Claude Builds Live Dashboards In Minutes

Free Resources & Next Steps

• Free Claude Cowork Webinar | 8/19:  https://bit.ly/ClaudeCoworkWebinar_819 
• Free Claude Cowork Masterclass: https://www.biggestgoal.ai/l/cowork-masterclass 
• Claude Cowork Starter Kit: https://www.biggestgoal.ai/l/cowork-starter-kit 
• Claude Cowork Workshop: https://www.biggestgoal.ai/l/cowork-workshop
• Get our free daily AI News Brief: https://bit.ly/BiggestGoalCommunity

About the Episode:

Executives sit on piles of data from Stripe, calendars, CRMs, and countless spreadsheets, but can't easily see it all in one place without making it a major project. AI has changed that, and Alane Boyd and Micah Johnson show exactly how to create live dashboards in minutes.

If you've ever felt frustrated jumping between ten different platforms just to get the data you need, this conversation will change your approach. The hosts get real about solving actual business problems with Claude's Live Artifacts feature.

In this episode, you'll learn:

• How to connect Stripe data using native MCP connectors for revenue tracking
• Why calendar analysis revealed shocking productivity insights in under 60 minutes
• The n8n + Supabase workaround when MCP connectors hit data limitations
• Real examples of dashboards built live during the recording session
• How to move from scattered data to unified insights without hiring developers
• The difference between static reports and AI-powered live analysis

If you're ready to stop drowning in data and start building dashboards that actually update automatically, press play now.

Upcoming Education Sessions:

• Free Claude Cowork Masterclass

Learn More about our AI Leadership Workshop

Winning businesses aren’t just working harder, they understand how to use AI strategically. Sign up for the AI Leadership Workshop, a 3-hour expert-led mastermind designed to give leadership teams clarity, alignment, and a practical path to stay competitive.

Book a call to customize your workshop today!

This episode is brought to you by Biggest Goal, the company Alane Boyd and Micah Johnson co-founded.

Fifteen years of scaling their own companies taught them that adoption is a people problem long before it's a technology problem, so we train executives and their teams through workshops, cohorts, mastermind groups, and hands-on implementation support.

For more info about our programs, join our free community at your.biggestgoal.ai.

Thank you for supporting the podcast!

Transcript

Copy

Alane Boyd (00:05)
Executives sit on piles of data, Stripe, Calendar, CRMs, and so many spreadsheets, but can't easily see the data all in one place without making it a major project. AI has changed that. Now we can create a dashboard in minutes, and in this episode, we will tell you how.

Micah Johnson (00:28)
All today we're gonna talk about a brand new feature in Claude Cowork called Live Artifacts, which when I first heard about this feature, I was like, what the hell is that? That doesn't mean anything to me.

Alane Boyd (00:42)
I mean.

I mean, it is kind of weird because it doesn't tell you what it actually is.

Micah Johnson (00:48)
Yeah, I think that's the problem. But I understand where they're coming from. If you've used Claude for long enough, even when you've chatted in the past, and I'm sure, you've probably seen this as well. Like you chat with, gosh, do we call it old school chat now? Pre-Claude Desktop Cowork functionality, where you would have Claude create what was called an Artifact, which was basically like a...

Alane Boyd (01:02)
Yeah.

Micah Johnson (01:12)
pseudo document that kind of lived in the chat conversation.

Alane Boyd (01:17)
And so it's an Artifact that stays there so you can keep using it. And in this case, creating a dashboard.

Micah Johnson (01:25)
There's obviously plenty of use cases for this and we're going to talk about a lot of them in this episode, but it's live in the sense that it's pulling data live. And the big picture use case here is live dashboards that you can create as you probably heard from the intro to this episode, literally in minutes.

Alane Boyd (01:47)
There's so many pain points and I'll tell one of my pain points that I was running into with my team and where some of your day is literally just going to different places trying to get the data that you need or checking up on things. And so we needed a better way to, we use Stripe for our invoicing, subscriptions, things like that.

And that is how people register for our cohorts and our trainings. Well, when we started doing these trainings, we had to change the way that we are basically doing our bookkeeping, where it's unrealized until the cohort or training starts. So it just needs to stay in Stripe. That way, if there's any changes or anything that money hasn't been used yet, as in is where it's coming from in a bookkeeping accounting perspective.

in Stripe, there's not an easy way to see that number. You just see a total number. So you gotta dig or keep a spreadsheet is what we were doing. We were keeping a spreadsheet of all of the trainings that were purchased and holding onto that total before transferring it into the bank account. And so I had the idea, Stripe has an MCP server that's a Connector in Claude Cowork, Micah.

