DesignBiscom
The designer used Biscom to generate high-fidelity visualizations from quick sketches to create 'eye candy' for stakeholder presentations, which helped secure project approval and accelerate the decision-making process.
“I would normally make quick sketches to then put normally into BISCO and probably visualize that really fast in terms of like creating that eye candy for a good presentation where we want to get this project approval and moving forward.”
ManufacturingQAD Redzone Champion AI (Line Lead Agent)
An AI agent analyzes production data to predict the top two issues a line lead will face and their impact on OEE, allowing the human to focus on preventing disruptions rather than collecting data.
“They are actually able to use AI to try to figure out, how is that production line going to run? What are the top two issues I'm going to experience? We have an agent called the line lead agent, champion AI agent. And that agent is working with the human line lead to help that person, think through what are the top two issues you're gonna have? How is that going to impact your OEE?”
Engineering & simulationMaterialise AI agents / APIsone day to build a new application
Materialise wrapped its SD case (MagicScript) in APIs and fed them into AI agents to automatically generate automation scripts for custom applications, reducing development time from hours/days to a few hours.
“we put that as an inputs into AI agents. And then you can just ask to build automation scripts. And there is something, you know, measure script, something you do in a couple of minutes to a few hours. So basically we have already been demonstrating that in, in one day, we can build a new application with customization.”
Running the companyIBM Watson
PSA Systems is rolling out IBM Watson as an enterprise AI platform to automate internal workflows, specifically starting with project management and accounting agents to help staff work better and control the business.
“we're rolling out IBM Watson into the company as our AI platform, company wide. So our 1st 3 agents are done and we're keeping generating agents … helping the staff work better and control the business better.”
Engineering & simulationFlow Engineering
AI agents automatically verify hardware designs by detecting how changes in CAD, simulation, or firmware break specific engineering requirements, aiming to reduce development time and cost.
“our AI systems engineering agents can listen to change in CAD, in simulation, and get in firmware, and then automatically realize that a change in one place affects a change in another, which effects a change in another, that breaks a requirement.”
Marketing & salesAI cart builder90% of member service inquiries are now self-serviced through AI
Thrive Market uses an AI-powered cart builder to help new members build their grocery orders based on their data and health goals, which has transformed the onboarding experience for new customers.
“our 1st orders now, members can get help building a cart with an AI cart builder that's just transformed the way the onboarding happens for new members. Um, you know, 90% of our member service inquiries now are self-serviced through AI.”
Running the company
Duro uses AI to simplify the PLM onboarding and configuration process by having the AI read documentation and API specs to set up workflows based on natural language intent.
“AI is an obvious, you know, um, um, path to do that because you can tell the A model what your intent is and the A model can do the dirty work for you, read our documentation, read our API, understand the configurability options…”
Engineering & simulationSlack, Jira, GitHubbuilt ERP without software developers
The team used AI coding tools to build an internal ERP system without software developers, dispersing product management needs into Slack channels that fed into Jira and GitHub, allowing the production team to iteratively design the software they needed.
“we're doing by just dispersing the needs into effectively things like Slack channels, where we can backchannel things into into Slack, have that go up into Jiro, have that fall down into a GitHub repo, and have this iterative cycle where the team is truly designing the N ERP that is required for them.”
Engineering & simulationDuro Platform / AI Interface
Duro is developing an interface where users can ask an AI agent to build a plugin for a new CAD or ERP product, allowing the agent to generate a basic plugin in minutes.
“within our own platform, you'd be able to use our interface, our model and say, hey, I want to, I just purchased this CAD product or this ERP product. I dont see a plug-in for it. Can you help me build one for Dero … hopefully within a couple of minutes, you got a basic plug-in, you can be off and running.”
Logistics & fulfilmentMobidas
Zollner is piloting AI agents to assist with supply chain work and administrative tasks at its Silicon LVD5 facility in Europe to manage high costs.
“we started the pilot in the Americas and Mobidas in the Silicon LVD5 facility because we have high costs over there, and then we we we figure out how the agents, the AI agents, can help us in the in the supply chain work and the administrative task.”
Running the company
The company used AI coding agents to burn tokens and build a digital twin model of the manufacturing process, aiming to optimize details critical to volume production solutions.
“strangely enough, that's that's what I'm spending my nights and my weekends on right now is is, uh, burning a lot of coding agents tokens to build a, uh, you know, a digital twin model of all these things.”
