Latest / Engineering Influence from ACEC / Revolutionizing Civil Engineering: Bentley's AI-Powered OpenSite+
Transcript
- 0:00Music.
- 0:06Welcome to the Engineering Influence Podcast from the American Council of Engineering
- 0:11Companies, brought to you this week at the 2024 Annual Fall Conference by Bentley Systems.
- 0:18And very pleased today to be joined by representatives from Bentley to talk
- 0:22about a new suite of software, OpenSight Plus.
- 0:26It's the first to market AI-powered civil engineering software specifically
- 0:31built for site design and land development.
- 0:34Today, we'll be speaking to their product development team to explore why it
- 0:39addresses engineering challenges and how it equips engineers to deliver high-quality
- 0:43projects up to 10 times faster than their existing workforce.
- 0:49And to be joined in this conversation, very pleased to welcome Francois Vallaud.
- 0:55With VP of Civil Engineering, and David Sillemeyer, Senior Product Manager.
- 1:00Thank you, Francois. And David, thank you for joining us. For having us.
- 1:03So this is really something, and I actually think this might be something that,
- 1:06you know, we just, like I mentioned before we went on air, that we had a student
- 1:12who was learning Bentley and, you know, Bentley systems.
- 1:16And, you know, just thinking about the fact that when she graduates next year
- 1:19and joins a, actually she was a geotech, so land development probably,
- 1:23she'll probably be using a suite of services like this.
- 1:26And AI has been sweeping the industry, creating a suite of software like this
- 1:32is very timely, especially for the industry.
- 1:35I mean, really, what's the differentiator here? What makes this product stand
- 1:39out and really be a benefit for firms engaged in land development?
- 1:44Right. So OpenSight Plus is the first ever AI-driven site civil product.
- 1:50It's definitely dedicated to engineers and drafters that are trying to really
- 1:55create site designs, whether that is a, you know, into the commercial,
- 1:59industrial, or residential sector, and differentiated by the fact that it offers
- 2:04a lot more productivity than any cats of or other. You mentioned the student of, as you just said.
- 2:10A student out there when they get out of school they don't know a lot
- 2:13right so we built these next generation tools that are easy
- 2:16to learn easy to use ai driven so we
- 2:19built a co-pilot experience that means you
- 2:22could do things like interrogating documentation and requirements of
- 2:26figuring out what is needed in there so that is
- 2:29helping you know especially new worker that
- 2:32essentially don't may or may not know the regulation and the reason they work
- 2:34and yeah own in and to what to do then can modify the model it can interrogate
- 2:40the model by talking to the software this is a new thing right so before we
- 2:43had our keyboard our mouse yeah you could use your voice to connect stuff and
- 2:47then the second thing it does that does that i think is very very powerful is automating drawing.
- 2:53Drawing automation drawing creation takes
- 2:55up to 50 percent of a project that five yeah that's
- 2:59a lot of time that you are spending placing text on the
- 3:02piece of paper essentially in the computer in the cat's offer
- 3:05so we're replacing all of that by an ai
- 3:07engine that is able to automate all this
- 3:10for and and will optimize for your technical
- 3:14requirements eligibility as well as steps this was not possible before aesthetics
- 3:19it's not something that a computer understands but ai does understand research
- 3:22yeah right so it does a lot for the engineers to help it really quickly draft
- 3:27design optimize the layout and then produce the drawings at the out of there
- 3:31yeah so it's it's a game changer.
- 3:34Really, it sounds that way. I mean, one of the things early on in the discussion
- 3:39with AI was all about, you know, of course, it's maximizing efficiency.
- 3:43And a lot of it was, you know, dealing with more of the management side of the
- 3:48house, of course, trying to use it to actually maximize efficiency and marketing
- 3:52and generating, you know, proposals and search and QBS searches and things.
- 3:58I mean, RFQ searches and things of that nature.