Micah Johnson (02:59)
Yeah, and Live Artifacts came out at the end of last month. It was great timing. And I will say there is a piece here. You can upgrade your Stripe account and get some additional functionality. You can buy add-ons for QuickBooks and have some workarounds with this. But hey, you know, we're an AI company. We need to experiment with this stuff and...

The ultimate end result was we didn't have to spend any more money. It took us less than an hour to get this all up and running. There were a few workarounds that we'll talk about here in a second, but we wanted to really put this to the test and say, can solve this problem with a specific feature with just going from idea to execution?

Alane Boyd (03:42)
you

I know it's so cool because as a non-tech person, I got a lot of the way there with the dashboard that I created using the native Stripe MCP Connector that is in Claude Cowork. And so there were some other things that I wanted to see, unpaid invoices, subscriptions, all of this data, including the unrealized versus realized revenue so we knew when we could push money through.

What I kept running into though, as I was working and building out this dashboard, and you're just writing low level prompts, just telling it what you're wanting and playing around with the visual data as it creates it in real time with you, is that it kept telling me some of the data that I needed from the metadata on what was purchased, you couldn't pull from the MCP server.

Micah Johnson (04:37)
Alane, I'm going to have to call you out for a second because you started that whole thing saying from a non tech person and then you went on to explain. Well, I used a Connector, I tied into an MCP server. Oh, I couldn't get the metadata, like let's be honest. You are slowly turning into a tech person.

Alane Boyd (04:41)
You

⁓

No, I just sound really good when I say these things. you know, when I think the biggest thing and on a side tangent, Micah, is that you have to try things and not be scared of trying it. And so that's what I've gotten really comfortable with. It's like, well, I'm not going to break anything. I'm building on top of the data that exists. I'm not changing the data that exists. So I just get more.

Micah Johnson (05:20)
Yeah.

Alane Boyd (05:21)
comfortable trying things without feeling like I'm gonna screw up.

Micah Johnson (05:25)
I love that as just a quick point to make is, and I think this goes in with anything related, maybe just technology as a whole, but not being scared to try things. And then when you try it, it's like it starts clicking. I don't know. Hopefully that's not just me, but that's how I tend to learn this stuff just try it, see what happens. If I get an error,

Well, then I learned from solving that error. So let's talk about the roadblock that you ran into with this. You have your MCP, everything's connected, and it starts telling you, we don't have the metadata for that.

Alane Boyd (05:53)
Yes.

Yeah, so I did hit a roadblock. So I wasn't able to finish my dashboard the way that I wanted to with getting the realized versus unrealized data in there because the MCP could not pull that metadata that I needed. So I called you, like one does at Biggest Goal, and I say, Micah, this is the roadblock I'm hitting, and you figured out a really cool workaround.

Micah Johnson (06:25)
from a technical perspective, the issue was that the MCP server capabilities and tools didn't have enough functionality to get at the data that Alane wanted to pull. And so we had to figure out a workaround and the workaround, this applies to a lot of things. And it's essentially, if we can't get it through the direct connection with Claude,

through like an MCP server, the direct Connector, we've talked about those in previous episodes, then what we ended up doing was building an n8n workflow that can connect directly up to our Stripe accounts API, which is the full capabilities and Stripe was originally created with developers in mind. So it's got an extremely powerful API and can do all kinds of things. So what we did was use the n8n workflow to connect up to the API, pull all the data that we need,

transform that data and then save it in a database. And the way that this worked was we saved it into a secure database that Claude could connect with through an MCP server. So then we passed the baton back to Alane and Alane, you were able to essentially pull part of the data you wanted directly from Stripe and the other part of the data you wanted from the data that was in the database.

That stays up to date because the n8n workflow just runs behind the scenes and keeps the database updated.

Alane Boyd (07:41)
Mm-hmm.

Yeah, it is so cool. It was so easy to do from a implementation point on my side. you gave me the file. I worked with Claude to download. Like it was so simple for the handoff. And then what I did is I just took what I had built with the previous one with the other data points that I needed and just incorporated it into the new master Stripe dashboard.

Micah Johnson (08:11)
Yeah, so what I love about this use case is, Alane, you started with an end objective in mind. You already had the vision of what you wanted in the dashboard, what metrics, how you wanted to see it, and then you worked backwards to figure out, how do I get that data into it? It required really no development because even the side that we did on the n8n workflows and getting into the database, we leveraged Claude Code to help us with that. So that helped us

Alane Boyd (08:18)
Mm-hmm.

