Integrated BIM tools, including Revit, AutoCAD, and Civil 3D
Professional CAD/CAM tools built on Inventor and AutoCAD
Any referenced datasets can be downloaded from "Module downloads" in the module overview.
Transcript
00:01
Hello, everyone. This is Prab Poreddy. I'm senior group manager in Autodesk Construction Solutions.
00:07
And today, I'd like to talk about how ACS uses machine learning to solve some of the critical problems in construction industry,
00:16
and demonstrate that with a use case that we've actually solved.
00:20
And also, show how customers are able to derive value from it with clear metrics.
00:28
let's dive in.
00:30
In this session, you can expect to see clearer explanation of a particular problem,
00:38
that we have selected to solve using machine learning technology.
00:43
And particularly, this is implemented in a suite of products, called Pype products.
00:50
Quick history here, Pype is a company that risk has acquired in 2020.
00:55
And that is how myself and my entire team has joined Autodesk family.
00:59
And we are so thrilled to be a part of this amazing family here.
01:04
At the end of this session,
01:06
you should be able to walk away with a good understanding
01:10
of how machine learning is being used to facilitate communication between various teams within this construction lifecycle,
01:17
and also, how that can encourage trust building within the teams.
01:25
So now let's get the context of construction industry in terms of its size.
01:34
It's a huge industry with almost 1.3 trillion spending each year. That's just within the US.
01:43
And so that's a huge industry with its own share of challenges.
01:49
Namely, there is a good amount of rework that happens, mostly because of quality deviations.
01:56
That is basically a term that means deviating from the original design intent.
02:02
And the other challenges are labor shortage and job site safety and data silos and poor communication,
02:11
between various teams, performing or working on any given project.
02:17
So, now let's focus on rework,
02:21
and how big the problem is and why does that happen and how we are able to actually address it using machine learning.
02:28
So, on an average, rework costs approximately 5% of project budget.
02:35
So, in a 1.3 trillion industry that would approximate to almost 50 billion.
02:41
That's 50 billion with a B. So, that's a lot of waste and rework happening there.
02:48
So, it demands that there be some mechanisms that I either prevented or reduce it.
02:55
Before we talk about the solutions, let's understand what causes this rework in the first place.
03:01
Any of these things can cause.
03:04
It could be changes or omissions or errors in the design phase,
03:09
architects or omissions or errors in the construction phase of the project,
03:15
or some issues in transportation or operability or the fabrication of the products.
03:22
So, you could ask,
03:26
isn't there any mechanism or process put in place in construction industry to address this situation.
03:32
And for that, my answer would be yes, there is a concept called submittals,
03:36
that is put in place as a quality control mechanism to prevent discrepancies from creeping into the projects.
03:44
The actual project that's causing rework or other impacts later on in the project life cycle.
03:50
So, in a nutshell submittals are documents that a contractor would submit to an architect for an approval.
03:59
And only after that approval, the actual construction activity happens. So that's in an actual what submittal is.
04:07
And the purpose of this submittal process is to identify potential issues,
04:12
that between designers and builders before they could occur on the job site.
04:17
As you know, it's always much cheaper to address an issue before it actually happens on the job site.
04:26
So, submittals and submittal workflows are very key for a successful and profitable construction job.
04:35
And what are the risk of missing the submittal?
04:38
As you can imagine, it could lead to costly reworks and costly rework means someone has to pay for the wastage.
04:48
And that's where teams tend to sort of look at each other and try to blame each other.
04:55
That's where the trust factor comes into play and they tend to lose trust.
04:59
And some of these scenarios could even lead to litigation.
05:04
So, we thought why not actually address this problem in a very holistic way,
05:09
and not just actually look at it from an operational standpoint.
05:14
And now you could be asking where does a I come into this picture where does it even belong,
05:21
I'm gonna explain right now.
05:23
So, we call this feature Pype AI.
05:27
Pype AI is a feature that uses machine learning algorithms,
05:32
to detect any submittals that are not called out by an architect.
05:38
But it would be deemed good practice for a general contractor and architect to evaluate,
05:46
whether such submittals need to be included to prevent costly reworks later.
05:52
So, when this feature is used early on in the construction projects, bidding phase or planning phases,
05:59
it would actually serve everyone involved in the project
06:05
to predict proactively what sort of issues might up if certain submittal is missed.