- 4:01But I mean, what i what well i was told at the time this a couple years ago
- 4:05was getting to the point of ai since a drawing is like the next step of actually
- 4:09being able to assist the end user of producing.
- 4:13Plan and having ai drive that yeah that's a real breakthrough i mean i think
- 4:19actually using ai to produce something that's like a drawing that is more than just you know,
- 4:24you taking in text or information and putting out
- 4:27an answer you know how long in the
- 4:29development cycle was that i mean that's not a simple thing to
- 4:32accomplish no for sure and the the neural network
- 4:35that we have built in there built over a long period
- 4:38of time keep on optimizing it yeah we also
- 4:41are very proud of the way we deal with our customers that's also
- 4:45something that we would differentiate what you talk about the implementation right yeah
- 4:48one of which is how do you train these sentence right we
- 4:51cannot as an as a software company and
- 4:54we should not any software company should not do that yeah like
- 4:57some other competitors but use their customers data to train
- 5:00their own and that's a big and that's a that's a huge thing right huge yeah
- 5:04we decided not to do that at all so unless we have specific data agreement there
- 5:09where others are just using the the end user agreement and basically sort of
- 5:13sweeping your data yeah from their system which is not a right thing to do ethically, right?
- 5:19Yeah. It's probably breaking this data engineering firm, put the data in there.
- 5:22It can not even be their own data, maybe their owners.
- 5:25And so, and then they're worried about IP being generated and so on.
- 5:29So our data policy is that we won't, we don't, and that's making it harder on us, right?
- 5:34From a technological perspective, we don't use our use data to train data.
- 5:38Yeah. We find other ways.
- 5:40Yeah. That's critically important, especially for firms that are doing business
- 5:44with government agencies or, you know, the fact that that data security and
- 5:49knowing how it's being used and the constraints that are being put on it,
- 5:53I mean, that's essential.
- 5:55You know, David, you know, what are your thoughts here? I mean,
- 5:58how is this really going to impact the end user?
- 6:01So for the end user, this is a game changer, no doubt. Prior to doing software,
- 6:06I actually owned an engineering company myself.
- 6:08Like Francois talks about, plan production is absolutely a massive take of my overhead.
- 6:14And now with even the workforce issues that we're seeing, there's not enough people to do it.
- 6:17So now you compound that plans are becoming more complicated,
- 6:20having to have more deliverables, more sheets.
- 6:23So when we started creating this product, it was about reducing that pain.
- 6:26So using AI, like Francois said, we're not training on their data,
- 6:31but we're creating a system where they can add their data in their area and
- 6:35then have it learn to create plans the way they do.
- 6:38So every engineering firm is completely different. And Having that ability to
- 6:42press a button and then 10 seconds later, look at my sheet, my way. That's huge.
- 6:48Yeah, it really is an important thing because it allows an engineering firm
- 6:53of any size to really create a library of data and a process that is unique
- 7:00to their way of doing things.
- 7:02And you're providing the platform for them to structure that data to really
- 7:07produce something that would be completely unique to another firm using the
- 7:11same software platform.
- 7:13Yeah so then so we talked about the copilot stance and
- 7:16we go back to that because we talked a lot about the drawing production yes fantastic
- 7:20it is a huge huge improvement at the word book
- 7:22yeah but the copilot also is the way this works essentially you
- 7:25have you create a new project you say my site my acme site layout yeah you go
- 7:30ahead and do that you can upload all the documentation that you have from the
- 7:33different sources in there yeah right it will index all that data within your
- 7:37own private space and from there you can start interrogating can i build an
- 7:41hotel in this area yeah you know How many parkings do I need?
- 7:44Is my design containing enough parking lots to actually meet my requirements?
- 7:49What are the setbacks? All those questions that may be in the head of the engineer,
- 7:54and he may not, or she might not remember what's out there.
- 7:58And so we allow to tap into that documentation.
- 8:01Instead of reading 800-page documents,
- 8:03he just goes in there and provides evidence-backed answers as well.