Micah Johnson (08:39)
rapidly create the workflow, to do the heavy lifting, to get the data out of Stripe, to do all that stuff. I don't think we used any code at all, but it's very minor lift at all. And really all through prompting and some drag and drop interfaces in n8n and a database system, Supabase. We've talked about that in the past as well

which is a fantastic database solution and designed with AI in mind. So now you are just connecting all these dots. But the point that I was trying to make before I got off on my tangent was you started with the end objective in mind. I'd like to share a story of mine using this where I started with a, I have no idea where this is going to go.

Alane Boyd (09:21)
you

Micah Johnson (09:23)
But I wanted to experiment with these dashboards in Live Artifact. So I started with some of the most boring data known to man, which is my calendar. And ⁓ there's a lot happening. But when you think about it, it is not exciting data, right? It's like event name, event date, duration, like start date, end date, times, attendees.

Alane Boyd (09:32)
Boy, there's a lot happening there.

Micah Johnson (09:45)
It is not exciting by any, it's not like, what's the unrealized revenue we're gonna get this month? ⁓ That seems more exciting to me. But anyway, so I started with, hey, let's connect Claude up to, it was already connected to my calendar, but let's pull that into a Live Artifact. And I just had it visually show me my calendar. Cool, I can already do that in my calendar. That's not very exciting. So then I wanted to push it a little further.

Alane Boyd (09:51)
Yeah.

Micah Johnson (10:13)
And I said, well, what if we visualize this so it automatically strips out all the blocks that I have throughout my day so that I can get focus time or space between my meetings or drive time, anything like that. So I prompted it and Claude figured out a way to automatically strip out all the blocks. And so now I've got a much more clear version of my calendar, which is already a bonus. But then I started looking at that going, man,

Alane Boyd (10:39)
Mm-hmm.

Micah Johnson (10:42)
my goal is to have a maximum of external meetings a day maximum. And so let's throw that as the threshold. So I go back to Claude and I say, hey, give me an indicator anytime my day involves more than four external meetings. Well, then I had to figure out, what's an external meeting versus internal meeting. And I worked with Claude and categorized all the meeting types. Now this is without me touching my calendar at all. This is

pure random calendar data with crappy names and stuff coming in from Calendly and stuff that I'm adding, And I worked with it and it figured out, well, let's categorize certain things as drive time. certain things as internal meetings. certain things as educational or like sales.

type meetings and then everything else gets into generic meetings. So now I can take that and I can see a threshold of how many external meetings am I having a day, which days are over those as I'm looking forward in my weeks. But then I looked at that and said, wow, it'd be really cool to break this down by hours in a day and assume that I'm going to work in eight to five. And then what does that break down based on drive time, meeting time, sales time, internal meetings, whatever.

And how many hours per day does that lead? What am I really working with? And again, this is without me editing anything. This is without me doing any time tracking outside of this. And suddenly I'm starting to paint this visual diagram with charts and percentages and pictures, and it's color coded. And I can see by the day and by the week and in a complete chart how much time I have to actually get quote unquote work done.

Alane Boyd (12:02)
Mm-hmm.

Micah Johnson (12:26)
versus different meeting types overall. the final thing that I added to this was, let's look at, second to final thing, let's look at if my quote unquote work time is less than three hours a day, let's make a big red border and highlight these days during my week so I can understand I'm already maxed out. Let's not add more to the plate here. And immediately I have this amazing analysis

of my week and how it breaks down. And it's already caused me to look at my days differently and look at how I spend my time and look at when I should or shouldn't book more meetings in a specific day. I can make that judgment call based on data rather than just gut feeling and hoping it gets in there. But it's actually changed how I feel about my days. It's really crazy.

Alane Boyd (13:19)
Yeah, I thought it was interesting because we kind of ran that same experiment on mine and what we realized is like, my gosh, Alane is in way too many things because she every day is red. where we a lot of times we do base things off of feelings. Well, I feel like I spend a lot of time doing this. I feel like I spend a lot of time doing that. Well, when we have the data in front of us, it was this is where we need to find a

resource for or problem solve where we're spending our time if we're not seeing that as a valuable time spent for one of us.

Micah Johnson (13:54)
Yeah, absolutely. And so you can make better decisions. It's that old adage of, well, I don't know how old it is. It's the adage of you can't improve what you don't measure. And so I would have never thought about looking at my calendar data in any of these ways. and all of this is less than 60 minutes to put together and have this spectacular, beautiful visual with charts and everything. And like,

I don't know, 60 seconds to apply it to your calendar, It was just like, just do this, but change the calendar or add the calendar. There's no way I could have justified that type of development of calendar data analysis or hiring a data analyst to analyze my week and days to give me this to make better decisions.

