YT YouTube2Text

YouTube2Text → Transcripts → Endra Rethinks MEP Design with AI

Endra Rethinks MEP Design with AI — Transcript

by AEC TechTV · 4,927 words · 697 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:03[music]
0:08AEC [music] Business, the construction
0:10industries innovation and technology
0:12show.
0:14>> Hello and welcome to the AEC Business
0:17podcast. Today we'll discuss AI powered
0:20MEP design with Nicholas Linen,
0:24co-founder and CEO of Endra AI, a
0:26Stockholm based startup that just raised
0:29$50 million in series A funding. And I'm
0:33your host, Arie Hastan. Nicholas, I
0:37understand you're extremely busy, so I'm
0:39really happy that you could join us.
0:41>> Thank you, Arie. Thank you for having
0:43me. I I think you've been uh giving uh
0:46many interviews recently.
0:48>> Yeah. There's been a hectic schedule the
0:50last [laughter] couple weeks.
0:51>> Yeah. Yeah. You you have a background in
0:54technical security solutions uh with
0:56Sector. Uh what made you start up uh an
1:02MEP design platform company? Was it uh
1:05specific specific pain point or was it
1:08just that you saw that there's a huge
1:10market in
1:12>> um Yeah. I think there is a there's a
1:14mix. So my segue into this industry is
1:16that I started as you said a company in
1:18in in the security system space. So we
1:21were an design build integrator for
1:23large scale enterprise uh security um
1:26security systems. Uh so we were part of
1:29building out the uh data center buildout
1:32early back in 20 from 2014 to 2022.
1:36uh and um that is sort of my segue
1:38because we were designing our own uh we
1:40were designing our own solutions and had
1:42an in-house design team worked in sort
1:45of we that that was working in Revit and
1:47and all these uh in in these sort of
1:50design and engineering uh parts of the
1:52workflow. So that is sort of my segue
1:55into into this space. Um and and yeah,
1:59we were kind of excited about it because
2:01it's it's it it given that we were a
2:04design build contractor or an integrator
2:07for these systems, we our business
2:08wasn't designing solutions. Uh we were
2:12we wanted to design the one the ones
2:14that would later install and then we
2:15wanted to install them as fast as
2:16possible in order to send our last
2:18invoice to to get our money and get on
2:20to the next project. Um and that's sort
2:23of the nature of that business and and
2:25once we saw the bottleneck inside design
2:28workflows uh and also tried to procure
2:32the the the the best plugins on the
2:35market the best sort of point solutions
2:36for each and every task. Uh we we over
2:39time understood that this is um this is
2:42this is an industry needs sort of um
2:44more of a platform type of grip um to uh
2:48to to to get to work. So uh we started
2:51and as well with the mindset of doing
2:53this for security systems then uh at
2:56some point in time my dad actually who's
2:59an architect said that if you want to
3:00have real impact on how buildings are
3:02being designed you should should
3:03probably not just look at the uh
3:06security systems um part of things you
3:08should you should you should should sort
3:10of consider doing the full MEP. So uh
3:13from very early on we've had a full MEP
3:15scope um and uh and then yeah that is
3:19that is how we're sort of segueing
3:21ourselves into whatever we do today.
3:24Well, we have had um um MEP software
3:30for for many decades already, but why
3:34has it been a challenge for developers
3:37uh to to actually make it better than
3:41than what we currently have? Is it
3:44something different from architectural
3:46design or or structural design that MEP
3:49has not been in the vanguard of of of
3:52maybe of BIM?