- 8:07What I mean is that if it says, yes, you can build a no-del in this zone,
- 8:10it will show you where it's found in the document. and then you could read it yourself out of it.
- 8:15Yeah. Because RAI strategy is not about replacing it around an augment.
- 8:19Yeah. And making them much more efficient in what they do.
- 8:23Something about the co-pilot that's, to me, very valuable. As an engineer,
- 8:27we are very worried about getting bad information. Yeah. That's why we got E&O insurance, right?
- 8:33Now the point, we had this question here. Well, where's the data coming from?
- 8:36This is not the co-pilot like you use at your office, just searching the internet.
- 8:41This is very specific on the data that
- 8:43you're searching so if it found an answer that's a
- 8:46good answer you can feel you know reliant that
- 8:49you've now got the right information to make a decision you're
- 8:52not going to worry about it later on because it's the data that you put into the
- 8:55system that's a proprietary to you that actually is generating the answer yeah
- 8:58so that actually that's right i mean from a from a i mean when you're looking
- 9:02at from a business case i mean david as a as a former owner i mean the amount
- 9:07of of savings and time i mean for an average Average firm utilizing this.
- 9:13I mean, have you guys done a kind of an analysis about what a usual cow savings would be for.
- 9:19A project on this instead of the traditional way?
- 9:22Yeah. So again, projects are all different sizes, all different.
- 9:25So when we talk about the 10X return, these are numbers that we got from the
- 9:29user that are working with us to build the tool. Yeah.
- 9:32They looked at it. They've seen what we do. They're touching it.
- 9:34And they've come back and said, you know what? This is 10 times faster.
- 9:38So it depends. If you're a little more savvy on the software,
- 9:40you may see four or five times, which is a huge number.
- 9:43If you're a little further behind, you might see 10. And for some people,
- 9:47they think it's probably going to be greater than that for its overwork.
- 9:49Yeah, it saves time. But also think about the also the mental space for the people.
- 9:56Yeah. Right. How many projects we all worked on and essentially ended up producing
- 10:00the drawing over the weekend because they added there was a last minute change. Right. Yes. Yes.
- 10:04Hopefully this person can use it on Friday and use the AI, validate the plans
- 10:09and ship it to the client as opposed to have to spend the whole weekend.
- 10:13Right. Yeah. And how much does that do to you, to your, the quality of life of your people?
- 10:16Yeah. So you don't necessarily have to squeeze the lemon, but rather give a
- 10:20bit more freedom and also use the engineers for what they're great at.
- 10:24Mm-hmm. Change and design.
- 10:25Yeah. And then remove the, you know, any of the mundane tasks and the tedious
- 10:30stuff, right? Out of the work.
- 10:32That's what we're trying to do. And I think that that really is the kernel of
- 10:35the importance of something like this, which is AI driven, is that it allows
- 10:38the engineer to be an engineer. It allows them to focus on their core competency
- 10:42instead of the other things that have to overlay that.
- 10:45And also from the client side, I mean, being able to receive a plan at speed
- 10:52and knowing that it's a quality design that took in all those variables because
- 10:57the suite of software is already,
- 11:01to your point, already has that kind of framework that it will ask and analyze those questions.
- 11:12Preemptively in the process of actually developing the plant.
- 11:14So one thing you got to look at from a client standpoint and why this brings
- 11:17value to the engineer or their client
- 11:19is in land development especially, some people don't realize the minute they
- 11:23come to you with a property, they're putting money down.
- 11:26There is a financial ticker on that project that minute.
- 11:30So if it's a bad project, they want to know it because they want to limit that
- 11:34risk and they want to get out. If it's a good project, then let's get it done faster.
- 11:37But they can lower that financial risk and they appreciate engineers to do that.