06:13
So, they can act to make an informed decision and share the cost or account for that cost in the contract.
06:19
So that's a very healthy conversation to have at an early stage of the project,
06:24
than actually waiting until it pans out to be an actual problem on the job site,
06:29
which is never a good time for anyone to be actually handling the situation.
06:35
So, in our product order, specifically the screenshots here illustrate how the users are able to use the feature.
06:45
So, on the left side is a screenshot that shows how subtitles are identified in a given project manual.
06:54
That's a pdf document and they are highlighted in green on the left side,
06:58
and after reviewing the submittals that an architect has explicitly called out for?
07:04
How can a contractor or in fact it could be an architect use this technology
07:09
to see what are the other recommendations that the Pype AI can make by leveraging
07:17
thousands and thousands of projects within the industry that are of similar type,
07:22
and see what are the best practices and what other submittals that recommendations might suggest.
07:28
And the engineers are going to be able to, project engineer,
07:32
general contractor may be able to take some of these recommendations to the OAC meeting and talk with an architect owner,
07:41
and come to a consensus on how to go about some of these important witnessing submittals.
07:47
And next I can show you how our users have actually gotten value already by this feature.
07:54
So, the percentages that you're seeing here are they reflect the various submittal types,
08:01
and how often they were recommended and how often they were actually included in the submittal log
08:09
for other projects that have been processed in our products.
08:12
So, as you can see warranty at the bottom is predominantly the submittal type that architects seem to have,
08:20
not always explicitly called out and the contractors were able to add that to the summit of log and add value.
08:28
And as you can imagine we can actually take this technology and apply it far beyond submittals.
08:35
We could even prevent some sacrifice or change orders being initiated and does help the construction industry in many ways.
08:43
We can even detect design changes with drawings from one version to another so that there are many other applications.
08:50
But for this session, I wanted to illustrate how we have solved one critical problem submittal so far,
08:56
and I hope this has been a useful session for you, informative session for you,
09:01
and thanks for your time and have a wonderful rest for the day.
09:03
Thank you.
Video transcript
00:01
Hello, everyone. This is Prab Poreddy. I'm senior group manager in Autodesk Construction Solutions.
00:07
And today, I'd like to talk about how ACS uses machine learning to solve some of the critical problems in construction industry,
00:16
and demonstrate that with a use case that we've actually solved.
00:20
And also, show how customers are able to derive value from it with clear metrics.
00:28
let's dive in.
00:30
In this session, you can expect to see clearer explanation of a particular problem,
00:38
that we have selected to solve using machine learning technology.
00:43
And particularly, this is implemented in a suite of products, called Pype products.
00:50
Quick history here, Pype is a company that risk has acquired in 2020.
00:55
And that is how myself and my entire team has joined Autodesk family.
00:59
And we are so thrilled to be a part of this amazing family here.
01:04
At the end of this session,
01:06
you should be able to walk away with a good understanding
01:10
of how machine learning is being used to facilitate communication between various teams within this construction lifecycle,
01:17
and also, how that can encourage trust building within the teams.
01:25
So now let's get the context of construction industry in terms of its size.
01:34
It's a huge industry with almost 1.3 trillion spending each year. That's just within the US.
01:43
And so that's a huge industry with its own share of challenges.
01:49
Namely, there is a good amount of rework that happens, mostly because of quality deviations.
01:56
That is basically a term that means deviating from the original design intent.
02:02
And the other challenges are labor shortage and job site safety and data silos and poor communication,
02:11
between various teams, performing or working on any given project.
02:17
So, now let's focus on rework,
02:21
and how big the problem is and why does that happen and how we are able to actually address it using machine learning.
02:28
So, on an average, rework costs approximately 5% of project budget.
02:35
So, in a 1.3 trillion industry that would approximate to almost 50 billion.
02:41
That's 50 billion with a B. So, that's a lot of waste and rework happening there.
02:48
So, it demands that there be some mechanisms that I either prevented or reduce it.
02:55
Before we talk about the solutions, let's understand what causes this rework in the first place.
03:01
Any of these things can cause.
03:04
It could be changes or omissions or errors in the design phase,
03:09
architects or omissions or errors in the construction phase of the project,
03:15
or some issues in transportation or operability or the fabrication of the products.
03:22
So, you could ask,
03:26
isn't there any mechanism or process put in place in construction industry to address this situation.
