AI Is Rewriting BIM & CAD! Qonic and Raven

Episode Summary

AI is moving from chatbot demos into the actual substrate of CAD, BIM, and AEC engineering workflows. In this episode of AI Across the Product Lifecycle, Michael Finocchiaro sits down with Chloë Guidi, software engineer at Qonic, and Moritz Rietschel, founder of Raven, to unpack what changes when AI stops being a bolt-on and starts driving modeling, data validation, and design decisions inside engineering software.

The conversation covers AI-native versus AI-augmented applications, the economics of LLM-powered software, MCP and bring-your-own-model patterns for enterprise AI, why hardware and BIM may follow coding into their own "ChatGPT moment," and where open standards like OpenUSD fit in the emerging AEC data stack. Both guests argue that startups with modern data architectures can outpace incumbents on digital transformation — but only if the tools connect cleanly through open standards and quality data.

Qonic is building a modern, cloud-based BIM platform with its own solid modeling kernel, aiming to make BIM more accessible, performant, and data-rich. Raven is building AI-first workflows for complex tools like Rhino, Grasshopper, Revit, Tekla, and Archicad, helping users work across fragmented design environments.

We discuss:

• Why the first ChatGPT moment felt immature but still important
• How AI coding tools are changing software development inside startups
• Why AI-native applications are different from applications merely “built with AI”
• Where AI actually sits inside CAD/BIM workflows
• The cost and business-model challenge of LLM-powered software
• Whether engineering and BIM are heading toward their own “OpenAI moment”
• Why young engineers should not ignore AI anxiety, but should sprint toward the tools
• How startups may accelerate digital maturity faster than incumbent platforms
• Why open standards, data quality, and tool-to-tool workflows matter for the next phase of AEC

This is a practical look at how AI is moving from demos and copilots into the actual workflows of design, modeling, data validation, and engineering decision-making.

Timestamps

00:00 — Intro: AI Across the Product Lifecycle 00:46 — Chloë Guidi introduces Qonic 02:26 — Moritz Rietschel introduces Raven 03:30 — First reactions to the OpenAI moment 05:46 — How AI changed software development at Qonic 06:53 — Moritz on early LLM experiments for CAD reasoning 09:02 — How AI is changing team workflows and code review 10:14 — Where AI helps, and where humans still need to check 12:02 — Where AI sits inside Raven and Qonic 12:48 — Raven as an AI-native CAD workflow application 14:14 — Qonic’s vision for intelligent BIM tools 15:46 — The economics of AI software and LLM costs 16:36 — MCPs, bring-your-own-model, and enterprise AI accounts 19:08 — Identity, traceability, and MCP limitations 19:45 — Will engineering have its own OpenAI moment? 20:45 — Why coding got the first AI breakthrough 23:10 — Mistral, engineering AI, and domain-specific models 24:45 — Digital twins, OpenUSD, and feedback loops in BIM 26:38 — Open standards and tool-to-tool interaction in AEC 27:51 — Advice for young engineers worried about AI 31:14 — How to convince AEC decision makers AI, BIM, and digital twins are connected 34:12 — Digital maturity in AEC customers 37:18 — Why startups may accelerate digital transformation faster than incumbents 38:29 — Raven, Rhino Inside Revit, and AI reducing workflow friction 39:59 — Qonic, accessibility, licensing, and better BIM data quality 42:10 — Where to meet Qonic and Raven next 44:01 — Wrap-up and sponsor mention

Please don't forget to click on this link from our sponsor AWS for access to an exclusive webinar! https://pages.awscloud.com/awsmp-gim-yngd-webinar-aim-enterprise-ai-and-data-leader-panel-lt-panel-1.html?trk=730a334a-28e3-4e91-960a-fc94de422926&sc_channel=el

---

Full Transcript

Speaker

And we're live. Uh welcome once again to another uh podcast of a across the product life cycle. U I'm Michael Finnecker. I'm your host and I'm joined today by Moritzel. Did I get that right this time? I think I did. Uh of uh of Raven uh which is a really cool uh CAD package. and also Khloe Gwidi of Khanic which is a well Mors you guys are in um Austria and Germany if I recall and uh um Khloe you're in Belgium so where it's a European uh European edition right now um so why don't you guys introduce yourselves tell me about um KIC and Chloe a little bit.