3:54Um yeah it's that's um that's a good
3:58question on why and I think the best
4:00sort of to answer that is is the one who
4:03sort of has been in involved in that
4:04space for many years but haven't done
4:06it. Um but um the core issue that we see
4:11today is that the underlying data models
4:14in in in uh in in let's say Revit is is
4:18just not granular enough. you're not
4:20able to build the level of automation,
4:22the level of um the level of um hard to
4:26solve the really hard tasks inside that
4:29um inside that environment. So the first
4:31thing that we had to do was to build our
4:34own platform outside Revit. So we had to
4:36start building our own 3D library and 3D
4:39format a 2D format and 2D library uh in
4:42order to get this very granular
4:44understanding of what a building is for
4:46a computer. So you need to like the
4:48computer need to understand what um what
4:51what what a building is what a room is
4:53and what what what a f what what a
4:55furniture type is. Uh and once you've
4:57like done that to a very deep level to a
5:00very deep as well like a geometric deep
5:04level um you're then able to start
5:06building these type of agents and um
5:08these type of um sort of very fast
5:11design approaches on top of that. And if
5:14you have but if you have the core if the
5:15core foundation is not where it should
5:17be like it is currently in the the
5:20software is being used it's um it's very
5:23hard to to to to do something
5:25meaningful. Uh so I think to answer your
5:28question I think there's been sort of a
5:31there's probably a mix between risk it's
5:34very expensive uh you have to invest a
5:36lot of money and really have an
5:38ambitious approach on on on building a
5:40platform where you can bundle a lot of
5:42these point solutions onto one single
5:44source data model that that that manages
5:46all relations to how MP systems work
5:50with each other not only sort of um HSC
5:53mechanic or HSC electrical and and and
5:55and plumbing or public health you need
5:57to one single source data model where
5:59that understands how these are relative
6:01to each other. Um so combination of that
6:05it's very hard it's been it's it's it's
6:07very expensive to build it's um it's
6:09it's um you have to take on a lot of
6:12risk and and that there are so many
6:13point solutions and um I think there's a
6:16combination of that that has that that
6:18that that sort of stopped meaningful
6:22innovation in this space.
6:24Yeah. Yeah. I I I I agree. Um so you
6:29chose AI as the technology uh to tackle
6:32these issues. What does AI enable in MEP
6:37design space that couldn't we couldn't
6:39do with traditional CAD beam software
6:42previously? Is it is it uh as you said
6:46the the management of the totality
6:50uh and or is it speed or what is the new
6:54edge so to speak?
6:56>> The new edge um would is definitely that
6:59you can structure. So if you build a
7:01data model that that we we are doing
7:03like a separate data model from from
7:05Revit or from whatever uh authoring
7:07software we're being used you can build
7:10a custom data model and then you can use
7:12LLMs in order to ingest the data from
7:14the legacy ones and structure that into
7:17exactly this data structure that that's
7:19needed inside Android. So that's the
7:21first thing which which sound like a
7:23small thing but it's like the core
7:25foundation of if you want that type of
7:28solution to work. uh and it's very
7:30important to get to to to structure it
7:32in the right way. So that's definitely
7:34one thing. Uh we work in what we call
7:36spatial AI as well, meaning that we're
7:38trying to solve spatial problems, which
7:41is a very hard uh thing. So let's say if
7:44you route um a conduit or uh HVAC duct
7:48through a building uh how do you do that
7:51without getting clashes without without
7:54sort of clashing with structural objects
7:56clashing with um w with other
7:59installations inside this building and
8:01um uh and that's a very hard pro problem
8:04to solve. um do you need to add
8:07reasoning processes? You need to add
8:09special deep special understanding of of
8:11of of spaces inside this room and um
8:15those are the type of things where you
8:17can you can you can uh you can start
8:19doing now um not on the sort of AI side
8:23but more on the compute side maybe as
8:26well. We're a cloud platform. So we
8:28leverage um like big amounts of compute
8:32and uh uh what we can what we how endra
8:37works is that you if you're looking at
8:39an electrical system to a building and
8:41model everything from a conductor level
8:43to the electrical board panel to the
8:47let's say the receptacle or the outlet
8:49outside in the building and and is is
8:52circuiting everything and once
8:54everything is done inside the den model
8:56where you have the outlet the receptacle
8:58inside let's say a meeting room. Uh you
9:00have the cable that goes to that one.
9:02That cable goes down to electrical panel
9:05that goes to a breaker and that goes to
9:07a transformer. And once you can control
9:10everything in that environment and
9:11compute everything, it becomes
9:13incredibly incredibly powerful. Um so um
9:18and and and and that's not a that's not
9:20a sort of an AI
9:22feature but what it is is that we can we
9:25can compute the most optimal solutions
9:27uh for our users uh having everything
9:29inside one data model uh which which is
9:32not possible today in Revit or it would
9:34take you a lot of lot of time to do it
9:36in inside Revit. So um those are the
9:40types of things that we're currently
9:42working on. Andreas not only an AI
9:45platform, we we we we leverage physics
9:47based modeling. We leverage simulation.
9:50Um we leverage AI a whole lot. Um and
9:54but and and as well we're we're doing
9:56this inside one single source data
9:58model. And that is sort of the product
10:00strategy that we're we're having moving
10:02forward.