Alane Boyd (14:24)
Yeah, it was really fast.

Micah Johnson (14:43)
But in 60 minutes of work without any additional costs, hell yeah, I can make a business case for that. where this gets really powerful, and then I'll stop with this story, is now that we have that calendar data all analyzed and organized and it's happening all on the fly because it's a Live Artifact, well, we can start pulling in data from other sources. So let's get back to the revenue side of things. Now is where it gets exciting.

Alane Boyd (14:53)
You

Mm-hmm.

Mm-hmm.

Micah Johnson (15:10)
Okay, how much am I doing in meetings related to revenue? And we can have these two separate data sources that in one Live Artifact dashboard can be correlated and analyzed in specific ways. And we're getting the benefit of programmatic, you know, it can run Python scripts, but it's also getting the benefit of AI. And so it's a very, very interesting feature that

Alane Boyd (15:15)
Mm-hmm.

Mm-hmm.

Micah Johnson (15:38)
I think is going to tell a bigger story as this feature improves over time.

Alane Boyd (15:43)
Yeah, that was actually gonna be one of the things that I commented on before we transitioned to your example is the Stripe dashboard was the first piece of this. Well, we also have QuickBooks online and we run invoicing and money management for the business through that. So I don't want just to be looking at things from a Stripe perspective. So it's going to turn into a master revenue dashboard for the company, but then we can also start

having the AI analyze it and start acting like a CFO type person where we can look at it. Every time we click on that Live Artifact or that dashboard, it refreshes the information. So as long as the MCP call or the API call, if it's a more complicated version, has updated data, then your dashboard is gonna be updated accordingly.

Micah Johnson (16:35)
Yeah, it's absolutely changing the game, think, Alane, like when you really start to think about the use cases of this. So let's give one more super simple use case that everybody can maybe relate to. And as we were talking about different use cases that we wanted to share, where there are tons, there's a pain in everybody's ass, which is I've got to go and check my.

project management system and see what tasks I have to do today. I've got to check my calendar. I've got to check my emails. I've got to check this. I've got to check that. I've got to check Slack. you kind of said this at the beginning, you're checking all of these places, but you don't need to anymore. And this is a beautiful use case. In minutes, you can literally go to Claude and say, the assumption here is it's already connected to these different data sources, but just say, hey,

Pull the latest deals of mine, pull the latest emails that are unread, pull my tasks due today or overdue, pull my calendar and put it in an agenda. And you design a dashboard like this once, and then it's there for you. And the data and all these systems change, but you're only going to one place to see it all.

Alane Boyd (17:27)
Mm-hmm.

Micah, I've got a bonus one that I didn't tell you about. Because I literally did it as I was setting up my recording studio. So it took me less than two minutes. So I have a ton of travel coming up. And whenever we were talking about the calendar and our email and everything, check in, a little light bulb went off in my head. like, I could have a travel dashboard. So I have it checking the next week of travel for me.

Micah Johnson (17:48)
⁓ my gosh.

Okay.

⁓

Alane Boyd (18:13)
and it's pulling my, where I'm going, pulling my airplane arrangements, my confirmation numbers for my hotel, if I got a rental car or not, and built out everything, like a little mini dashboard for everywhere I'm traveling in the next seven days. And if I didn't do a rental car, it says not found. So I know like, ⁓ maybe I missed it, or maybe I'm not getting a rental car for that one, but I'm looking at it right now. It is so cool.

Micah Johnson (18:41)
OK, all right, so I already know the answer to this, but I'm going to ask it as a bridge to what we want to talk about with some of the issues. I would love a dashboard like that. Also for my upcoming travel, can you share it with me?

Alane Boyd (18:55)
Okay, so the dashboards are personal to you.

Micah Johnson (18:59)
I get it. I get it.

Alane Boyd (19:01)
but I can give you what I told it and it can create one for you.

Micah Johnson (19:05)
Yeah, so we do need to call this out. This is a very early stage feature inside of Claude Cowork, and there are no share Alane and her amazing travel dashboard, I hope that's what you call that, Alane's Amazing Travel Dashboard. There's not enough, anyway.

Alane Boyd (19:19)
That's exactly it.

Micah Johnson (19:22)
There's no ability for her to share that with me directly, even on a team account, even on an enterprise account. You can't share it. It's an individual dashboard has made for herself. There are workarounds. Like you can ask Claude to package it up and write a prompt. And we have tested this out and it does work pretty well. You can get a package and a prompt and say, all right, now run this. Give this to Claude Cowork as a task and run this prompt.

Alane Boyd (19:40)
Mm-hmm.