Speaker

Uh yeah so hi I'm Chloe. Uh I've been working as a software engineer at KIC uh for about four years now. So I started there in 2021 and that's also uh when Konica was founded. So I've been there since the early beginnings. Uh before that uh I worked at another company which was uh or is still uh working on a cat platform. So I've been

Speaker

in the industry for about nine years now as a software engineer. So that's been very interesting. Um maybe a little bit about Connik. Uh I don't know if everyone knows what it's doing. Um so um yeah at Conic we really believe uh that we should uh democratize the BIM process because yeah the traditional tools they require way too expensive hardware and it's really heavy software so we really started from scratch uh to build a BIM platform um that is cloud cloud-based um that is uh modern and userfriendly um so uh we really started from scratch to build that uh such that you can uh or everyone can model freely that they can view massive models even in the browser performantly uh and that they can also enhance the data uh they can uh verify and validate the data so all of that uh in one platform so that's what we build at com

Speaker

yeah and I think you guys wrote your own graphics kernel too which is really cool

Speaker

yeah solid modeling kernel actually yeah

Speaker

yeah it's really cool Jacob showed me spinning run around and get me a a bit sick, but that was really cool. Um, so tell me more. It's about Raven.

Speaker

Yeah. So, uh, I'm Mitz. I was I'm one of the founders at Raven.

Speaker

We first brought out Philip and Max also. So, there's three of us.

Speaker

And we first came out with a product a year and a half ago. So, a lot younger than therefore also we started out with a product that was AI first right start. So basically we had um this insight that a lot of the tools that are out there are um really useful and very powerful but uh quite hard to like use and manage and um you know get the skills to do to use them. So when we started out with Raven it was really uh supposed to help you use uh complex tools like Rhino and Grasshopper more efficiently and easier uh and and quicker. And so now we've built out this platform that's like um very very good at you know supporting your existing workflows with AI tools. So they work across like the term plugins everything that I know grasp connects to Raven can help you with. So connect to Revit or tech structures or archive pad and all these other things. So that was that's good.

Speaker

Awesome. Um I always start with the first question about um looking back. So, you know, we we've sort of had a we've had a moment, right? The open AI moment. There was before 2022 before everybody saw this chat window and after um were you guys skeptical or bullish when it when you first saw that? Did you think like, oh my god, the world just changed or were you like, I don't know, this is a bit of a fad. You know, either of you can pick that one up first up to you. I guess at first um it was uh I think I would say I was curious more so than bullish or think it was a fat and then it was very bad at first. I just distinctly remember that it would get all these things wrong and there were all these funny ways of like making it tell you some something wrong or like you ask it any kind of like factual question and it would just like hallucinate them something. So it was fun. I think at first was fun. Yeah.

Speaker

How about you Chloe?

Speaker

Yeah. I think for me it was a bit similar like I was intrigued so I really wanted to try it. Yeah. Especially as a software engineer I always had the feeling okay something major would come someday. So was really yeah you had this feeling like okay maybe this is it. But indeed yeah while trying it you quickly saw the limitation limitations especially for coding. Yeah we realized okay it's it's too soon. Um, so we could use it for for text generation and and stuff like that, but uh other than that it was a bit too soon to really start using it in in the company. Uh, but yeah, we kept an eye on it.

Speaker

No, but I mean did you guys see that this did you see it as something as a game changer from the beginning or you were like well you you already said you were sort of a wait and see, right? Because it was so immature and I suppose it's been such a surprise. the the the the rate at which it's improved has been uh really spectacular. Um and uh you mentioned, you know, you're you're a software engineer, Chloe. I I I'm imagining that it's changed completely now in 2026, the way you develop software. So that that's sort of the second section is how how has it changed in terms of the tooling and also your behavior as a programmer? How how has it changed uh for you at at KIC?

Speaker

Yeah. Yeah. Exactly. In the beginning of of this year, I think everything changed. Yeah. You realize okay the yeah the complexity of of the task that it can do has has gone up way more and it can solve tasks that would take us a couple of hours and that's the point that we also realized okay we have to really start adopting this in the company. So that's when we um started to use uh cloud code for software development. Um so it's been working really well and and it really accelerates our process. Um but we still have the feeling like okay we we are creating a very complex BIM application. Um so

Speaker

it really helps us but we still have to be very critical about the outputs because we need to develop something that is performant and scalable and yeah it's not like a a small tool. So um we have to build something that is very good and so it helps us but we also have to be be very critical about the outputs.

