Jake's workspace
Jake HellerJan 9, 2026

Underwrite Deals in Excel Using Endex AI

Underwrite Deals in Excel Using Endex AI

Jake Heller·Jan 9, 2026

Discover Index, an AI agent that integrates directly into Excel, revolutionizing real estate underwriting. This tool, backed by $14 million from OpenAI, automates complex analysis, verifies data, and streamlines the entire underwriting process, allowing users to analyze more deals efficiently.

Transcript

00:00
I got early access to Index, which is an AI agent that lives in Excel.
00:07
They raised $14 million from open AI, and it is unbelievable.
00:13
And I'm so fired up.
00:15
I got early access to this last night.
00:17
And this changes the game for underwriting.
00:21
without a doubt.
00:22
It is so brilliant.
00:24
It's like PhD level Excel analysis, you know, not purpose built for real estate by any means.
00:31
Purpose built for Excel.
00:33
And, obviously underwriting in Excel is a big part of real estate, so I'm going to show you guys a couple of different use cases.
00:39
There's infinite amounts here.
00:40
But this is an underwriting model, which I already have.
00:44
Let's just say it's your template.
00:45
Okay.
00:46
This is one I got offline.
00:48
there's a lot to it.
00:48
It's very thorough and comprehensive.
00:51
Okay, so, let's assume that we get a new deal.
00:55
All right, I'm gonna go.
00:56
I just downloaded this file, but I lost it.
00:58
So bear with me here.
00:59
I am gonna go in.
01:00
Yeah, okay, Bear with me, guys.
01:03
I need to find this omn.
01:06
wonderful.
01:06
All right, so we have this om, right?
01:09
Taylor sent it to us.
01:11
Actually, I don't know if Taylor sent it to us, but we got this om.
01:14
It's an apartment building, right?
01:15
19, 29, 16 units.
01:17
We're going to take this om.
01:18
we're going to throw it in our chat.
01:21
An index.
01:21
It's going to take a second to load.
01:23
Okay, but you, you communicate with this the same way that you would with Chat gbt.
01:27
And also it connects with your.
01:30
All of your company files.
01:32
It could search the web and all the user files.
01:35
In other words, the that I upload into this chat here.
01:39
Okay?
01:39
But what I'm going to do is I'm going to take that OM, and I'm going to say, please underwrite this deal for me using my existing model here.
01:46
Okay?
01:46
And it's going to go in, it's going to update all this.
01:49
And what's nice is you could ask it.
01:51
So before it actually does any editing, you could do some planning, which I think is pretty critical.
01:55
But, for the purpose of this demo, we're just going to go ahead and ask it.
01:58
I'm going to say, looks like it already loaded.
02:00
So I'm going to say, please see attached offering memorandum.
02:05
Memorandum.
02:07
Okay, I want to underwrite this deal using my real estate acquisition model here.
02:20
The plan is to execute a value add strategy.
02:27
Ask me any questions if you'd like or just get started and underwrite the deal.
02:37
Sometimes I just like to type.
02:39
So that was a bit too much.
02:40
But anyways, again we could just hit ask and it wouldn't edit anything.
02:43
Or we could edit and it would just go ahead and start doing it.
02:46
So, this agent's capable, right.
02:49
The same way that these other agents are capable of going onto the Internet.
02:53
So we could verify comps, it could do analysis online.
02:56
It could essentially do anything.
02:58
But it is brilliant.
03:01
So it's going to go, through.
03:02
It already reviewed our om, so now it's going to start populating the model with deal data.
03:07
Okay.
03:07
And now, I've only tried this a couple of times, so I need to get a lot better at it.
03:12
But OMs are built by brokers.
03:15
Brokers want to sell deals, so it's important.
03:17
And, and what's so cool, look at this.
03:19
So it creates a to do list here.
03:20
So it's going to review complete model structure and input sections.
03:24
You could follow along here.
03:25
So it's pro preparing property deals input.
03:28
Okay.
03:28
But it's going to go through all this now.
03:29
It's, it's thorough.
03:30
It's not the quickest.
03:31
Okay, but now while this is working, you could do other stuff.
03:34
That's what makes it so brilliant.
03:36
So talk about being able to underwrite 10 times the amount of deals that you normally could.
03:41
You certainly can with this.
03:43
And I'm thinking like big picture, the ability to just streamline and automate the whole process.
03:50
So not even going in and taking the OM and plugging it into here, but the OM automatically get plugged into here and then this effectively being your analyst, it really is like a junior analyst not replacing anybody's capabilities because they.
04:04
There needs to be a little bit of guidance.
04:06
Going back to what I was saying about the om, it's broker talk.
04:10
So like you want to verify all that information and do some research.