- 11:43And then we'll go back to those engineered over and over absolutely so
- 11:46how is is there any built-in interoperability with
- 11:50other bentley softwares or is this more of a standalone product
- 11:53that given the fact that it does deal with data collection
- 11:57and ai is it is it a it doesn't communicate with other software suites or is
- 12:02it its own standalone program yeah it's fully interoperable not only with bentley
- 12:06but we're also with autodesk with cesium with google esri yeah all the things
- 12:10you might need as an input right if we're doing the site signal you might need
- 12:13the footprint from the architect.
- 12:15That might come from a reddit file yeah you might get the roadway
- 12:18design from an open roads you know kind of the tools
- 12:21or you know so we gather that information we make
- 12:24it part of the project we call that you know informing the
- 12:27digital twin yes your digital design digital delivery
- 12:30because this is all working in 3d as well yes so
- 12:33it goes into does the the site planning and and
- 12:37the drawing production all of that is done in 3d and automatically creating
- 12:40the drawing so this is very powerful in that
- 12:43topic so when it comes to digitally deliver not only yeah
- 12:46a lot of clients started to ask for the models right it has
- 12:48a whole bunch of exports as well to other formats to ifc for
- 12:52example that's required or two other formats so it
- 12:55kind of really allows to cover the whole workflow yeah from
- 12:58the concept through the final design through the digital delivery drawing production
- 13:02all those stuff right the whole gamut in one package and that's very very powerful
- 13:06yeah An example of that is one of the biggest pains inside the site design industry
- 13:11is creating a design and then having to go to a third-party hydraulic design software,
- 13:17whether it's OpenFlow products or like Autodesk.
- 13:20Everybody loses data and accuracy.
- 13:22Well, we decided when we started building this, the OpenFlow technology is going
- 13:26to go inside of this tool.
- 13:27So if you don't leave the tool to do storm drainage design, it's right there for you.
- 13:33Makes it for a more clean delivery. Yeah, one-stop shop, pretty much everything
- 13:36right there, completely interoperable.
- 13:39So it's definitely, you know, you don't have to worry about,
- 13:43to your point, any data loss or any problems with shifting over to other softwares or what you have.
- 13:49For firms out there who want to learn more about the product,
- 13:53where should they go on the Bentley site?
- 13:55They could type in Bentley OpenSite Plus into any good search engine.
- 14:00They'll find reference to that. Currently, OpenSight Plus is available in early
- 14:04access, only a limited number of.
- 14:07Bentley clients we're keeping that small because
- 14:10we want to work with firms that are really interested in and
- 14:14working with us and yeah adding this and and developing this we've
- 14:17been working also with our first mover advantage clients so i'm going to name
- 14:21a few so pannoni hdr the video speed burns and mcdonald and others that have
- 14:26allowed us to each to talk about them yeah openly and then we're developing
- 14:30that fantastic piece of software and we want to have a bit more people to join us to be engaged with us,
- 14:36especially large firms in North America, involvement sites, so I'll be a perfect, perfect partner.
- 14:42Fantastic. Well, I would encourage any firm out there in land development to
- 14:47check out OpenSite Plus from Bentley.
- 14:50It's a good example, a great example of how AI is transforming the engineering
- 14:55industry, how it's increasing productivity and allowing firms to do more with
- 15:01less time, and a lot engineers do what they do best,
- 15:04which is design and deliver for their clients.
- 15:07So, Francois, David, thank you so much for joining us.
- 15:10I hope you have a great conference. I hope you get some people,
- 15:13a lot of people at your booth this week.
- 15:15And I also appreciate you sponsoring the podcast programming this time.
- 15:20It's always great to have Bentley part of the program. Thanks for the partnership.
- 15:24And thanks for having us. We appreciate it greatly.
- 15:27Yep. Thank you so much. And again, this has been Engineering Influence,
- 15:30a podcast from the American Council of Engineering Companies.
- 15:33Coming to you from our 2024 Fall Conference, sponsored by Bentley.
- 15:37And we'll see you next time.
- 15:38Music.