Speaker

Exactly. Yeah. Um same for you Morris or has it been a little different? Um so I mean before I get into maybe uh you know building the car that's like uh for me two and a half or three years ago I was um in graduate school so I was actually doing a lot of research very early on

Speaker

and back then um I was I would build very complex systems to like get to do 3D reasoning and to model cat and do all these things and so uh I seem to remember um in early 2024 so well over two years ago um I built like a first uh system that would bring that lens into into CAD. Um and it was really complex and the results were very very like you know you run it 10 times and before you see anything that looks like something. Um but that was interesting because I I could kind of like guess at the future a little bit you know like at the time I was just poking around at this weird new thing that's an LLM and like trying to find um strategies to elicit the kind of thing that I was interested in which is CAD CAD work and and reasoning. And so uh I could watch this progression of like every new model that came out and especially the job um in late 2024 with the using models with like 01 preview became available. It was like it was like a giant leap for everything that I had done. And then so much of the scaffolding that I had built would just collapse into the model and I thought that was interest. So by late 2024 I was like a full viewer because I had seen sort of all the work that I had done just sort of collapse into the model and model would do it on itself and that was a big big change. So I would say you know in 2023 um or 2022 when we first started interacting with HP and you' ask it casual questions you get them wrong and that was kind of an early moment but then within the year 2024 I saw it just explode in its capabilities and so much work that I had done just months prior they become kind of unnecessary that you know then I was like okay this is like it's not stopping anytime soon and it's going going to go pretty far uh which still doesn't mean that you like you know write software so much because I guess I wasn't interested in it using it to write software. I was interested in seeing sort of the capabilities work in CAD directly, not not to make CAD software. So that's sort of the world that my mind was at. Um two years ago,

Speaker

how has it changed the way you guys work as a team? Because I'm trying I'm trying to imagine there you must have agents showing up now for scrum meetings, right? I mean since they're the ones actually doing the code. How how has it changed? I mean, are you guys both using um agile or or waterfall or some um some kind of mashup of the two? And how is AI changed that? How how is AI

Speaker

how are you putting guard rails on it as a developer and still leveraging it on a daily basis?

Speaker

Go ahead.

Speaker

Well, I mean um it's a very interesting question and it's also something they have to continuously re-evaluate, right? So,

Speaker

not just as the models and the harnesses change, but also um maybe you learn things using them, you figure out what works and what doesn't. And I guess I'm just reading a lot more code than I used to because uh there's so much more code being produced, but now um like Chloe said, right, you need to be very critical um of the output and and really evaluate it thoroughly. So, that's like a big change that I see. Um it's a lot more review uh which is also a lot of work. Yeah. Yeah. True. Yeah. It's something for you though.

Speaker

At the moment, it's not like we we really let it just solve tasks without us looking at it. So, that's something for example, we had a look to to incorporate it in in our exception follow-up. So when there are issues uh coming by customers that that are using it um to have a look like okay this is uh something that the customer was trying um maybe uh it can just solve it by itself because these are sometimes small issues uh but then we uh quickly notice like okay it's not always solving the issue correctly and or it's maybe solving some no reference exception to maybe not go too technical but yeah that's not the actual issue. So, uh that's when we realized okay we should use it a bit differently. Um so now we are using it to make summaries and to maybe assign the correct person to look at it and to propose something. Uh but then someone still really checks it and has a look whether it's actually correct and but also there it really helps us to to remove some some work that otherwise had to be done manually. So there it really helps us. Uh but we still have to interfere ourselves.

Speaker

So it's good to get rid of some of the mundane stuff and yet you still have to keep an eye on it because

Speaker

it could still give you a pink elephant once in a while which you don't want.

Speaker

Um that's cool. And I was um that sort of leads into the next section is um I think that a lot of people that are not very close to the startup world or close to engineering don't realize how much AI is a part of everything and and so I'd like to ask you guys like in your software in Raven and in KIC where does the AI sit? Is it at the user interface level in terms of you know there's a chatbot and you're asking it to do stuff? Is it part of is there a foundational model where you've taught it how to build a building or to you know draw do Rhino um or is it the entire DNA maybe you've also put it all through up and down through through the stack where where is AI actually implemented in the software how in other words how's the DNA of software in 2026 changed from the DNA pre2022 or even 2024 because it's really the cloud code and I think it was anti-gravity might have been the first like pretty awesome agentic coder and then quickly replaced by clock. Anyway, I'll let you guys answer. Morris, you seem to have a want to talk. I always um Okay. I mean, look, we built a software that was from the start about um interacting with AI and so obviously it's part of every every bit of the software. Um what I think is is interesting is that there's kind of two ways to deal with coding um models and then one of them is like oh I am making the software already now I can code with um agent so I can make the software faster right which is like very good that's what's happening everywhere everybody's going to be more efficient write more software but the real like to us the real thing that changed or like the thing that we are seeing and how we think about this problem and because we're such a young entry I guess also space is we are working with those models. We're not using them to like build a thing and then ship something that has no AI in it, right? Uh that's kind of like we're not just building something faster than was already possible before, but we're just doing something completely different. Like the product that we do makes absolutely no sense without AI. So yes, AI is at every level. And I also think that that's just a different kind of category, right? It's like an AI application and not an application written with AI. So um that's just what we are like our DNA is uh making AI work for you in your design workflows and we have AI at every level.

