Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:04):
The other
thing that's been exciting to seeis kind of the convergence between
how things that are importantin construction
are working their way upstream into designso that, you know, we can eliminate
some of the errors in the wastethat happens throughout a normal process.
I hope five years from nowwe are talking about a world
where data is going from
(00:25):
format in concept all the way to tandemand a built environment
where you're using Bluetooth sensorsor whatever
to tell what traffic flow is really likeat this intersection.
AI is actuallyone of the great equalizers.
It doesn't matter the size of the firm.
I think everybody should lean into it,because there is an opportunity
(00:47):
for everybody to understand the influencesof AI in your specific, business.
Let's get right into it.
Can you startmaybe by telling our Viewpoints listeners
a little bit about you, your journeyand your role at Autodesk?
Sure.
So, I have been with Autodeskfor over 21 years.
(01:09):
I joined the company throughan acquisition on the manufacturing side
of the house, but then about 12 years ago,I came over to the AEC side,
and I've been working in AECfor the last 12 years.
And so it's been really exciting for meto be part of an industry that's going
through so much transformationand has so much opportunity to,
(01:29):
you know, really deliverbetter outcomes for, you know,
for you and your customersand really, really for the planet.
So today I'm responsible
for our all of our architecture,engineering and construction solutions.
And the other thing that's beenexciting to see is kind of the convergence
between how things that are importantin construction are working their way
upstream into design so that, you know,we can eliminate some of the errors
(01:53):
and the waste that happenedthroughout a normal process.
Thanks, Amy.
I know I'm looking forward to ut,I know there's a lot there,
especially in the construction side.
Definitely looking forward to that.
So, Ryan, maybe, for Viewpoint listeners,a little bit about your role of VHB
and what you're doing to help our clientsreally leverage some of the design tools
and specifically AI.
(02:14):
Sure. So, Amy, thanks for joining us.
Ryan Noyes, I've been with VHBfor almost 28 years now.
I am a recovering designer,started my time doing highway design,
and moved into our technologyenablement team.
And that move I now lead a teamthat is focused on
how we do digital delivery here at VHB.
(02:37):
So teams partnering with clientsas well as internal teams on preparing
for digital delivery in the digital futureand specifically around AI.
My team spends a lot of time focused on
how will our current
partners, yourself is one of them,how will we be presented solutions
(02:59):
that are focused around using AI toolsto make better decisions on our projects?
You know, really looking to
remove some of the fear aroundleveraging AI on projects and,
get people comfortable with that ideathat this can be
another assistant to haveas we move forward in design.
So, you know, Amy,it was really kind of interesting
(03:21):
that you started your,your journey, on the manufacturing side,
an industry that has always beena little bit ahead of us as far as moving
to models as the deliverable and moving
towards generative design.
I’m really curious,how do you see our industry transforming?
(03:44):
Are there parallelsthat you see from your previous roles?
Well, that's a great question.
You know, manufacturing was driven to 3D
parametric design out of necessity,you know, a lot earlier than AEC.
Part of that was, you know, manufacturing.
You're looking for volume.
And so everything is not a one-offkind of project.
(04:06):
And so, it was absolutely required.
And it's been really excitingto see how those principles
have been appliedto, you know, AEC over the last 20 years.
And for me,
the final, maybe not the final, frontierbut one of the exciting
frontiers is getting more manufacturingprinciples deployed on projects.
You know,I think our MEP coworkers have been doing
(04:28):
offsite fabrication for a very long time.
And we need more and more,you know, components in the building
to becomeor in an infrastructure project to be
manufactured and assembled offsiteand then brought onto the job site.
Right.
It has amazing impacts on cost,schedules, safety, sustainability.
(04:48):
So that's really an area where I seea lot of customers experimenting,
but not as much,you know, at scale deployment of that yet.
So I think that could be a placewhere AI helps us, for sure,
but is definitely something we needthe industry to be thinking more about.
Though it's definitely coming faster.
And I'm sitting there, you talk about AIand the onslaught of it.
(05:11):
You know, you're outrepresenting the entire industry across,
you know,
what are some of the things
that really keep them up at nightrelative to hearing about AI?
So there's this difference between hearingabout AI and really where it is today
versus the journey where it'sgoing to be over the next five, ten years.
Yeah.
You know, I think there's sort of a cyclethat people go through.
