuff that I care about. So there's all kinds of noise and information in
that. Um but on the other side of that, like we all have our very clearly defined businesses and within those clearly defined businesses are very specific plays. And those plays have very
specific checks in favor and variables and characteristics. And so my idea is like if the AI can understand what my businesses are and what those very specific checks
in favor and characteristics are and then on the other side I could feed it information um somehow. Maybe uh maybe it would become a really good resource page for me for either
finding setups or even just, you know, grading things that are there. That's a pretty cool idea. I like it. Especially on the grading setups front. >> the I think the reason I thought about that is just because like that's been my
be particularly good at it. Like I'm really I have my doubts. Um but I I go there because like that's what my help me do anything, it might as well be
now. So it would be like how can I figure out how to get it to help me in that area. So you know, I'm like and I'm kind of on the fence with this stuff. Like I acknowledge it's
unbelievably powerful and it clearly based on your demonstration, you can do is awesome because for even stuff like TraderView, like you can do a better version just by yourself. I mean, this
is this is why the software names are quick have been in the in the the Um awesome. And I can see so many applications, but we've been doing this for so long that we're hooked up with so
many things already that are sort of like integrated and helping us. like my dad who trades but doesn't have the infrastructure that we have. I'm for for you." Right? Someone who doesn't even have
TraderView, right? Yeah. But I also feel like trading is simple and it and it should be simple. And of course like we want to do a really good job of being alerted to
information quickly that's going to help us. We want to do a really good job of finding our best setups. But when it comes down to it, like at the end of the it's like whether you put the risk on or not on like a really good setup and it's
like how you do it. And it's like, you know, it could be the simplest thing that you that you miss or that you don't do or that you do based on really really do or that you do based on really really simple things like
right? So given how simple trading is and how simple it should be, like that's my big sort of reservation or like question mark with this stuff because the last thing that I want to do is like get in
the weeds, like deeper and deeper and deeper into this like never-ending sort of like complex um building out this like system that that seems really fancy, that makes me feel feel I'm doing like all kinds of
work, but in the end, it's like what's the ROI? Like how's it actually going to help me like crush that really simple moment more than I am now. So, I don't want to lose
sight of that. I What I want to do is like get into this. I want to experiment with it. I want to use it, but I want to do it kind of parallel to that thought of that thought. Have you heard this study of tens of
millions of traders which shows that only 1% of retail traders actually make Don't be a statistic. Visit smbtradingfloor.com to greatly increase your odds. And you can spin your wheels if you're
not careful like so many directions you can take things and do way too much work. But I think like you just said, if you have the main goal of my entire trading
dashboard is these are the things that I want to be aware of that I want to be able to access quickly in or in like to review or like pre-market.
And this dashboard like if I focus on these things, that will get me to trade better in those simple moments discretionarily. So, never losing that connection where the whole idea is
you just are doing stuff to make you the best trader you can be. You're not looking for some flashy dashboard that has these cool moving charts and like, you know, heat maps and crazy stuff. I I really like what you've done with it
in terms of personalizing it to you and having it really focus on those points think you've You've done a good job of that, which is great. element as well. Yeah, sure. I mean, like everyone's
like if I can just keep it super simple. Like again. That's the last thing I want to do. if I could just have something to check in with like during the day
that knows my process and my tendencies just like you. Like I might use it more intraday than say after the close. might be like here's the setup I'm looking at.
seeing. Here are the things kind of going against it. And going against it. And um for some reason like I intuitively feel really cautious with this setup, but I
I'm second-guessing myself. And maybe that's just a terrible path to take. Maybe talking to an AI [clears throat] about a goddamn trade it could be here? >> of the world. Like yeah, like exactly
I'm just I'm just like we're experimenting here, right? We're just This is a thought experiment and I totally understand the potential pitfalls and sort of like how crazy this sounds. But I'm interested, right?
Because if you can make it simple and mathematical and you can use it as a computer, right? Like it's like, okay, well, if it has these checks in favor like so did this this and this example.
And that these these great examples didn't have this one either and it was environment, right? It can go through your like right away. Like that that that stuff might be helpful.
My mind's going to you have a breakout day. You're getting involved in the morning. It doesn't close above the breakout level or close is right at the breakout level, which tends to be a tough decision on close.
You have Claude has all your trades and it's like, all right, like how many times did a breakout trend trade that closed below the breakout level work for Right. Right. >> you maybe just get stats and it's like
you've taken 20 breakout trend trades the past year. All of let's say 70% Yeah, so >> And none of them closed below and
second. All right, lower the grade, maybe get flat. That's that's a that's a very specific example of what I'm talking about. That that That's great. Um That's just where my mind went. No, that that's that is
that's an actual example that that is that is right along the lines of what I'm talking about. And I think that that just speaks to how it's all about the inputs. So, like the better it knows
closed above or below the breakout level. Like it has to know that or it has to be able to backtest. So, the data that you have that you're working with, if you're able to use an API and input data, if you're able to
do really good reviews, really good playbooks that have all this information in it that you can give like it's just it's all about the inputs, the quality in my opinion that you're giving it that's going to then give it the power,
right? Because just cuz it can process doesn't mean it's going to be smart. too many decisions on its own without your >> I don't want to make I don't want it making any decisions. Like that's my
That's my thing. But I sometimes might want to know um about past setups with like, you know, sometimes we'll run a backtest. It's not that different, right? But like to do that, you've got to go write a script
and then do a backtest and even that has its own nuances and complexities that kind of are are you going to really get what you want to see? And we might come back and say, hey, you know, every time an IPO
right? Um so, it's not that different than that. All right, just to summarize, we went over a four-step workflow that you can use on any Claude project. I happen to use it on a trading dashboard. And
this is the Trading Floor podcast. Uh we're still on Spotify if you choose to listen there. And in the comments, let us know what you're building.