Speaker

How about for and KIC?

Speaker

Yeah. Yeah. I think um yeah our point of okay we are not not an AI application but I think our point of view is a bit similar that um we really want to embed it like in the product like not um just one one chatbot or or just for example other tools they use it to generate some 2D images or we really want to go further than that because yeah architects they we we believe that they don't spend their time modeling but they lose a lot of time in doing quality checks uh doing class detection checking the quantities are right.

Speaker

Um so if we can incorporate the intelligence really in the whole application and in the tools while they are modeling um then that way they can they can solve and and win a lot of time um because nowadays yeah they have to model then build then then fix it and then the fixing is always uh afterwards. So if we can incorporate that in the tools they can a lot of issues can be prevented. Um and this is what what we want to do with comic is is to make these tools intelligent enough to already take into account uh all the compliances while while modeling or to already say without someone asking for it maybe this quantity is not correct. Uh have a look at it. uh and this way uh this can really help uh help the users uh to to get better models and get better uh quality of the models.

Speaker

Another thing I've been thinking about a lot um because of all these changes is also the the economic model, right? Because it's expensive. I mean cloud code I don't know how much it costs me more than I even want to say online uh every month. Um how how are you guys dealing with that? because that's as a startup using AI that's definitely a consideration like how are you going to offset the cost of the cloud because obviously these solutions are running on you also have to pay AWS every month and you've got to play a cloud cloud code bill how are you guys leveraging that or you know you doing um bring your own LLM for users so so that they can get some of the things or thinking of doing your own financial model so you don't have to pay anthropic for every single click I mean how are you trying to rationalize that in in your products respectively I'll go first.

Speaker

Okay, go for it.

Speaker

Um, yeah, that's a good question. Um, so I think we definitely um are also checking MCP. So I think at that point they would use their own uh accounts or would be able to use their own accounts. Uh but for the other pricing I think we are still in are still developing um like more the the usage of of the LMA as well. So I think for pricing wise we still have to investigate a bit how we will tackle it but it's definitely a relevant question um how to how to set the pricing.

Speaker

Yeah. And Morris, how about you?

Speaker

U obviously very central to our business case um and offering. Um there's a couple different things. Uh one of them is uh we don't have uh the ability that you can bring your own keys as a normal consumer but as an enterprise can. So uh because enterprise um have enterprise of you have these contracts with like Azure OpenAI or um Google Gemini enterprise contracts. So they can do use that. Um but on the other hand like one thing that makes Raven very compelling to most of our users is that it's very easy to just set up and use right. So don't need another subscription. You don't need to hook it up or something. You can just like download it and and get get in. So uh for us like maintaining that ease of use is super important. And um I think there are a couple things to consider like the way that it works right now is not necessarily going to be the way how it works in the future. Maybe um the adds that are really expensive right now you will get the same level of intelligence cheaper or through an open source model or something like that and maybe eventually you can self host on a you know good computer the model that can achieve the stuff that big models R5 do right now especially in the last few months you see we've seen a lot of really cool releases from Google for example Gemma 4 stuff like that um where you can really see that they managed to scale down and maintain a lot of the reasoning capabilities they lose a lot of the factualness but we don't really mind because we're doing the cat stuff, right? So, um, as the field is developing, I think, uh, we're just like we have a solution that's worked for us right now, but we're always monitoring sort of how that changes. Um, and also how sort of the spending structure of AI tools and and firms in the field kind of adapt. Uh so yeah, hard to say what the future brings, but I think there will be a lot of different models and lot of different ways to leverage tools like Raven and um I think a maybe less of a concern in in construction. Um but nonetheless, I think another issue is the lack of traceability and identity management with MCPS. I mean you mentioned MCPs Chloe but there is no protocol for or standard for tracing the who used the MCP and when and what data they used right I think that uh there's something fundamentally flawed in the current implementation that that needs to be fixed in terms of identity and authentication um I wanted to ask too so you know we've now four years into this AI revolution right since open AI uh opened the floodgates Um I I I think I may have seen an open AI moment in the manufacturing world at prove it when I saw them deploying entire factories with OT with using Kubernetes and Git within minutes and I was just blown away. I'm not sure I've seen that engineering. I don't think we've really had a before and an after moment and maybe we won't. Um so do you guys think that um we're going to have an open AI moment engineering and or BAM and uh what do you think it's going to require to get there? it because and and do you think it's like six months away, two years? I mean, it's hard to say, right? Because it's everything's changing all the time, but it feels like our jobs and our industry is hard, right? CAD is not easy. And and so I'm wondering how you guys uh what you how you react.