(05:33):
Especially in theI think the first question is, oh my gosh,
is it going to take my job, you know,and I think all of us
hopefully have realized now
AI is, might change a job,but it's not going to take your job.
Then I worry is AI going to take my IP?
Will it take my secret sauce?
Will I lose my competitive advantage?
And then I worry like,oh gosh, my data is kind of messy
(05:56):
and I have data,but my data is not structured,
you know,and if I go back five, ten years,
do I really want to use that datato do the things I'm doing today?
And it just kind of goes,you know, on and on and on.
And the other question then becomes,well, will my customer be okay with me,
you know, using the data on the projectto deliver better outcomes.
(06:17):
And so there's this likesort of constant journey.
But I think what's helping usa lot are all the,
you know, that commercially sort ofconsumer-like use cases and the use cases
we find within our Microsoft applicationsare really helping people realize it's
not as maybe scary or as,and that there's a ways to go.
(06:37):
Right?
Or it's doing some pretty,you know, summarizing meetings is great.
But we have a ways to go beforewe're getting, you know, deeper
into some of the more knowledgebased, activities that our people do.
So, Amy, I think it's interestingthat you went to kind of that fear
and the transparencyand what happens to my data conversation.
(06:59):
I thought it was really interestingat Autodesk University this year
that was really a focus for Autodesk.
Was that
decision really something you got toso that you could drive adoption?
Like we need peopleto have confidence in this?
Was that how you positioned it?
So we, you know, trust is really importantto us with our customers as it is,
(07:19):
I think, with all of youand your customers.
And so, we knew that one of the important
hurdles that people had to get overwas really understanding
how their data was being used.
You know, what the risks were,if there were any.
And us being transparent about
what we were doing with the data.
So we've always had I think, like
(07:40):
many of you,we have internal governing bodies
that govern both the internaluse of AI at Autodesk as well as,
you know, lots of approvalsand checklist of anything we're doing
withinthe products is ethical and appropriate.
And so we have a lot of thosethings happening.
And now we're trying to expose that
more externally with these,they almost look like nutrition labels.
Really helping you understand:
what is the feature? (08:03):
undefined
How does it use AI?
And a lot of our use casesdon't even involve project data per se.
Some of them use synthetic data,surrogate models,
inferences, all kinds of thingsthat are just really good
(08:23):
for everyone in a, in a very zero-riskkind of way.
Right to that point (08:29):
even to VHB this is
one of the biggest challenges, you know?
Good data yields good AI, right?
What is VHB doing to help
drive the data standardization of VB?
Well, it's interesting, Dave,there's a couple things around that.
I think when you look at ourkey initiatives, it's really,
(08:55):
it aligns well that our data initiative
and our AI initiativeessentially both started at the same time.
It was almost like we realizedyou can't do one without the other. And,
you know, you've heard me say before
that data is the engine that feeds AI
or the fuel that feeds the AIengine rather, and really, with us,
(09:16):
it is about standardizingwhere we find our data and really trying
to formalize the processaround documenting authoritative data.
So leveraging toolsthat are our standards in that industry.
You know, ACC Amy, is a toolthat we're using more and more here
at VP around submissions, around,
(09:38):
tracking of milestones on projects.
And I think that is helping,Dave, versus the lack of clarity
that we've always had around,oh, well,that file's got a date in it.
Does that mean it wasn't good at that date
or that was the datewe submitted that file?
You know, I don't know, Amy,
hopefullyyou're finding we're not the only customer
(09:59):
that has that problemthat we're trying to address.
Well, on the journey to digitize,
for most of our customers,having access to the most
current information is one of the,the most important things.
And I think one of the great use cases ofAI is really search and retrieval,
and it does it a lot better
than any human could do a lot fasteracross multiple systems.
(10:22):
And it can helpyou look for those discrepancies right
before you go offand make new decisions or,
you know, spend money on things.
And so that's I think is one of the waysthat also gives people confidence.
And one of the cool things aboutsome of that is like if you
when you get the information,you know, having that attribution
(10:43):
and being able to know, okay, you found itall, but don't just give me, you know,
just give me the answer.
Give me the, the backgroundso I can go and find it.
And I know you guysare passionate about that
as part of your AI journey as well.
Yeah.
So, I'm interestedbecause I've heard Ryan mentioned this.
But also you just brought it up, earlier,the new data nutrition levels.
(11:07):
Maybe if you can talk a little bitabout that.