04:14
So you'll want to guide the agent on all of that as well.
04:16
Right.
04:17
Like hey, look at their assumptions.
04:19
Is that market, is it not?
04:20
If not, go online.
04:22
Let's verify this for Los Angeles type deal, whatever it might be.
04:25
But it's editing ranges so you'll see.
04:28
This is a deal in Los Angeles.
04:30
It already plugged in the address here.
04:31
Look at this monster.
04:32
It's insane.
04:33
It already updated the amount of units 16.
04:36
So it is underwriting this deal for us at the moment.
04:39
That's exactly what it's doing.
04:40
It's entering operating expenses now.
04:42
So brilliant.
04:43
It's going in here, it's making changes, entered gross buildable area and year built into the property.
04:49
How insane is that?
04:50
Look, look at my ir irr changing.
04:53
It's so brilliant.
04:54
This is like when you talk about AI and all this, like this is exactly the stuff that's been missing.
05:01
This is like we have our model, we don't want to change it.
05:04
Let's assume we're an institutional shop.
05:05
We've been building and iterating our model for 50 years.
05:08
We're not just going to abandon it.
05:10
We're not going to log onto another interface onboard another tool.
05:13
This integrates so seamlessly with our process already.
05:18
And so that's why like something like this works.
05:21
It's, it's not, it's not cheap.
05:23
Okay, it's, it's not cheap.
05:24
Set vacancy assumptions.
05:26
I think it's 300 bucks a month by the way.
05:28
Yeah.
05:28
So look, you'll see like we just have huge return numbers because it's probably just going to take all the broker data.
05:34
So what I would ask it afterwards is like, let's figure out what the actual numbers are.
05:39
property taxes, 1.25 mill rate for la cap reserves 200 bucks a unit.
05:45
Like these numbers all kind of feel right.
05:47
120 divided by 370.
05:49
I'm not sure what that percentage is, but yeah, look, I mean if you're just going to take what the broker 90,000 cash flow.
05:58
If you're gonna just take what the broker gives you.
06:00
Yeah.
06:00
Then any deal makes sense.
06:01
So we just need to figure out.
06:06
Yeah.
06:06
I mean look, the expenses are low.
06:08
No doubt about it.
06:09
Those are broker expenses.
06:11
It looks like it's.
06:11
Is it still working?
06:13
No.
06:14
So it's completed.
06:15
So I've successfully underwritten the deal.
06:18
It basically just took all that information, it plugged it in there.
06:21
Year one.
06:22
This, this would be awesome, wouldn't it?
06:24
Value add strategy.
06:27
so it didn't, so it didn't add any renovation costs.
06:32
So we're going to need to spend money on reno costs to execute a value add strategy.
06:44
Let's do some on what that would look like both interior and exterior.
06:56
Now again you get out what you put in.
06:58
I, the prompt in the beginning.
06:59
What I would probably do is use like chat, GBT or Claude to come up with a prompt, a really detailed one.
07:06
Like really?
07:07
Or I mean you could do it all in here actually I don't know what I was thinking.
07:09
You don't need to edit it.
07:10
You would ask it.
07:11
So you plan all this out, right?
07:12
Like hey, I'm trying to execute a Value add strategy.
07:15
It's in L.
07:15
A, you know, five year hold.
07:17
Here's the types of returns I'm looking for.
07:19
You could say what purchase price would give me this type of return.
07:23
so look, we asked it analyzing renovation cost impacts.
07:26
So it's estimating renovation cost.
07:28
I'm calculating typical interior and exterior renovation costs.
07:32
Appropriate for a 1929 built multifamily property in South L.
07:36
A.
07:37
Insane.
07:38
So watch.
07:38
soon it'll populate all year.
07:39
But it's going online and it's researching all this for us, which is so unbelievably brilliant.
07:47
This, it could do anything.
07:48
It can go online.
07:49
It could look for comps.
07:51
It could then take those comps and plug it into our model.
07:54
We could build pro formas from scratch.
07:56
One of the use cases I did previously was build me a institutional style development model.
08:03
And it built like 13 tabs.
08:05
It looked a little bit, it took a little bit of time, but it built like 13 tabs.
08:09
so many different formulas.
08:11
It was so awesome.
08:12
I couldn't believe it.
08:13
Using market data and rent up.
08:15
Look at this.
08:15
Look.
08:15
It just updated everything.
08:16
How insane.
08:17
Literally and then it adds a note too.
08:20
I need to move this over a little bit.
08:22
So it adds a note.
08:22
22,000 a unit.
08:24
Look at this.
08:24
How insane is that?
08:25
It actually adds a note.
08:26
22,000 unit times 16 units based on 6k per 100amonth rent increase.