Speaker

Can you guys hear me?

Speaker

Oh, yeah.

Speaker

Okay. Sorry, I just thought the audio cut up. Okay. So, um I think it's going to happen uh because but the reason why you haven't seen it happening is also because nobody has had a focus on it. So, um it happened in coding and you know like I mean it's still ongoing but um and a lot of that has to do with the focus um that the labs put on coding. So, um there are many different things you can do with the reasoning model to uh like you know train it on on different tasks but it's very easy to do it on coding tasks because they're verifiable. You can uh you know run a reinforcement learning loop over c a set of problems. You can judge whether a code is better or worse than another code using in some way anyway. So there's a lot of style questions and evaluations that are easy to do for code. We could do them for engineering specific engineering tasks. It's just that the focus hasn't been there. But I think in the last six months you've seen a lot of money going into sort of physical AI startups like Bezos has this like entire like projects with like billions of dollars in funding. So I'm not saying that they're going to like solve CAD or something, but in the past year the focus has been elsewhere or like two years. For example, OpenAI didn't put such a big focus on coding. And that's why Anthropic um has sort of like outgrown them, right? Because entropic went all in on coding. they don't really mind all the other things that OpenAI is still publishing like oh we found we solved a new uh like an existing other problem or that they're interested in like protein folding and all these like very broad research terms whereas at topics just sprinting forward with just coding I mean they're not even good at vision they just do coding but it really paid off for them because they were they picked a wedge handle and went all in and so now everybody else has to be great at coding because you're compared to that rugby um and that can happen for any other domain. So what we've seen with Gemini is that they started focusing on SVG generation a few months ago which I think is like a step into the CAD world. Of course they use it for like web interfaces and stuff now but structured SVG is really 2D CAD right so going in this direction um I don't know where it comes from I don't know if it's going to be one of the existing labs or something but the focus on engineering will happen because people realize that there's this big market there um and it will be different like the AI will be will have very different capabilities from what we see right now um it's just hard to know you know where it's going to come It's interesting because um uh you guys probably saw that last week MEAI, which is an Austrian uh surrogate modeling startup that I've actually had on my podcast already, uh they were acquired by Mistral.

Speaker

And that's especially interesting because Mistral is providing the AI backlane for Daso Systems uh portfolio as well as uh contact software, the German PLM. So it seems that at least Mistral is trying to get that wedge in terms of engineering which is super interesting. And the other thing you made me think of Moritz is that um I saw a post that someone was reading between the lines of what Jeff Bezos was talking about in terms of Prometheus because that's what you were referring to and he actually is talking about CAD. So is a little, you know, I'm not sure what's happening there, but I think that that whole open USD, you know, because you're trying to model the the digital twin of the factory closed loop optimization, they might actually be looking also at uh at that CAD foundational model too. So it's going to be an interesting year, right?

Speaker

Yeah, absolutely. Absolutely. I think I mean it makes a lot of sense if you think about um running these loops where you have the agent model something and you can evaluate it and then improve like you can do that for path. you can have it model uh you know solve an engineering problem and then run it run some simulations on it and something like that. So Emmy M is like it makes sense to build these loops. Question is like who's going to do it in what industry for what vertical but it's totally possible with what they're doing today with coding you could just map it to a different discipline. Yeah, and Chloe, I'd direct that to you too. And and I would think about um the other big acquisition last week was Twin Thread being acquired by Aviva. Twinthreaded was already a partner of Aviva, but the outright acquisition came as a bit of a surprise. Um which signals that OpenUSD is really becoming a big deal. And I'm wondering um in the so you know OBSUSD is all about building digital twins and having a feedback loop that's in more or less real time. Is that something also the BIM world is looking at? Are you guys trying to get to the point where you design the building, the building gets built and then as uh the building's being used, feedback's coming in and know well maybe the next time we design a building that in that similar area for that same customer, we might need to make these other changes. And so you're getting sort of a feedback loop. Is that something that could happen? BIM or or not?