I think it's interesting
from the perspective of the correlationof that versus the data.
Yeah.
I mean, this is something we made up,but it was more like to have
a consistent way of describinghow we're using AI.
I think one of one of the challengesin any software company is there's
(11:27):
a lot of different ways to deploy AI.
And so you want to be able to repeatedlytell your customers
how you're using data.
So one at one example is we we have generative design capabilities in the products.
So for example, a generative
design featuredoesn't really train on any customer data.
That's all algorithms.
(11:49):
A lot of software in the backgroundproducing
those generative examples that we'recoming up with for our customers.
So, the nutrition labelwould clearly say what method is it using.
Because there's dozens of differentAI methodologies.
What type of data is it using?
Where did the data come fromand the kind of checks things off.
And so you can look across our portfolioas we build all these out and see,
(12:14):
where are things, you know,how is data being, how is data being used?
On the other hand, we have AI that helps
answer health questionsright through Autodesk Assistant.
And so that is ingesting like all of the
documentation we've made over the yearsand all of that.
So that's a different type of data.
And so just being clearon the broad spectrum of data
(12:37):
and what we're using, where, and how.
You know, Amy,you mentioned the assistant.
And we've been lucky enough to be testingsome of your newer assistants.
Like you recently releasedthe SEC AI Assistant,
which is connecting to datathat we're now storing in the cloud.
And it's connecting to our data.
(12:58):
Have you found that that is where
the biggest demandis from your clients around AI right now?
Your customers?Is it around those kinds of tools?
You know,I think it's a journey right now.
A lot of people want automation.
So I would say in terms of,you know, automating a tedious tasks,
whether it's on the design side,
(13:21):
the handover sideor the construction side, I think anything
to do with counting the counting things,aligning the, you know,
one of the problems that we solved inAutoCAD was block alignment, right?
You might use five different toilets,or the same toilet, but you might use
five different blocksto represent that toilet.
So we can go through a drawing and alignall of that.
(13:43):
That's just general productivitygoodness that you want to see more
and more of that so that your peoplecan focus on the high value tasks.
So I think everyone loves the automationand the productivity enhancements.
And after that,then they are starting to think about
how can I create insightsfrom the data that I have?
And that's where the Assistant startsto come into play.
(14:04):
Like can I look at history and predictmaybe RFI
if we always have problemson these projects with and
we have these types of RFIs, and can wesomehow figure out how to make changes
earlier in the project on the next projectSo we don't have the same problem?
So I think you go from automationto kind of
(14:25):
search and retrieval,to then creating insights.
And then for me,
what's going to be great in the futureis then, you know, taking action.
So to tell,you know, to use an agent to tell it to go
and make a change on your behalfso you don't even have to go in there.
You know, it can go across a projectand make a change for you
without you having to, you know,have a human do all that work. So,
(14:49):
as you look at the
entire suitethat you support and bring to a project,
whether it's info drainageto do drainage design, whether it's Revit
to do structural modeling,whether it's civil 3D
to do the work that we do around sitedesign.
What's Autodesk’s strategy around
using an AI assistantor tool across all those platforms?
(15:14):
You know,that's a lot of different data types.
Yeah.
So we're building this thingwe call the AEC Data Model.
So part of what allows a lot of the valuehere is getting the data
out of the proprietary file formats
that we've all used for decades. So,
right now we are working to populatewhat we're calling in the AEC Data Model.
(15:34):
We started with Revit data.
But by doing that you've nowgot a granular data model that you can,
you know,
interrogate withouthaving to like open Revit and use the API.
So it's very, very much connected to docs.
And so we started with Revit
and will be adding moreand more types of data to that.
And what's interesting thereis, this will be an open ecosystem
(15:57):
so, you know, otherscan use the data model if they choose.
But also you'll be able to take that dataand mash it up with your other data.
All your data in SharePointor your ERP system or what have you.
Because it's not just right, it'sthe some of the insights of the future
will come from combining the business datawith the design data.
And I knowfrom talking to all of you in the past,
(16:19):
we don't think
you should have to haul all this datainto one common location to do this
kind of kind of work, because you're goingto use multiple systems.
That willget expensive and time consuming.
So the way we can connect all these thingsusing an AI will be, will be great.
But it has to be on a granular orvery fast basis and a standardized basis.
(16:40):
It’s exciting to see the unified approachcoming in and actually connecting.