10:03>> Okay. Yeah.
10:05Um you and many other developers who
10:10work with BIM always talk about Revit
10:13and it's a legacy system and we still
10:15use it. Uh but do you see maybe a future
10:20where
10:22it might be challenged by AI based
10:25systems?
10:28Um
10:30I I I would say my sort of my first
10:32answer would be no. um it may be
10:36challenging the way where users doesn't
10:39spend as much time inside Revit as they
10:41do today. So they may spend uh some of
10:44their time in in in other in other
10:46software platforms like Kendra. So we
10:49want to be a workplace where for MEP
10:52engineering organizations that means
10:54naturally that that we want to move time
10:57that they spend inside Revit to in
11:00spending that same time inside Endra.
11:03Um, but it doesn't mean that we're
11:05building a new Revit. Uh, we we're very
11:07much um we we we very much rely on on
11:11the user having a Revit license. Uh,
11:15technically the user can run a project
11:16inside Android without a Revit Revit
11:18license. So, we support IFC as input and
11:22as well as output. So, you can run it
11:23without wi- without without Revit. But,
11:26um, almost nine out of 10 users I think
11:29at least in the western markets are are
11:31using Revit. Um, so it's um um
11:36yeah, I I
11:39I'm not sure if if it'll be I I don't
11:42think there's a reality where Revit is
11:44is sort of replaced. Uh I think it'll
11:47definitely stay um sort of at the at the
11:50point that that it is now and it'll like
11:53new companies coming up and and and and
11:57in the space will will work with Revit
11:59and and integrate towards Revit in order
12:01to um to make their their um their core
12:06products work. So I think it'll be like
12:08an orchestration layer where Revit is is
12:10is sort of a center point and and and
12:13that's also I think that aligns quite
12:15well with with Autodesk's uh strategy as
12:18well. So um yeah.
12:21>> Yeah. So uh based on what you said I
12:25understand that
12:27your platform can do a lot of things
12:30that had be had been traditionally done
12:34almost manually by designers. So what is
12:37this the role and workflow for designers
12:40on your platform?
12:43>> The role for designers on our platform
12:46is that we build the whole workflow as a
12:48human in the loop. Uh so the um the the
12:53the users are always in control at all
12:56times and we don't want to build any
12:57blackbox answers for for them. So the
13:00engineers play a crucial role in in
13:03understanding what Android does. uh
13:05valuing that and uh and and editing if
13:08if needed and everything is editable in
13:10inside Android. There's a full user
13:11interface where the user can do whatever
13:14he or she wants to do. Um so the
13:17engineer play a crucial role in that. I
13:19think engineers over time will do more
13:21with less. Um I think I think they will
13:25um they will be able to start offer new
13:27products to their customers. Um but they
13:30will be very central inside this um
13:33inside the whole um design process and
13:35engineering process and uh um they're
13:39they're yeah they're just going to do be
13:41able to do more or less spend time on on
13:44on on more high value task. I still
13:47haven't met uh an electrical engineer
13:49that loves to um place 2,000 receptacles
13:54or outlets inside a 3D model. Um so it's
13:58it's it's a lot of things that they
14:00don't like naturally don't love to do um
14:04that we could sort of help them remove
14:06so they can actually to focus on on uh
14:09on really hard task and like designing a
14:11building is one of the most iterative
14:12things that you can probably do in the
14:14world like just look at whatever other
14:17industries and and designing a building
14:19like objectively looks very hard many
14:23stakeholders there so many different
14:25consultancies that has to coordinate
14:27under a short period of time. Uh having
14:30someone to tell them to work even faster
14:32and and it's it's it's like a it's like
14:34a hard thing to do. So in order for us
14:36to help them get rid of some of the
14:39manual and tedious tasks, I think it's
14:41um that that will be incredibly val
14:43valuable to them over time. Um so
14:47>> so at at what phase of the of the design
14:52or construction process
14:54do you does your platform uh become
14:58useful? Is it already at the let's say
15:02conceptual or layout basic layout phase
15:05or is it more towards the detailed
15:07design? Um, so we're we're we're um
15:12let's let's let's let's um reference
15:14towards let's say Arieba stage uh in the
15:17UK. I'm not sure where your listeners
15:19are, but let's say they're we're working
15:21towards that we're talking about Aribba
15:24uh Aribba scale. Um
15:27we want to we were focusing to be we
15:29want we want engineers to set the
15:31strategy. So they want to be in the
15:33early stage and set the strategy for the
15:35building and the design intent. and that
15:37we want the engineers to work with Endra
15:39as they go to to a more deta design. So
15:42we want to support up until in the UK a
15:44reva stage four design uh and uh and and
15:48and and really be the sort of right
15:52before it's being issued for
15:53construction. Uh that is sort of where
15:55we um we um we want to sit and we work
15:59with mostly engineering consultancies uh
16:02some of the largest consultancies on the
16:04globe and um and and and they usually
16:08they tend to sometimes they go and do
16:11like the actual like the the
16:12construction documents but uh um in many
16:16many cases they they work uh up until
16:18permit and and and that's equivalent uh
16:21to rep stage
16:24Okay. So, so contractors, MEP
16:27contractors might be not be using your
16:31platform.