Speaker

Uh for me as a software engineer it's maybe a bit difficult to to reply uh on that question. Um I guess so I think um yeah if you have all this information and and now you can you can incorporate that uh in your flows and and learn from it. I think it's definitely something something you should be able to do. Um is it something that that we are looking at? I think I think not yet. Uh, I think we're not there yet. Um, I think there's a lot more things that have to be done maybe first. Uh, but yeah, that's that's just um that's just how I look at it. I would have to check with the with the real uh like the more product people to to know what they think about it. Um, but yeah, I think it should definitely

Speaker

Well, I didn't give you

Speaker

I didn't give you a chance answer the initial question about just the open AI moment for engineering or for BIM. Do you think we're there or you think it's coming or you think it'll just there'll be many moments and not just one? How do you look at that?

Speaker

Yeah, I I think like like Moritz was saying like I think definitely something is changing. Um also because I think for for the BIM world then you have like we are also focusing a lot more on these open standards. So you have also much more the data becomes more open and compared to the tools where like all the data was kept inside. Um so this also makes it a lot more possible to to really train on the data and and to make the yeah make the data more more more um yeah more that you you have more information so you can also build better tools around it. Um also I think toolto tool interaction becomes possible that way and I think that also becomes very powerful that you can scan a PDF and uh get all the specifications out pass that to the BIM BIM tool which can also uh leverage that. So I think yeah you start to see these possibilities and I think that the customers also start to see them. Um so that way I think bit by bit I don't know whe whether there will be like a really turning point maybe there will be but uh definitely something is is shifting that's something we really see

Speaker

and um I think that the in the demographics of the people that watch this show I have maybe 20 to 25% are really entry level um and I can imagine that those kids are having a bit of anxiety because they're like you know AI is going to take my job. Um, in terms of you as a software developer, clearly and you Morris as a founder, what kind of advice do you give to that younger emerging generation, the ones who are coming out of university looking for jobs? How what do they need to focus on so they're not immediately replaceable by an agent?

Speaker

I would I would always say like stay curious, try it out. Um, don't pretend like it's it's going to go away because I don't think it is. it is is not going to go away anymore. Um, I think as long as you as you learn to to handle it and and learn how to how to um yeah, how to say uh learn how how to use it properly and also I think maybe also be critical about that because you see a lot of things passing by but um it it doesn't always work like there's a lot of tools for example that say ah we can do this and that and then you try it out and it doesn't really work. So I think it's important um that you also see the the benefits of some tools but also realize okay some tools don't work. So um but you can only figure that out by trying it and um by staying curious. I think that's important.

Speaker

So fluency in the AI tool is super important, right?

Speaker

I think Morris Morris, what what would your advice be? Um I I I would say I get it. I get the anxiety and I also don't know if anybody really has the answer. Um yeah, that's all honestly. Um I do think that engaging with the tools is probably the most important thing. um if if tools are going to change your workflows that it might make you feel like they're going to take your job, but what's probably going to happen is that the the job and like the skills and the things that you actually do are going to be slightly different. So like Chloe said, instead of writing all the code, you're going to have to review a lot of the code. So it's super important that you create a coding and you understand the code, but the job looks very different, right? So I think that's true for a lot of areas and I don't see currently you know um at the current capabilities they are replacing engineers. I I don't see that. Um but I do think if you if that's something that you are worried about then you need to probably sprint to the front of of this wave and and and figure out sort of what this how do the skills change and what is going to be really important. Um yeah, so that's what I think and I know that it feels very um scary to recent graduates because they think that the entry- level jobs are the ones that are falling away. I also want to um remind people that young people are usually very good with new tools. They have very good like meta model with how stuff works. They can get experience. They are still very adaptable to new skills. So, um I I say embrace that as sort of your strength instead of um it being sort of your um opponent, you know, like AI is the thing that you're you can be good at. You can be the person that's good at this. Uh and that's going to be a different great answer. Thank you, Morris. Um, actually we do have a question in the in the chat from Janelle Pierce and she asks, "Uh, any advice on how we can better influence AEC decision makers to move beyond viewing AI, BIM, and digital twins as isolated technologies and recognize them as fundamental to the future operational resil resilience, efficiency, and competitive advantage.

Speaker

What do you think?" Yeah, I was checking Mo is maybe lost. He's not moving anymore.

Speaker

Um

Speaker

Oh, he's not connected. Okay,

Speaker

that was a long question. Um

Speaker

Oh, sorry. Well, I think she's just saying how can we influence decision makers so that they can realize that we're not, you know, these all these things are connected. The digital twin and the infrastructure and the eye. These aren't three separate questions. are the same question about improving resilience and efficiency and competitive advantage for uh for firms, right?