Ryan, you went through several of thedifferent applications that Autodesk has.
This may be a completely random
question for me because I'm curious,what has been the best success story
that you have seen at Autodeskrelative to some of them?
(17:00):
Ryan, you mentioned that,you know, informa, for example.
One of the areasthat I'm just extremely intrigued,
is your energy optimization,the rapid optimization tool.
You know, it's a big pushthese times these days in the industry.
So where are you starting to see, like,the biggest push and the biggest momentum
and some of the biggest success stories?
(17:23):
You know, I would say,
there's a couple thingsI'm really proud of.
So one thing I'm really proud of iswe started, gosh, 3 or 4 years ago,
using the usage insights
to give users advice on and to detectwhen a user was frustrated and,
you know, kind of get in thereand give them feedback.
And we've now made that more,nearly real time.
(17:45):
So we know our products are complex.
The more we can doto help users on their learning journey
and even the expert usershelping them understand
maybe there's other ways to do thingsthat they weren't aware of.
So I'm really proud of that.
I'm also,
really proud of what you talked aboutwith the energy modeling and the carbon
(18:06):
and all those things, becausewe've been able to take something that
used to have to go out to a consultant,you know, come back,
I don't know, a week, 3 or 4 days laterand put it in, throughout the portfolio,
but also in the very early phaseof the project so that people can see
in nearly real-time.
as I make changes on that conceptualdesign, the impact
(18:27):
that it hason noise, and other sustainable,
environmental conditions on the projects.
And so that I think is very coolbecause we have the,
we had to have the detailed capability
in order to create the surrogate modelthat we used to do the near real-time.
So we're really democratizing the abilityto do that type of analysis with AI.
(18:49):
That's exciting. It's exciting.
You know, it's,when you started talking about the
sustainable side of thingsand you talk about carbon counting and,
you know,this is actually one of the biggest pushes
that we're making, as you know, VHB,we're looking at how do we actually help
make, and change the carbon footprintand the carbon impact that we're doing.
So, it is a great partnership, Ryan,that we have with Autodesk
(19:12):
and trying to push this insideour footprint.
So, what are we doing at VHBto actually help that as well?
Where do you see some of the areasthat VHB is going
relative to the embedded carbonsand the calculations and so forth?
It really
started with our change in approachto design, where
when we made the commitmentthat we will be building models,
(19:33):
we will not be doingtraditional plan development,
even if the client isn't requiringthat model.
You know, if we're going to commit
to sustainability,we have to commit to having data
that we can understand our impacton our world around us.
Right now, we're doing a very detailedcoaching program, Amy, that I'm sure
(19:57):
you're aware of with our staffteaching them how to maximize,
you know, the tools that we have availableto us and part of that is
we are bringing in our sustainabilityleads at VHB to talk about embodied carbon
and doing embodied carbon calcsas part of our design process.
You know, Dave, so it really isgetting people to understand that
(20:19):
the data is not just valuableto the client, it's
also valuable to usthroughout the process of design.
So, Amy,I think that it's really interesting
to see how we're going to be ableto position some of these tools to overlay
when we have that ecodata model available to us,
(20:39):
so that we can get AI to give us thatdata back I'm not going to say instantly,
because there's work that needs to be doneto interrogate the model, but,
much faster than we could do it before.
Yeah.
Well, and I think one of the great thingsabout AI is it's very,
amenable to experimentation.
(21:00):
And I think that's one of the thingsas I talk to
customers is really important,is there's a lot out there.
We don't know which ones are going to endup being the very best for our use cases.
I think having a
sort of a learning mindsetaround these things and experimenting,
and then what I've heard from customersis they start out
kind of with a bit of a Wild West,in some ways.
(21:21):
And then they start to determine, well,which of these things actually deliver
value to my company, which of these thingsdeliver value to my customers
and then start to hone in on where it'sworth
their overall investment to advance.
Then keeping an eye out onwhat's just happening across the industry,
because this is, in my career,the fastest changing environment
(21:44):
that I've ever seen.
Thank you, thank you.
Ryan, so relative to VHB.
Similar question to you.
What is VHB doing in some casesto educate and inform internally
because we have a lot of differentapplications,
a lot of interest inwhat Autodesk is doing.
So how are we educating folksinternally, at VHB,
(22:06):
on what Autodesk is doingand getting that message out, as well?
