Jake HellerApr 20, 2026
Build a Back of Napkin Deal Model with Shortcut AI
Build a Back of Napkin Deal Model with Shortcut AI
Jake Heller
See how Shortcut AI, an Excel super agent, can build a professional back-of-the-envelope BOV analysis workbook for multifamily investment opportunities in Los Angeles in just minutes. This tool levels the playing field, making complex underwriting accessible to everyone.
Build a Back of Napkin Deal Model with Shortcut AI
Jake Heller
See how Shortcut AI, an Excel super agent, can build a professional back-of-the-envelope BOV analysis workbook for multifamily investment opportunities in Los Angeles in just minutes. This tool levels the playing field, making complex underwriting accessible to everyone.
Chapters
00:00
Introduction
Introduction to using AI for real estate underwriting and deal analysis.
00:53
Configuring the AI Agent
03:04
Building the Model
06:38
Reviewing Results
08:00
Final Thoughts
Transcript
00:00
We're gonna build a quick back of the napkin model for a Los Angeles real estate broker to quickly analyze deals.
00:10
Now, I'm gonna go over to my AI girlfriend Claude, where I had already created a prompt for this.
00:17
We're gonna plug it in and we're gonna watch Shortcut go to work.
00:21
Shortcut AI, which is an Excel agent that can effectively do anything within Excel.
00:28
We could build pro formas for any asset class.
00:31
We could build comparison models, any type of models, or Excel work we could do utilizing shortcut.
00:38
We could extract P, Ls and OMs and rent rolls and essentially whatever we want.
00:44
Now I, I just have the free version of shortcut AI here, so you get 250 daily credits.
00:50
But let's put it to the test here.
00:53
We want to create a professional dynamic back of the envelope BOV analysis Excel workbook for multifamily divestment investment opportunities in la.
01:03
we gave it a bunch of criteria as well.
01:05
And so we're going to watch Shortcut go to work.
01:08
I will hit send there and I'll just talk you through what is happening here.
01:18
Essentially it's, it's going to understand the task first and then what it'll do is create a to do list, for us here, which we could easily just keep track of.
01:31
Shortcut is 1 of 2 Excel super agents that I use.
01:38
Hang on, it's just asked us some qualifying questions here.
01:41
Which, which it's great.
01:43
for the unit mix input, how would you prefer to structure it?
01:46
simple table with unit type rows, detail table.
01:49
Either is fine.
01:50
That's nice that you could just select a button versus typing in for operating expenses.
01:55
Would you like to input both options single total line?
01:57
Yeah, the more the better.
01:58
Right?
01:59
What default hold period should the model use?
02:02
Let's use five years for the sensitivity analysis.
02:05
Which variables are most important?
02:07
Let's do all of the above.
02:09
And then what's nice, it's, it's asking any specific branding colors you'd like me to use.
02:14
So maybe you could put in your, your logo, your brand colors, all that good stuff.
02:20
Waiting for clarification here.
02:21
We just gave it.
02:22
It it's good.
02:23
The more of an understanding that it has this, the, the better the model will be.
02:30
obviously we want to build it specific to our needs.
02:34
so let's just watch it go to work.
02:36
I think it's only going to use 50 credits for all this, which is nice.
02:39
So that means we could do five similar type tasks for free per day, which to me feels like quite a bit.
02:48
So that's great.
02:49
Shortcut AI also has a prompt library.
02:51
So I think what's so interesting about these tools is it levels the playing field.
02:56
What I mean by that is that the barriers to entry for underwriting for all this, are, are so much less.
03:04
Watch it go to work here.
03:05
I mean it's already built these tabs.
03:08
Boom, boom, boom.
03:09
we're keeping it simple in this particular case, but it looks like it's going to start with the input, assumptions and then it gave us 10 tasks that we're going to do here.
03:18
So once all that's done, we'll be done here.
03:20
but going back to what I was saying, if I could find my train of thought, yeah, so you could effectively delegate a lot of the underwriting tasks to folks that are maybe not as qualified as an analyst, but now have the tools and resources to be able to do a lot of this.
03:38
What I mean is that if you have a va, an executive assistant, anybody for that matter, maybe they don't even have a real estate background, you could have them underwrite deals or at least take a first pass at underwriting deals with this software, which is so great.
03:51
What I also like about the software is that you're not, you're not abandoning your existing workflows and processes.
04:03
Right.
04:03
You're not learning another interface.
04:05
This is all in Excel, which we have known to come, have come to know and love.
04:12
78 contact context is the AI's memory of your conversation.
04:16
When it runs low, agents responses can degrade.
04:19
Start a new chat to continue with the fresh contact.
04:22
That's all fine.
04:23
review changes.