16:32>> Um, not right now. uh we will like we we
16:36haven't sort of set out that strategy
16:38yet but currently we work mostly with
16:40MEP uh engineering consultancies and uh
16:44um it's both um we have a lot of
16:47interest from contractors but we have as
16:51of sort of right now and this changes
16:54quite rapidly but as of right now we're
16:56we're working with um with engineering
16:58consultancies.
17:00>> Okay. Um on your website you're
17:04reporting 70 times efficiency gains
17:08uh and cutting design time from months
17:10to hours.
17:12Are these anecdotal results or do you
17:15see that these could be scaled uh
17:18scalable across all projects and
17:20customers in the future?
17:22>> Um so it it obviously depends on on the
17:26project uh type. Um, but I think
17:29headline here is that it's it's it's
17:31it's it goes in like dramatically faster
17:34versus doing it yourself. And um um and
17:38and and and if that could be scaled,
17:40yes, definitely. Um we in all the deals
17:44that we do, we uh invite the customer to
17:47do what we call a customer validation
17:49project. So or C multiple customer
17:51validation projects and that's when the
17:53customer pilots and on their actual
17:56projects that they have uh that they
17:58have done themselves uh previously. So
18:01they have all the data. They know how
18:02long time it took. They know what good
18:04looks like and they can compare that to
18:06whatever end produced and still which is
18:09uh really killer metric but we're still
18:13100% from everyone we have invited in
18:15order to do this customer validation
18:17project. We have still converted 100% of
18:20of the customers uh from that process to
18:23assign did. So it um and if you go on
18:27our website you can you can you can see
18:29at least some of the logos that we're
18:31working with and these are as I said
18:33this is some of the largest consult
18:35engineering consultancies on the planet
18:37and um um so it's it's it's it's um it
18:43it I think that sort of is the best
18:46answer to that um that the ones who have
18:49tested it have all um went through and
18:52actually purchased this as So
18:56>> um as we have discussed uh our business
19:00is very much project oriented. Every
19:02project is like a new new new new thing
19:06uh almost like a prototype. Uh do you
19:10>> Yeah. Yeah. So, so the question was that
19:14with your on your platform you can
19:18actually start building more like
19:20processes uh than than than projects and
19:23and and that's I think that's really
19:26valuable. Yeah. Um
19:29so if I understood correctly you started
19:32with fire alarm systems.
19:35Uh what's up next uh in your road map?
19:39Uh
19:40>> yeah, we we we we have been most public
19:43I think around around our fire alarm uh
19:45module. Uh having that said, the most
19:48amount of work has gone into uh our
19:50electrical module or power module. Um
19:54and um uh that is being released for
19:57wide access September 14th. Um we are
20:00doing we we're launching it in Las Vegas
20:03the day before Autodesk University. Uh
20:05we have rented this really cool venue.
20:08Uh, and the event is called AP Poke. So,
20:11anyone listening, you should go on our
20:13website, android.ai/epoke.
20:15Um, you find all the links in there, um,
20:19to to sign up. And, um, so we're we're
20:22releasing it in Vegas, uh, September
20:2414th, and we have some really cool
20:26speakers doing it together with us.
20:28someone like Gary Casper, we have Gary
20:31Kasparov uh and as well Amar Hanspal
20:35who's former CEO of Autodesk and we're
20:38announcing more and more speakers as we
20:40go now. So we're um it's going to be a
20:42cool night and it's it's like day zero
20:44for Autodesk University. So the day
20:46before Autodesk University um and
20:49September 14th is the launch of the
20:51electrical module. So, um um yeah, it's
20:56it's it's this autumn uh for wider
20:58access and then we're releasing it in in
21:01uh in in in um selected parts of Europe,
21:05I should say. Uh the UK and the US.