Speaker

Yeah, definitely. Yeah, I think um if you if you use it and you can enhance uh the quality of your model and and yeah, in BIM context of course um and your model can become the single source of truth. I think a lot less mistakes are made um for for construction world as well. So I think if you start using that and and you see that I hope the decision makers also uh see the added value of that um and then you can incorporate it but you can also start doing it bit by bit. Yeah we have customers that start using it for for one project and and then see uh if it if it actually works. Um so maybe with some proof of concepts. Um it's not like you have to immediately shift your whole company. It's also not not what we do. So you can you can start trying it out and um try it on one project and then um if that works incorporated further that's

Speaker

so it's really an argument of ROI then of just trying to make them understand the ROI that they're getting out of it

Speaker

eventually.

Speaker

I think I think that was would be the best way to work. Uh yeah, no one wants to use something that yeah is adding more complexity or um also I think using tools that really uh make it transparent what is actually worth doing what it's actually doing. I think that's also important and that's also something that we want to focus on that the output is is clear and that you don't have to uh spend more time on checking whether the result of the IO is actually correct but that you can quickly see what it actually did and that you still have the you can still grasp what it's what it's outputting.

Speaker

Well um I asked in the in LinkedIn I have not heard back from Morris. We hope that everything's okay with him and he'll join us. Um so then the I and it the the next question is more about digital maturity. Um like when you're working with the customers um and typically AEC is one of the least mature uh industries. I I usually think of digital maturity for companies being on a scale of like one to five where um five is the most um uh the most mature you know like adaptive uh digital twins and one is people are still using email and Excel most of the time. Um I suppose that uh most of the customers uh you're dealing with Hi Morris, glad to see you back um

Speaker

to open a new and like reactor studio as a new person somehow.

Speaker

Um so I was asking about digital maturity. So when I think about digital maturity, I think about um companies that are not mature that are using Excel and and and email still for collaboration and being like a one and I think of the mythical company that has all autonomous agentic digital twins, you know. Um where does most your customers sitting? Are they closer to one or some of them at three? Where do you see your customers today?

Speaker

Uh I have a question. I think um yeah sorry I'm just small jumping um uh we see everything um and it's interesting because what what we see sometimes is that people who are at very low digital maturity will use AI as like the moment where they do embrace it in a way you know that so even though they maybe have been very sort of low tech for a while now they sort of see that as like a leaprog opportunity right so I think that's interesting

Speaker

okay

Speaker

even for people like even for less people for example because it's unlocked on the like sort of skills network.

Speaker

Yeah, I was going to get to that part in the second half of the question but what about you Chloe? Did you have some feedback on uh on the maturity of the customers you you talked to?

Speaker

Yeah, I think for us it's also I think in general yeah we also see see a lot of different things but in general it's also low to medium I would say. Um but I think it's also related to the traditional tools that there are always like only a few experts in the companies that that can actually use these tools. Um so I think um

Speaker

by creating these modern tools um this this shift is also starting and and also through AI they they start to see maybe a bit like what we were discussing before like they start to see okay this can actually solve uh some workflows and solve some tedious errorrone tasks that we've been have having to tackle. Um so bit by bit this shift is coming. So um there's really a momentum going as far as we can feel. Uh so I think this digital maturity will will definitely improve. Um well my my thesis is that um companies that want to move that needle further to the right towards a five are much more likely to do so using startups that are powered by AI like KIC or or Raven if they're depending on Autodesk or or ZMANS or PTC or D. So to do that digital transformation I think 2 3 years later they're still waiting for it. You know even now they're waiting right because we AI has been out and all we're getting is more or less co-pilots which are so 2024 now. Um so I'm wondering have in your experience when when the customer sees all that they can do inside of Raven or Conic thanks to this AI is there a bit of an epiphany? Do they go talk to their colleague? Like, did you see this that what we can do if we, you know, had a more we're more mature and we didn't have data silos and the guys in engineering talked to the guys in manufacturing and the designers talked to the people in the field, you know, we'd actually be able to do business better. Have you ever seen that had an an, you know, a real impact on the business? Oh, that should go for you. Um,

Speaker

uh, yeah. Yeah, for sure. So, one big thing that I can mention here is Rhino inside Revit workflow. So, bringing Rhino stuff into Revit and stuff. So, uh, Rhino has built this Rhino inside integration and uh, it's very very good and very very hard to master it because not only do you know need to know all the stuff that you already need to know about Revit, but now you also need to know the stuff about Grass. So it's kind of like this o overlapping expertise field that's very small. Um but it's extremely useful because all of our customers in the AC space have these workflows. Um now with Raven you can generate new uh like files code that does this transformation and then also uh you can use other people. So you don't need to be the one expert that that like can author script themselves and use it, but you can take someone else's script and like ask a question about it and say even though the tools have been around, just being able to use them with Raven by your side is like a mass bottl. So that's one of the things where it's like, oh, you know, we've been trying to even the dig digitally mature ones like, oh, we've been trying to push this and push that and there's all this friction around uh creating new tools and you know, less data silos and so on. But uh AI is kind of like uh the oil in the gears that can like actually make the machine run how it was supposed to uh on the power on on like the PowerPoint slides from 2020.