03:32
And for that, my answer would be yes, there is a concept called submittals,
03:36
that is put in place as a quality control mechanism to prevent discrepancies from creeping into the projects.
03:44
The actual project that's causing rework or other impacts later on in the project life cycle.
03:50
So, in a nutshell submittals are documents that a contractor would submit to an architect for an approval.
03:59
And only after that approval, the actual construction activity happens. So that's in an actual what submittal is.
04:07
And the purpose of this submittal process is to identify potential issues,
04:12
that between designers and builders before they could occur on the job site.
04:17
As you know, it's always much cheaper to address an issue before it actually happens on the job site.
04:26
So, submittals and submittal workflows are very key for a successful and profitable construction job.
04:35
And what are the risk of missing the submittal?
04:38
As you can imagine, it could lead to costly reworks and costly rework means someone has to pay for the wastage.
04:48
And that's where teams tend to sort of look at each other and try to blame each other.
04:55
That's where the trust factor comes into play and they tend to lose trust.
04:59
And some of these scenarios could even lead to litigation.
05:04
So, we thought why not actually address this problem in a very holistic way,
05:09
and not just actually look at it from an operational standpoint.
05:14
And now you could be asking where does a I come into this picture where does it even belong,
05:21
I'm gonna explain right now.
05:23
So, we call this feature Pype AI.
05:27
Pype AI is a feature that uses machine learning algorithms,
05:32
to detect any submittals that are not called out by an architect.
05:38
But it would be deemed good practice for a general contractor and architect to evaluate,
05:46
whether such submittals need to be included to prevent costly reworks later.
05:52
So, when this feature is used early on in the construction projects, bidding phase or planning phases,
05:59
it would actually serve everyone involved in the project
06:05
to predict proactively what sort of issues might up if certain submittal is missed.
06:13
So, they can act to make an informed decision and share the cost or account for that cost in the contract.
06:19
So that's a very healthy conversation to have at an early stage of the project,
06:24
than actually waiting until it pans out to be an actual problem on the job site,
06:29
which is never a good time for anyone to be actually handling the situation.
06:35
So, in our product order, specifically the screenshots here illustrate how the users are able to use the feature.
06:45
So, on the left side is a screenshot that shows how subtitles are identified in a given project manual.
06:54
That's a pdf document and they are highlighted in green on the left side,
06:58
and after reviewing the submittals that an architect has explicitly called out for?
07:04
How can a contractor or in fact it could be an architect use this technology
07:09
to see what are the other recommendations that the Pype AI can make by leveraging
07:17
thousands and thousands of projects within the industry that are of similar type,
07:22
and see what are the best practices and what other submittals that recommendations might suggest.
07:28
And the engineers are going to be able to, project engineer,
07:32
general contractor may be able to take some of these recommendations to the OAC meeting and talk with an architect owner,
07:41
and come to a consensus on how to go about some of these important witnessing submittals.
07:47
And next I can show you how our users have actually gotten value already by this feature.
07:54
So, the percentages that you're seeing here are they reflect the various submittal types,
08:01
and how often they were recommended and how often they were actually included in the submittal log
08:09
for other projects that have been processed in our products.
08:12
So, as you can see warranty at the bottom is predominantly the submittal type that architects seem to have,
08:20
not always explicitly called out and the contractors were able to add that to the summit of log and add value.
08:28
And as you can imagine we can actually take this technology and apply it far beyond submittals.
08:35
We could even prevent some sacrifice or change orders being initiated and does help the construction industry in many ways.
08:43
We can even detect design changes with drawings from one version to another so that there are many other applications.
08:50
But for this session, I wanted to illustrate how we have solved one critical problem submittal so far,
08:56
and I hope this has been a useful session for you, informative session for you,
09:01
and thanks for your time and have a wonderful rest for the day.
09:03
Thank you.
How to buy
Privacy | Do not sell or share my personal information | Cookie preferences | Report noncompliance | Terms of use | Legal | © 2025 Autodesk Inc. All rights reserved
Sign in to start learning
Sign in for unlimited free access to all learning content.Save your progress
Take assessments
Receive personalized recommendations
May we collect and use your data?
Learn more about the Third Party Services we use and our Privacy Statement. May we collect and use your data to tailor your experience?
Explore the benefits of a customized experience by managing your privacy settings for this site or visit our Privacy Statement to learn more about your options.