There's a few different things, Dave,that we do around that.
I'd mentioned that we have that
advanced training that we do,which definitely,
my team alone cannot support2200 staff members.
So I think do a good job, though.
Yeah, really good job, my friend.
(22:27):
Thanks.
But, you know, the reality is,if you really want to drive,
change, driving that changefrom within the teams and giving them
an understanding of what these tools doand how to leverage them is important.
So, you know, we do everythingfrom training those advanced users to also
we run open sessionstalking about new tools
and advancementsthat that are coming into the tool.
(22:50):
You know, I think the AI Office Hours here
at VHBgets a pretty significant attendance.
And we run that a couple times a monthwhere it'll be focused,
not only on technology staff
bringing tools to the users,
but also on userssharing their experiences with each other.
You know, I think,
(23:12):
while everybody always trust technology,Amy, I don't want you to think that
people doubt technology and upgrades.
And is everything going to work? Italways works.
But, hearing from that end user that
this is really somethingthat applies to me and was valuable,
we have developed some of our ownassistants, Amy, similar
to the Autodesk Assistantthat allows us to ask a question to a tool
(23:35):
and have it providea client-focused answer.
And, you know, validating thatand working to tune those.
I think that's one of the big things
that has really changedpeople's perspective here.
Dave, I don't know what you think, on AI,is that the recognition that
it's not “well, this is what you getand you can't make it better”
(23:57):
but you make it better by using itand testing it
and finding where it makes mistakesand training that model.
So, you know
have you seen that
same sort of reaction, Amy, with “oh,the first version of this tool
really wasn't great, but the feedback madeversion three much better.” Yes.
Well, and and quickly,which has been great.
(24:18):
So, like we started with,
the GitHub copilot for software developersand, you know,
the first release of that was very much
not appreciated,but the people that stuck with it
and kept going there, you know, yes,they're contributing, making it better.
And they're actually seeing real,real impact.
I think that's important.
(24:39):
Certainly anything we release,but also with all the other external ones,
the rapid development hereis really great.
And they get better by design.
That's kind of how they're meant to be,which also means for all of us,
you and my team included,
who are building these things, it'snot a set it and forget it kind of thing.
Right.
They can also wander off and hallucinate,if you're not, you know, if
(25:01):
you're not careful about
constantlyunderstanding how they're performing.
So that is a different mindsetthat we bring to developing these things.
It's really kind of interestingthat you see that internally, too,
as a software company that, you know, oh,
first one,nobody was really that thrilled with it.
But as we go on,
I always think of engineers being a bitstuck in their way, I don't think of,
(25:24):
software developers but…
Let's talk about that, because what I hearfrom most of the customers
I talk to is, you know,they do what you've been doing, right?
They find the evangelistsand they have the meetups
or the weekly tips and just tryto encourage people to use it.
And the thing I've spoken with themabout a little bit too is what about
(25:44):
the aging
population a work that, you know,has all of that amazing knowledge,
in their brains and it's in the projects,but that are maybe a little reluctant
to engagesometimes with some of these tools.
Even at Autodesk I can tell you like we,
you know, change is hard and exploringnew ways of doing things is hard.
(26:05):
So we're very much doing a lot of thingsto encourage people
to at leasttry these things out in some way.
And I'm wonderingwhat you all are doing there?
So we do try, right?
And that is a constant challengewith anything in terms
of new technologies, right?
In terms of the testing, trying,I would say back in 2019,
(26:26):
VP senior leadership,
really pushed,
trying to actually engageand lean into some of the AI
initiatives that were out there, testit in a, you know, a very safe sandbox
and making sure that it was protecteddata that were not out,
you know, publicizing things outand anything out there.
The interesting thing,
(26:47):
Amy and I have heard that relativeto the aging population and where,
you know, that legacy knowledgeis actually going to be lost, right?
When is the day that we're actually going
to create an AI assistant for that personthat's basically mirror matching and
replicating everythingthat came out of the individual?
And how does that happen?
I personally, you know, I believethat we're probably not too far away
(27:09):
from creating those type of agentsin the personalities.
And I can actually seea, an avatar of myself
in about, you know, five, ten yearsis going to have the same personality.
I apologize to all the listenersthat's going to have my personality on it!
But, you know, it's going to happenwhere we're,
you're going to start to actually getthat information in there and,
retain some of the knowledgethat would have been lost over time.
(27:32):
One of the....