Micah Johnson (19:51)
and it can replicate this dashboard. And so you can do it that way because of the strength of AI as a tool, but there's nothing built in to actually share it.

Alane Boyd (20:03)
Yeah, and I mean, being somebody that experienced that where you packaged it up for me and then shared it with me, it was so easy. So there are ways to share it, but if you make any changes to it, like I took what he sent me and then I developed it a little bit more with some of the other pieces I wanted in the dashboard, he doesn't have access to that.

Micah Johnson (20:23)
Yeah, it's not a shared database. So Alane gets version 2, and I'm still stuck on version 1. But pretty sure that's going to come down the road here inside of this. It would be insane if Anthropic did not add this feature to this. But that is one thing to watch out for. Another thing, Alane, think

Alane Boyd (20:36)
I think so too.

Micah Johnson (20:47)
you experienced this recently as well. You tried to create a new one and Claude's trying to be helpful, but when you go to create a new Live Artifact, it pre-prompts you, right? Like did this just happen to you recently?

Alane Boyd (20:53)
Mm.

Yes.

Yeah, it pre-prompts you and it puts in a bunch of old stuff that may, if you use like create in the past, it basically uses that same discussion. So my Live Artifact that the new one I was trying to build, it was getting confused. And so don't do that, at least for now, cause it's doesn't work right. Just go and create a new task and Claude Cowork and say you want to create a dashboard and it'll walk you through everything.

It basically builds out its process and how it's thinking about things first. And then it says, all right, I'm ready now to create the Live Artifact and it'll go and create the Live Artifact.

I have one more item on this, just talking about limitations, Micah, is that in any software build that you're doing, where you're connecting things, you are always gonna be limited by the data available in an MCP, an API, or a webhook. So you can't magically do every idea you ever wanna do. It's if the data is available by that software product. So where I hit a roadblock in the MCP, it was available in the API.

Micah Johnson (21:48)
Yeah.

Alane Boyd (22:12)
Had it not been, there's nothing you can do.

Micah Johnson (22:16)
Yeah, yeah, I think that's a really good call out. You know, if they don't make the data available, the data is not available to get, you know, there's some longer workarounds, there's manual workarounds, you can manually export stuff and put it somewhere and, and then can process it and get it into a database and then you know, but like, already, right? Like, how bad do want that data?

Alane Boyd (22:28)
Mm-hmm. Right.

Right. It got complicated.

Micah Johnson (22:41)
But thank you for calling that out, because that is super important. That is not the magic bullet of saying, ⁓ any data anywhere at any time, and now I can build a dashboard. It has to be accessible. And your user has to be able to get access to it, because this is still not free data, free for all. This is still authenticated, secure data that you're getting in, just like any other API connection.

Alane Boyd (23:06)
Yeah, and I think the last thing I would just say on this is that you might have to do a little massaging with working with Cowork on building this out. I was running into some issues with one of the dashboards I was trying to create and it kept pulling no data. So I had to just keep working with, I'm like, no, there's data in there, like try looking again. And so sometimes it just needs a little bit of reprompting or trying to ask it in a different way if you're not getting the result that you want. If the data is not there,

like in my Stripe situation, it came back and said, the MCP data is not there, I cannot pull that. So there was no massaging I could do to improve that result.

Micah Johnson (23:44)
Yeah, and basically what this is introducing is what it's like to work as a lead developer with junior developers. No, I'm just kidding. I'm not going go down that route. But essentially, it is just like working with a developer or a person to say, they might not see it the way you see it. You might need to say, well, let's reorder this. Let's change this. Let's pull this data from this other piece or whatever it is.

Alane Boyd (23:54)
Bye.

Micah Johnson (24:10)
It's going to try to do its best based on the prompts that you give it, but that's all you're really giving it is the prompts and the connections. And it's got to make assumptions. so like you were saying, Alane, it's going to make assumptions. And then you just correct it. You say, well, let's move this over here or move this down, or we don't need this metric at all, or make this a chart, or color it this way, or whatever. And you just work through your vision.

All right, so I would say the big thing from here is go, like we said this earlier in this episode, right, Alane? Like go try it. You're not gonna break anything. The worst case is you make a crappy dashboard, but go out and try it, get the feel of it, get the vibe, and chances are you're gonna be able to build something really cool.

Alane Boyd (24:58)
Yeah, just have fun with it. If you need help setting up your Claude Cowork because having the access to data is really important, Micah and I have an AI mastermind that we meet every other week with a team of executives and leaders and we talk about these kinds of things and not just talk about it, but we show how to do it so that you're never stuck and we always are a resource in helping those mastermind members get unstuck.