08:32
And then it has a source and everything too.
08:34
So awesome.
08:35
So it's saying, 31,000 per unit.
08:38
cool.
08:39
And 6.25.
08:41
Okay, fine.
08:43
The, these rents are, are too fat.
08:45
Maybe they're using section eight rents.
08:48
So let's say I think the OMN, used section eight rents.
08:55
Like, we're just underwriting.
08:56
We're not even leaving the platform.
08:57
Can you find me some comps for market rate rent in this area please?
09:09
And then let's average and plug into post Renault.
09:17
Right, so they have a 700 square foot two by two.
09:20
2451.
09:22
I don't see how you're getting.
09:24
You're not getting four bucks a foot on a 1929 building in South LA.
09:29
The fact that they're already at 350 something's interesting.
09:32
So I, I would want to look at the OM here.
09:34
But so it's still analyzing the rent impact.
09:37
So it's doing its to do list, verifying impact on returns.
09:42
It looks like it just wrapped up there.
09:44
The expenses are low.
09:45
It's so it.
09:46
Yeah.
09:46
So again, this is om math.
09:48
So it's just pulling direct from the om.
09:49
So it's 32% expenses.
09:52
We're going to update that as well.
09:53
But let's dig into this rent first.
09:55
So we just said, can you find me, some comps for market rate ran in this area, please?
10:00
So gathering market rent comps.
10:03
I'm so fired up about this.
10:05
Like, you don't need to know anything about Excel and you could just, It's magic.
10:11
So here's our little to do list.
10:12
Research market rate, apartment comps in South LA 9044 area code.
10:18
Calculate average market rents by unit type.
10:21
So it's compiling.
10:22
It's compiling and averaging market rents.
10:24
We'll minimize that.
10:25
I'd rather see all the good stuff.
10:27
But what you'll see.
10:28
Yeah.
10:28
So it looks like it just updated our rents.
10:30
Yeah, look at that.
10:32
So based on the rent comps, it found these are more kind of in line.
10:35
So.
10:36
Yeah.
10:36
What did it do?
10:37
Adds notes.
10:37
It averaged a bunch of comps for us.
10:39
And it's saying that these are where the comps are, which is wonderful.
10:44
Okay, so now we're coming back to reality.
10:47
All right.
10:49
also the om.
10:52
so as part of our initial prompt, we can avoid a lot of this by like the initial prompt.
10:56
So in other words, we could say like, see attached OM for reference.
10:59
Like let's do our homework online to validate the assumptions and underwrite this, you know, based on our criteria.
11:08
Or like real assumptions, just again, because broker math is broker math.
11:12
So.
11:13
But the next thing I'm just.
11:14
We're piecing this out here.
11:15
But the next thing is, can we validate our operating expenses, please?
11:23
They feel a bit light.
11:27
So now we're going to jump into that.
11:28
So go on our to do list.
11:30
It's going to review all operating expense line items versus OM and industry standards based on the om.
11:36
I'm wondering like what it says.
11:37
Just reviewing the OM here.
11:39
what was the pricing part of a portfolio?
11:43
Yeah, so 27.
11:44
Yeah.
11:44
So it's saying it's a 9% cap rate going in, which is nuts.
11:49
That's why you never trust broker math.
11:50
Like, look, I mean, this is the typical, you know, section eight.
11:54
We're getting these crazy rents and then very low operating expenses.
11:57
Very high cap rate deal on a 1920s building.
12:00
Like not today.
12:01
Not today.
12:04
Cool.
12:05
So it looks like it updated, it's closer now it's 36%.
12:08
I still feel like that's pretty crazy low for a 1920s building.
12:14
I'd probably bring it closer to like 4.
12:16
45, if not even 50%.
12:18
Yeah.
12:19
So it's saying going in 87 cap.
12:21
And you could like update manually as this is happening as well.
12:24
Which is awesome and realistic.
12:26
We're not selling a, a 1920s 1920s building at a 6 cap either right now.
12:31
I mean now we're at a 7 cap.
12:33
So now we're, we're a lot closer to reality here.
12:35
I, there's still, I mean I didn't check our growth assumptions there.
12:39
Six and a quarter interest rate.
12:41
Sure.
12:41
That maybe that's fine.
12:43
Doing a 75 LTC loan.
12:46
Interesting.
12:46
And it's cool.
12:47
It has these suggested follow ups as well.
12:50
Analyze how the updated Renault budget impacts the overall project irr.
12:55
Like there's really good suggested follow up questions.
12:59
let's just see.
13:00
But yeah, it's awesome.
13:01
I mean like what I just showed you doesn't even scratch the surface of the capabilities here.