If you think about them as,you know, coworkers, right?
You're kind of the senior personin training
and onboarding the more junior coworkerand they'll learn pretty quickly.
Yeah.
And Ryan, one of the biggest pushesthat we started probably about a year
and a half, two years, which is part, youknow, part that your teams involved in.
A lot of the folks
(27:52):
across the footprint of the companyis trying to actually standardize
the data that we collectso we can actually make better informed
decisions and get insightsthat you were referring to earlier.
Amy, we've started to actually automateall of our collection process,
you know, our field collection, our datacollection, to go away from the paper,
hopefully in the next year to two years,we won't even be talking about that.
(28:14):
We will be all digital formatat that point.
That will let us make better informeddecisions, better insights
into where we make our decisions.
So, Dave,
I actually would like to hear the answerfrom both of you on this question.
One, as we standardize our data, and two,
as AI becomes more able to assist us,
(28:37):
with not only just tasks that are builtinto the software, but with things like,
you know, fostering
a culture of citizen developerswhere it's no longer scary anymore
to say, hey, how do I do a Python scriptthat does this?
Are you seeing that kind of,
outside the standard tools,really become a drive here
(28:57):
and a push for VHB to have more peoplewho can understand an API
that Autodesk may provide us with, or that
Microsoft provides us with?
So, Ryan, I'm going to answer the questionby going back a couple years.
And we were presentingto the executive leadership team
and we asked the question,of several of the younger folks, you know,
(29:18):
where do you see the design team, Amy,going relative to VHB?
And they said this wasone of the responses: You may have today,
ten engineers or designers on this,and maybe one software.
We're looking and pushing that in,
probably five years or less.
(29:38):
you're now looking at upskilling
some of the different team membersto actually have different skill sets.
And you may end up having
five people, but,you know, some of them may be a majority
or 3 or 4 are on the, the software side,or at least on the prompting side of it.
It doesn't relinquish the rightor the obligation
of an engineer to essentially reviewand certify the documents.
(30:02):
So that’sstill going to have to be and hold true.
There's a lot of different questionsrelative to the full automation
and the requirements for insurability,licensure, the delivery and so forth.
But I do see a change, Ryan.
I see a change in
the team compositeand how they're actually built out.
I look at it as partnering
with Autodesk to understandwhere those workforces are going.
(30:24):
And we're going to have to evolve.
You know, that also starts, in my opinion,
going back to the middle schooland high schoolers.
And as they come up through the programsin the universities, that has to change.
Doing what we're doingtoday is not going to be
the same as what we'redoing in five years from now.
Yeah.
So as a software company,our workforce will transform also,
a little differently.
(30:47):
We are obviously hiring more data people,
you know, peoplethat can analyze the data, people that,
you know, understandhow to do AI at scale.
But we also when we do that, like as inanything we build,
we can't sacrificethe quality of the product,
all the ethical and trust and safety andall the things that we put in the product.
(31:08):
So, you know, all those thingsneed to move hand in hand.
Where I see interesting potentialtransformation is sort of the interaction
paradigms will change with the productsas we get more
AI kind of doing thingsthat you are all doing manually.
What is the new way that you're interactwith the product, right?
What is the prompt?
(31:28):
What are the prongs going to look like?What are the inputs going to be?
It's not just going to benatural language over time.
So we need to do more researchand really think about how
our user experience teamsare designing experiences for the future.
Maybe so you can have some of those,low-code no-code
citizen developerkind of things happening.
(31:51):
Amy you opened it up.
You opened it up.
So five, ten years for now,you got your crystal ball.
What does it look like?What does it look like at that point?
Oh, yeah.
Well, I definitely.
I mean, everyone's
talking about agents and the workflows,and so there'll be more and more of those.
I think to your,
to whichever one of you made the pointabout okay, safety's important.
(32:13):
We're not making a 2D image.
We're making,you know, a 3D thing that that,
has to be sound and performantand all that other kind of stuff.
So I never, I don't ever think.
I think we will get further, but we willstill always have humans in the loop
kind of making surethat we're doing all of those things.
(32:33):
I do, one thingI've been thinking about in regards
to this question, is I think that the
notion of, operating system
and platformwill become kind of a moot point.
I don't think I think all the desktopproducts will be the lighter.
You know, we'll have web, mobile,all the different kinds of, applications,
(32:53):
but I don't think it will matterso much as to where
things are necessarily running.