Speaker

Nice. I like that answer. How about you Chloe?

Speaker

Yeah. Yeah. Yeah. Yeah. Similar. But yeah, I think for us not not only AI but like also the fact that they are using user friendly tools uh also helps. Like for example, we also believe in like the fact that like this licensing model of that you um can have unlimited users to to make it really accessible for for everyone in the company. For example, we had a customer that like only one person was using it and then bit by bit the others we really started like to see a ripple effect where more more c more uh employees of that company were using it. Um probably because it was also looking way way more user friendly for them. Um so that also really helps. Uh but yeah of course through AI uh it becomes also a lot more easy to use. Um but other than that I think um also through I I think I mentioned before but that like through leveraging um this data quality and um you have less like garbage in garbage out than than with the classic tools and then they also feel more uh uh more maturity of the of the model and of the of the product. Um and therefore there is more interested interest to also really start using it. Um, so I think I think that really helps to

Speaker

if it becomes more reachable to have also more digital awareness throughout the the industry. Yeah.

Speaker

Oh, I don't

Speaker

I think

Speaker

Oops. How's that? No, I said I I think that validates my thesis that startups are a better place to to uh to count on digital transformation uh particularly if you want an accelerated uh version of that. Right. Um that that's about all I had in terms of questions today. I want I was curious um in terms of uh people wanting to meet you guys and wanting to learn more about your solutions where where can they do that over the coming months? I think uh for us

Speaker

go ahead

Speaker

I'll go first instead. Um for us I think so some important ones for us is uh in October we will be at the building smart summit in Tokyo.

Speaker

Nice.

Speaker

Uh there we will also have our own uh comic talks uh with some dedicated sessions about comic uh for who is interested

Speaker

and that's cool.

Speaker

Yeah. Yeah. Yeah. Yeah. But we also have something a bit more close by. Uh so here in Belgium in G uh in our hometown we will also do a conic talk event uh 27th of October. So

Speaker

everyone is definitely also welcome to join that one.

Speaker

Oh looking forward to my invite.

Speaker

Yeah yeah yeah you will.

Speaker

How about you Mars?

Speaker

Yeah. So uh we have on our website we have like a list of things where you can meet us. I'm not going to go through all of them, but um I can already say in the next um next week and on Monday we are putting on a Rhino user meeting in New York together with SOM and McNeel. So we love the Rhino user meetings in Europe. Um they don't really have it in the US. So

Speaker

uh we asked the McNeel guys about it and they said you'll have to make it happen yourself. So on June 1st if anyone's listening is in New York, you know, come by. Um awesome. And then we're going to be at a a couple other um you know conferences. is going to Acadia which is in September um in the US issu which is in Italy. These are kind of more academic conferences. And then um uh a tech of course in in Europe again.

Speaker

They're doing um the CD fam and DC.

Speaker

We're not going to be in DC but

Speaker

Okay.

Speaker

Well, um I hope that uh

Speaker

Sorry. Go ahead.

Speaker

No, I mean we always uh we have like if you want to meet us, go on our website, scroll down, there's like a list of events and you'll find something I hope near where you are.

Speaker

Right. I'm I'm still uh planning to do one in um another event in Munich before the end of the year. So, I hope that I'll see you guys there. Um and uh once again, thanks to the audience for joining. We hope we had a a good audience today. Um a shout out to the sponsor uh AWS. Please download the white paper. Um I think there's actually a video associated with that. I'll put the link in the in the comments. And um once again, a big thanks to Morris and to Khloe for for joining me today. I really appreciate your openness and and all your answers. Thank you very much.

Speaker

Thank you for hosting.

Speaker

Was very

Speaker

Thank you very much for hosting us.

Speaker

And so uh stay tuned. This will be published up on YouTube pretty soon. And um and we'll talk to you everybody later. Thank you very much. Have a great day. And oh for

Speaker

folks that are watching and on LinkedIn, you know that in about an hour and 15 minutes I'll have another one of these but with supply chain uh startups Omni and Ban. It's going to be a really cool discussion too. So, thanks everybody. Bye-bye.

Speaker

See you.

Speaker

Bye.