13:06
Like build any model you want utilizing this.
13:10
Right.
13:10
I need to build an industrial model in 15 minutes.
13:14
Institutional grade.
13:15
Sick.
13:15
All the formulas fully dynamic.
13:17
Never leaving Excel.
13:18
Brilliant.
13:19
Okay, if I'm a broker and I just got a rent roll in a P and L and we plug it into our model, bang.
13:27
It's spitting out values like, no problem.
13:30
this is just unbelievable.
13:32
I think it's so awesome.
13:33
I mean you're effectively taking the entirety of the Internet, bringing it into Excel and then the AI does everything within Excel.
13:41
So like the next wave of analysts, guys is what I'm saying.
13:45
Like this isn't, this doesn't replace an analyst by any means, but the best analysts are going to be the ones who know how to direct this effectively, who have the best judgment and could understand this.
13:56
Right.
13:57
The best storytellers.
13:58
Here's the story.
13:59
Bang.
14:00
Right?
14:00
This, this is the next wave of analysts here in real estate underwriting, not changing anything, but integrating seamlessly into your exact process.
14:10
This is it.
14:11
Okay.
14:11
Four and a half.
14:12
Yeah, we're still a little fat.
14:13
Like these numbers are just.
14:15
You're not getting a two and a half times equity multiple on this deal.
14:18
And like look, to actually get tenants out of these old buildings in LA too, like you got to spend money.
14:24
Cash for keys.
14:25
31,000 a unit.
14:27
Probably low.
14:29
Realistically, like are we getting, are we getting a 75 LTC loan on this deal either that, I don't know how I missed that.
14:35
No.
14:36
Are we getting 75?
14:38
No.
14:39
So that was the big change, by the way.
14:41
Right.
14:41
And you could even say something like, based on this, what Purchase price do I need to buy this at to get a 20 IRR levered IRR?
14:58
I'm telling you the use cases are infinite.
15:01
It's going to underwrite this deal and figure out what purchase price you need to pay.
15:05
Watch to get that 20 levered.
15:07
Irr nuts.
15:09
So it's taken a stab at just different purchase prices to achieve that 20 IR.
15:13
But it'll get there.
15:15
This company is going to be so big.
15:16
I mean I'm so stoked that I got early access.
15:18
It's just amazing.
15:19
Like any type of deal you want to underwrite you could do so if like I didn't have an understanding I've never done retail, I've never done industrial.
15:27
Like I've never done self storage.
15:29
And, and look, I could learn, build the model, buy one online, figure it out and stuff.
15:34
But like this saves you all that time.
15:37
This allows you to underwrite any type of deal, any type of asset class, any market and do it efficiently.
15:44
This like makes the barrier to entry so much lower.
15:47
A lot lower.
15:48
I love our little to do list.
15:49
The fact like it creates these to to do lists, it's unbelievable.
15:52
unbelievable.
15:54
All right, it changed our initial assumption here, which is wrong.
15:59
I don't want that.
16:00
I want it, I want it there.
16:04
I want to leave the debt assumption as is and tell me what price I need for 20% IR Bang.
16:18
And you could also chat with it like while it's working, which is super cool guys.
16:22
The use cases are just so endless.
16:25
This is such a game changer.
16:27
I'm not even kidding.
16:28
I'm going to go through a series of, of different use cases for this.
16:32
and look, I mean you could, you could run this like you can create a new workbook, you could edit only in a specific range.
16:38
You could trace like different cells, all the precedents, you could trace all the depend.
16:45
Looks like there's a lot that's dependent on that particular cell.
16:49
So if you go here, maybe it's easier to see this one.
16:51
You can see history.
16:53
So beautiful.
16:53
beautiful.
16:54
Okay.
16:54
I mean you get the idea.
16:57
I gotta be a lot more specific and to be honest, I'm in a rush right now, so I'm kind of rushing through this.
17:01
I'm gonna get you guys a way, more thoughtful video, right?
17:06
Just so you can kind of see the inputs and stuff.
17:08
This, this tool's a no brainer.
17:09
Like it's, it's so unbelievable.
17:11
I could see why they raised 14 million from open AI going to be a home run.
17:16
nobody's leaving Excel.
17:17
Not yet.
17:17
That is right.
17:18
So, we are onto something with this.
17:20
I'm so excited to just continue to test new use cases and capabilities, with this thing.
17:26
But we're talking underwriting on steroids.
17:30
This thing's just a monster.
17:32
Such a beautiful thing.
17:33
So really excited.
17:35
I'll continue to share more videos as part of a series for you guys.
17:38
This changes the game of underwriting full AI agent in your Excel model.
17:43
Bang.
Underwrite Deals in Excel Using Endex AI — Tella