I think we'll be in a situationwhere, you know, you do the workload
wherever it makes the most sense for,you know, for your people.
If they're on a jobsite with a mobile,an extremely powerful mobile device,
things can happen there.
You know, if you're in the office,if there's certain things that are,
you know, well-suited for the web,I do believe all the data
(33:17):
we're talking about will be normalized
and standardized in the cloud in some waythat helps us drive all these workflows.
I think
the notionof, you know, desktop or cloud or
any of that will kind of blurin a way that that is great for everyone.
So with that, Amy,do you (please tell me you do), but,
(33:39):
do you see a future where, you know, okay,
the scene between that design model
and the data used in construction…Youknow, one thing that always scares me
is that, you see the stats of 90+percent of our data that we use in design
just vanishes.
(34:00):
Do you see that day where, “Okay.
You know, I'm just going to bring you
into the project now, contractor,take that data and go with it.”
Is that something that you seein the five year future?
It's, well, it's
something I hope for in the industryand you know, in addition to that.
It's really about building a trustand aligning on the deliverables.
(34:23):
So that whatbecause I was in this discussion just
this week, is the architect'smodel should be the constructible,
you know, like, should that designbe the construction version?
And, and if we can use AI to kind ofcreate different representations
of the data so that it's purpose -builtfor the different
(34:43):
use casesand then bring that all the way out to,
you know, a digital handover,at the end of the project.
Then we've got the digital data being
the data of record as opposed
to, you know, drawingsand things that might get outdated.
So yeah.
I think AI will help us increase trust.
I think it will help us create
(35:06):
the different representations of the datathat we need for the different use cases.
And then I think it will help us createa closed loop of feedback,
from the very end to the beginning wherewe say, oh, yeah, don't do that anymore.
Right.
Like,let's talk about water penetration issues.
If we could track those, you know, throughall the decisions made in a project
(35:26):
and then gradually diminish because nobodylikes to work on any of that stuff,
and it would be greatto make that a non-problem in the future.
So, Ryan, you I'm coming to you nowbecause I want to hear,
and now that I get this recordedon, on the viewpoint,
where do you see WBS designservices aligning and going in the future.
(35:48):
You knowsame question that Amy had to answer.
You know, fair back to you.
So where do you see thethe VP positioning to go, you know, five,
ten years down the road?
You know,honestly five years down the road,
I think the first changethat we see with our design teams
is a change,
into a, a data steward model.
(36:09):
I'm going to call it,you know, really, especially for us, Amy,
you know, in the architectural spacethat's been there for a while,
like you have had to have someone who is,
really kind of curating that datathroughout project development.
And I think that that's a big changethat our teams need to get over the hump
into this model of,
(36:31):
I need to have curated dataso that we can effectively leverage it
for everything from stormwater,stormwater analysis
to traffic flow to, you know,construct ability concerns.
When we start
talking about clash detectionand other pieces like that doesn't happen
with each of our different partsof a project team working separately.
(36:53):
So I think that the firstthing we're going to see is
a new focus on
who is our datacoordinator on this project.
Because that data coordinatoris going to facilitate
what tools can be available on a project.
I hope five years from nowwe are talking about a world
(37:15):
where data is going from
format in concept all the way to tandemand a built environment
where you're using Bluetooth sensorsor whatever to tell
what traffic flowis really like at this intersection.
But I think you don't do thatunless you really have someone
(37:36):
who is managing the datain the expectation of that data
downstream, is the biggest change I thinkwe're going to see on our design teams.
So let me ask a question,because some of that
will probably requireworking differently with your customers.
And I wonder, what are you doing today
to help your customersbe more comfortable with AI?
(38:00):
Or are they already super comfortablewith AI?
Dave,you want to start answering this one?
The answer tothat is there's a range of clients, right?
Some are still on their very infantjourney on where they are.
So, the initial reaction is, ones
that are in that stagedon't overreact to the AI terminology.
(38:22):
There's a lot of folks like Autodeskthat are actually out there,
doing a lot of research and understandinghow the business is going to change.
A lot of our state agency partners,
are really, wait and see in some cases.
And then you have some progressive states
out there that are basically saying,we're going to go all in.
(38:42):
Texas DOTliterally a week ago, week and a half ago,
publish one of their AI master plans interms of the direction that they're going.
So it's interesting to see the range.