21:09>> Oh yeah, sounds exciting. Um
21:14um so um
21:19um so uh how do you see because now now
21:23you have a a really good insight view
21:26into how AI can improve our processes
21:30and in in in construction. How do you
21:33see the future overall in in
21:37construction? Because we have so many
21:40many challenges and one of the
21:43challenges is naturally that the as we
21:46said talked previously is the the
21:49collaboration and there are so many
21:51stakeholders with different agendas and
21:54different incentives and so on uh and
21:57the data is not really flowing between
22:01these participants.
22:03Do you see that AI could dramatically
22:08change the situation or is it a slow
22:10evolution as as as it has always been?
22:13It takes always 10 years to do something
22:16new in the construction technology. I
22:17think
22:20>> um well I think I think we're a
22:22testament that that it has that it
22:23doesn't need to take 10 years in order
22:25to do something. We've been doing it
22:27like we started Endra for like a year
22:30and a half ago. So um there has at least
22:33been some things done during this during
22:36this time. Um
22:39so is the question is around sort of if
22:41if if AI will will will help sort of get
22:45in like do innovation faster or whether
22:47like if data is that is that the overall
22:50question?
22:51>> Yeah. So so the question is that because
22:53we have so many participants
22:56>> which are not compatible in terms of
22:58data for example. Yeah.
23:00>> And and the processes are different. So
23:02do you see that with these capabilities
23:05that AI provides us, we could actually
23:08start collaborating better in the
23:12future?
23:13>> Yeah.
23:13>> And and and actually make some because
23:17it's it's still everybody's so siloed in
23:20this industry.
23:22>> So I yeah, I can speak specifically for
23:24MEP engineer u MEP engineering
23:27organization. I'm not very looped in
23:29like on the details of of of of the
23:32other trades even though we're we're
23:33interfacing with them. Um but for if you
23:36look at the the stack for MEP
23:38engineering organizations so there they
23:40have their their drafting software which
23:42is mainly Revit. There are obviously
23:44alternatives that but mainly Revit. Um
23:47then they they are using in either Excel
23:49or they use some point solutions to
23:50calculate things like cable sizes or um
23:54voltage drop and and and similar things.
23:57Uh and then they use um three three some
24:01sometimes at least um like third party
24:05um um software for coordinate spatial
24:08coordination and then where they sync
24:09their models and and check for clashes.
24:11So it's like three layers of of
24:14different st different tool stacks um
24:16which only the first one which is the
24:18authoring layer is the one that
24:20everyone's using and that's not like
24:22okay now it's it's possible to
24:24collaborate uh through ACC or pharma um
24:28but it's it's it's like it's it's three
24:30layers of of different tools uh none of
24:34which really speak to each other so yes
24:36obviously there's a there's there
24:38there's there there's an issue there Um,
24:42Endra bundles all of these three um
24:44without sort of with a small sort of
24:47flag that we're not building a new
24:49Revit, but we we're bundling all of
24:51these three into one one um one um one
24:56workflow for the user. And it's very
24:59important for us to to do that quite
25:02fast. uh because if we're not helping
25:04our users to to do everything inside
25:07that workflow, it doesn't meaningfully
25:10fast add value to them. So we need to
25:14bundle all of this that you should like
25:15end needs to calculate how big your
25:17cable should be. It needs to set your
25:20breaker rating like break size uh
25:22breaker sizing and and it needs to
25:24understand where where you're allowed to
25:27route conduits or or or or raceway
25:30systems or whatever. Um so it's it's
25:32like we need to do all these things into
25:36inside one platform. Um and and and I
25:40think that is what will move the needle
25:42for innovation in general here because
25:45another point solution like no no one
25:47need another point solution. I've never
25:49heard uh CTO or CDL ask us for another
25:52point solution or a tooling upgrade.