04:27
So it's working here, it's working hard.
04:30
How beautiful that is already like right out the gate.
04:32
So awesome.
04:34
And you can see all the work that it's doing, all the different cells it's building.
04:38
Input sheet looks thorough and clear.
04:40
Moving to detailed pro forma.
04:43
So now it's going to start building right here.
04:46
And it moved on to task two, as you can see.
04:52
And here we go.
04:54
Now these take a little bit of time to build, but relatively speaking, no time.
04:59
In other words, if I'm taking a crack at building this model myself, or even an analyst who spends all day in Excel, I assure you they're not done building this input tab or making this type of progress.
05:11
It's not possible.
05:12
I mean this is like PhD level capabilities in Excel.
05:16
In Excel It's a plugin as well.
05:17
So I mean, we're on their website here, but this plugs into your actual Excel.
05:22
If we were to open that up separately.
05:24
so again, those that have been following along for some time now know that I'm a firm believer that the tools that integrate seamlessly with our existing workflows are the tools that are winners.
05:36
This is a winner.
05:37
And watch, we have some of these division errors.
05:39
All that's going to get cleaned up.
05:40
It'll do.
05:41
All that for us might just be a zero division error.
05:44
Yeah, so anytime you divide by zero, you'll get that.
05:47
But, it's working.
05:49
So next it's going to calculate investment returns and pro forma analysis.
05:52
Cash on cash returns.
05:53
Bang, there it is.
05:54
I'm gonna stop playing around with it.
05:56
We'll come back once it's ready.
05:57
But just watch how beautiful this is gonna look when it's all said.
06:01
All right, so we're pretty much done here.
06:07
What it's actually doing is final verification.
06:09
So it's going through every single formula in cell just to make sure that it works and there's no errors, which is great.
06:16
You have the option of skipping that verification, if you'd like.
06:20
I wouldn't recommend it.
06:22
unless you wanted to go and battle test this yourself.
06:25
it didn't take that much time.
06:28
I'd say about 10 or so minutes total, maybe a little bit longer.
06:32
I don't know what the math is on however many seconds that is.
06:35
12 minutes, maybe, tops.
06:38
But look what it's built.
06:39
All right, so here's our executive summary.
06:42
Okay, and how do I move this off?
06:44
Let's push that off to the side.
06:46
Yeah, look at this.
06:47
All right, so it's built us charts, move a little slow here.
06:52
Color, coded, formatted perfectly.
06:54
Here's our input assumptions tab as well.
06:57
So this is where everything goes in as inputs.
07:00
Wonderful.
07:01
Pretty robust page here.
07:03
It's straightforward, easy to use, obviously.
07:07
blue being, you know, where do you input is my guess.
07:12
yep, so it seems.
07:14
Closing cost percentage, that's all fine.
07:16
So we fill all that out.
07:17
Here's the pro forma analysis here.
07:20
looks like it did 10 years.
07:22
We have a IRR calculation helper on the side here, whatever that is.
07:28
Got our cash flow, all of our key metrics.
07:30
Look how easy this is to read and understand.
07:33
exit analysis, investment return summary, additional metrics.
07:37
And then we have our sensitivity Analysis as well, all formatted beautifully here, but really, really just easy to build professionally done.
07:51
I mean this is something if you're going to do yourself would, would take a very, very long time to do.
07:58
much longer than, than we just saw here.
08:00
What's great too is, I mean you essentially do anything else while this is all happening too.
08:06
So again, AI I believe is here to multiply our capabilities, not replace anybody.
08:11
But instead of me spending the time building something like this, right.
08:14
I could have AI do it even better than I could and be jumping into something else.
08:19
Now if you're a broker and you know, you need to build something like this, like go make cold calls while this is happening, come back, check it out.
08:28
Right, it appears this is done.
08:30
No, it's still verifying anyways really thorough verification process which is great but really, really great use case.
08:37
Now this was just multifamily.
08:39
Let's, let's assume, you know, there's other asset classes which maybe you don't have as much of an understanding for.
08:45
Industrial retail, whatever it might be, we could build these models so easily.
08:51
It levels the playing field, lower barrier to entry.
08:55
this is a game changer whether we like it or not.
08:58
So, and, and for underwriting, you know, the existing deals, this makes it a lot easier.
09:03
Again you could delegate to an assistant effectively.
09:06
They could underwrite this stuff, add the OM into here, bang bang, have it fill out all of it out.
09:12
so the use cases are endless.
09:13
This is a really exciting tool to have in the arsenal.
09:18
Check it out, play around with it.
09:20
You need to be using it even if you just want to use it to extract rent, roll and P and L data.
09:25
This is a no brainer.
09:26
So really excited about this