One of the things
that we're seeingand I think it's a common theme here
today, is cautioning,
(39:02):
any client, whether it's public sectoror private sector, to just jump into AI,
but more so take a data readiness to itfirst.
You got to have the appropriate datagovernance in place.
You have to have the understanding
where it is, elsewiseyou be wasting time and effort
and a lot of manpower to actuallygo into that, into the AI portion of it.
(39:23):
There are some really progressiveclients out there
that we work withthat are looking at AI and testing it.
Some folks down in Florida,some of the agencies
down in Florida are testing AI relative to
crash detection systems and so forth.
How are we helping them, Ryan?
We're trying to actually get aheadof some of the advanced technologies
(39:44):
out there,the emerging technologies, understand it
and actually marginalize their risk
by testing the futureof where AI is going.
Because it is evolving so fast.
The biggest message isjust don't go too fast
without understandingwhere you are running to.
Yeah.
And I think, Dave, you know, aroundthat things like wrong way driving tools
(40:08):
and other tools that are very practicaland really help improve safety
and allow for betteremergency response are areas where I think
I've heard more successes with our clients
and I think, you know,avoiding the absolutes.
One of the things I worry about is,
at some point, if you're not leveragingAI, you will be falling behind.
(40:31):
Whether you're a DOT, you know, a companylike Autodesk or a company
like VHB, your competitors or customers,
people are going to expect more from you.
And if you're not on that bandwagonand haven't figured out
how to make it work for your business,you're not going to be as relevant
in the future.
You mentioned that the older generation,Amy, if you go to the other side of that
(40:54):
coin, you know, there's not going to bea graduate that starting college today
that doesn't come out of schoolwith an understanding of how to leverage
AI to help them eitherwrite their term paper
or figure out an approachto solving a problem.
If we made the blanket
(41:14):
statement of thou shalt not use AI at VHB,it would be incredibly
shortsighted and challengingto recruit the next generation.
Yes. Yes.
And Amy, Ryan, I know we're pushing upon the hour here and I appreciate it.
I agree with you both.
I think AI is actuallyone of the great equalizers.
(41:36):
It doesn't matter the size of the firm.
I think everybody should lean into itbecause there is an opportunity
for everybody to understand the influencesof AI in your specific business.
So I want to leave with one question.
We have a lot of listenerson this traditionally.
Let's leave our listenerswith one action item
they can take right now to be in placewithin a year.
(41:57):
What would you suggest?
What would I suggest?
Well,
I would, I would suggest that
if they're not doing something with aAI that they pick one.
They do
(42:18):
a very lightweight scan,try to pick one thing
that they think will have an impacton their business Make sure it's safe
to try from a, you know, ethical, trust,data perspective.
And, you know, get their people,especially their senior people,
to try itand to do what you guys have done.
Lead by example and then build that
(42:38):
environment within their companies sothat people can do safe experimentation.
I appreciate that the security andthe safety of leaning in and trying it.
Ryan, how about you?
So I think that there'ssome real opportunity, Dave,
to step backif you haven't got into AI yet
and look at what are the things that cause
(43:00):
the most painfor you during project development,
and look for that opportunity
to either reduce risk,and not talking about
reducing the risk of using AI,but reducing risk on a project.
Is there an opportunityhere for us to evaluate
(43:21):
frequency of change orders,
frequency of RFIs that come in, andwhere does that put this project at risk?
Look for that opportunity to leverageAI to provide insight
or streamline a process that you do
on every job.
I look at clash detection, Amy,to go to a ridiculously simple one.
(43:45):
Why are people doing this still
with a plan and a pen?
That is somethingthat we should be automating using
AI to help with that.
And I'm going to take the soft balland take the combination
of the two of yours response and say,
I agree, we need to identify one area.
(44:07):
Identify one area that you're really goingto try and lean into and understand.
Ryan, taking your your pinchpoints, your friction,
your area and take that onepoint and identify what it is
and then evaluateif there's an off-the-shelf
solution out therethat actually can marginalize your risk.
Because there are a lot of them out there,
(44:27):
that you could testin a very safe environment.
And I would encourage folks to listenand actually do a lot of research
on, exactly what Autodesk is doing,and other vendors.
Autodesk has a lot of informationout there, YouTube and other areas
to actually go and understandwhat is actually happening in the market.
(44:49):
Knowledge is power,and you cannot take that away.