25:55Like they're they're more hungry to to
25:58like do a platform shift than try like
26:01go into this more AI native um uh
26:05workflow as as an organization and in
26:08order to do that because that is
26:09ultimately what we are doing. for
26:11helping them deliver on their AI
26:12strategy and their AI transformation
26:14strategy. And in order for us to do
26:17that, it has to be a platform. We cannot
26:19do this with a with with a tooling app,
26:22right? So it's um it sort of naturally
26:26data has to naturally flow inside one
26:28platform and and I think the problem for
26:31the industry as a as a as a whole is
26:34that there are so many point solutions
26:36and that's sort of that's the underlying
26:38issue for for data not flowing um
26:41flowing between every everything. So um
26:46>> yeah yeah that that's the the reality
26:49now. Yeah. Um,
26:52I have been podcasting and blogging for
26:5514 years now and I don't remember if I
26:59have maybe just a few Swedish
27:04construction technology startups uh as
27:06as guests or interviewees.
27:09Uh, how how is the startups seen in uh
27:14in construction technology in Sweden?
27:17Um, so for for construction technology
27:20specifically,
27:22um, I there are definitely there are
27:25there are some some some some good
27:28startups in in in Stockholm. Uh, BCanta
27:31is one. Uh, they work with with
27:33contractors uh, trying to find issues.
27:36Arrive is another um, and and and we we
27:40we do and we have another one called AI
27:42Bob as well. we have like we have a
27:44couple ones uh that's really are are in
27:48the space. Um and um more more sort of
27:53broadly the startup scene in in
27:55Stockholm is probably one of like right
27:57as of right now it's probably one of the
27:59best places to be in the world. uh you
28:01have access to talent um that doesn't
28:04necessarily compete with Google, Meta or
28:06or or um SpaceX or or anything like big
28:09tech companies um as as they do in
28:12Silicon Valley. Uh but but but so so
28:15Stockholm is a great is a great place to
28:17be. Uh we have access to capital, we
28:20have access to vast engineering
28:22networks. Um and and and just look at
28:26sort of the AI application layer at
28:28least
28:29per capita were we're probably by far
28:32the best in the world of of delivering
28:34these um these companies. You can you
28:36can have you can just look at lovable
28:38Legora Asana uh and and you know you
28:41have so many of these um from from a
28:44city that that's that's a population of
28:46of a million people. So, it's um it's
28:49it's it's it's
28:51an incredibly good place to be right
28:53now. Um we the the startup ecosystem is
28:56is kind of mature and um and uh and
29:00yeah, money money is flowing into
29:02Stockholm and um it's it's a very
29:04vibrant and cool place to be.
29:06>> Yeah. And you you are a perfect example
29:09of that. [laughter]
29:11>> Thank you. I hope so.
29:12>> Yeah. Yeah. So uh practical question for
29:16our listeners, how can they join in?
29:20>> Um to how they can become a customer or
29:22>> Yeah. Yeah. So they go on our website
29:25andra.ai
29:27uh and then they can fill out the
29:30request to access form and we'll we'll
29:32be in touch with you.
29:33>> Okay. But you already have like 6 600
29:36companies on your weight list.
29:39>> [laughter]
29:39>> Yeah, it's I think that number is like
29:42like many months old now. So, it's it's
29:45in the thousands. Um
29:46>> Okay.
29:47>> But we're we're um we're trying to sort
29:51of work our way through um through
29:54through all interest that's coming. But
29:56but yeah, we're currently very
29:58bottlenecked by both the amount of
30:01people that we are. Uh we're growing
30:03rapidly. So like we're trying to hire in
30:05the fastest p possible. We went from
30:07like 10 people to 36 or something the
30:11last three months. So, we're we're we're
30:13scaling very fast, but not not not fast
30:16enough. Um,
30:18>> but yeah, we're that that is the best
30:20way to do or or you guys can reach out
30:23on my LinkedIn, Nicholas Lingo, and I'll
30:26try to sort of put you in touch with the
30:28right right.
30:29>> Yeah. Well, I would have a a lot more
30:33questions, but maybe we we we'll just
30:36leave those for the next time hopefully.
30:38>> So, Nicholas, this has been incredibly
30:41insightful and thanks for sharing your
30:43perspective on where AI powered MEP
30:47design is heading. Um, so
30:50I thank you and hope to talk to you
30:53again soon.
30:54>> Yeah, likewise, Arie. Thank you for
30:56having me.
30:58[music]
30:58>> Thanks for listening. Subscribe to this
31:00podcast and visit aec-business.com,
31:04the award-winning blog, for more news
31:06and stories.

About this transcript

This page contains the full transcript of Endra Rethinks MEP Design with AI by AEC TechTV, generated from the public captions YouTube serves with the video. The transcript has 4,927 words across 697 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.