Engineering & simulation
Touchlab trained AI models to interpret raw data from their tactile fingertips, enabling the detection of specific object properties like ripeness and texture.
“one of the really exciting things that we've been doing with these sensors is training AI models to interpret this data and we have some really exciting sleep detection that we've been we've been trying out with picking up ice cubes and picking up fresh produce, ripe blueberries.”
Engineering & simulationNVIDIA Isaac Simhundreds of thousands of simulations
The team used NVIDIA's simulation environment to train reinforcement learning policies for quadrupedal walking gaits, replacing hard-coded step sequences with AI-driven movement.
“In simulation, it is running many, many, many simulations, hundreds of thousands of simulations, and it is trying to hit those targets. … Over the course of every simulation, is essentially trying to hit all of those rewards. When it hits all of those rewards, then we have a functioning policy.”
Running the companyClaude
The team used Claude to compress long voice dictations of business operations into structured markdown documents, creating a legible knowledge base that connects to their data sources for inventory planning and ad analysis.
“I just was chatting with Claude, like, okay, hey, we want to do something that like gets us these goals … And what it ended up being is it's basically like a series of like 8 markdown documents. … Claude takes that and compresses it into like a formatted life, almost like a knowledge base.”
DesignDeepPCB (Cooper)
An LLM-based orchestrator named Cooper analyzes an underspecified PCB design file to detect missing context like net classes and decoupling capacitors, then automatically applies corrections such as changing layer types to power planes before routing begins.
“Cooper is just the LLM that sits on top of the system, right? Cooper is not an autoroter. It's an orchestrator of the DPCB routing engine. So now it's looking at your PCV and this unlocks a bunch of things because so many of our users, again, were frustrated by not being able to have the right configuration.”
ManufacturingSiemens10% increase in line performance, 13% reduction in waste, 7% energy reduction
Kellanova and Siemens built a digital twin of Pringles dough that continuously retuned recipes based on incoming material shifts to improve output and reduce waste.
“they had, um, sensors, lasers, cameras, feeding temperatures, humidity, protein, reading all into an AI model, um, plus a digital twin of the dough that continuously retunes the recipes as an incoming potato material shifts.”
Engineering & simulationAltium Requirements Portal (Engineering Assistant)
The AI agent analyzed a high-level email description of a wind-farm inspection drone to generate structured, well-formulated technical requirements with TBD placeholders for unknown values, enabling faster initial specification discovery.
“basically what you see now is the engineering assistant will now try to come up with actually well formulated requirements for this … It actually formulates requirements in the way that I and my company wanted these kinds of things to be formulated.”
Engineering & simulationThinkLabs AIinterconnection studies reduced from 6-9 months to 10 minutes or less
The company uses physics-informed AI models to automate utility grid planning studies, replacing manual trial-and-error with deterministic simulations that generate infrastructure solutions.
“studies that would take previously 9 months, as I mentioned, interconnection studies now takes us only a matter of couple of minutes, like 10 minutes or less…”
Engineering & simulationSiemens design software
Xometry embedded its pricing and intelligence engines directly into Siemens' design software, allowing engineers to receive instant pricing and partner matching within their CAD tools without leaving the design environment.
“we've actually now started to take our pricing and intelligence engines and embed them directly into design softwares. We announced a strategic partnership with Siemens. … making the manufacturing capacity accessible to all, but accessible right in the tools that they're working in…”
Engineering & simulation
Protolabs uses machine learning models to assess part manufacturability and simulate tool paths, allowing them to generate G-code directly from CAD files without skilled labor programming the machines.
“a lot of it is really around how we've built machine learning models to assess the manufacturability of the part and then to simulate the tool path that is required to mill that part in particular. … We then take that tool path and we generate the what's called the G code, which goes right into the controllers of these mills.”
Engineering & simulation
The team trained algorithms to distinguish between classes of litter and living organisms (animals/plants) by labeling images, enabling the robots to identify what to collect versus what to protect.
“So we do that through machine learning. We trained algorithms to directly distinguish between different classes of liter and we typically just label that to be trash. And other classes like animals or plants so everything that we need to protect down there.”
Engineering & simulationNo Magic VLA (Vision Language Action models)0 churn within deployed systems
No Magic uses VLA models as a secondary recovery workflow to allow robots to autonomously solve packing errors in real-time by interpreting visual input and generating action plans, reducing the need for human intervention.
“we have, uh, censoring technology that's, you know, with AI that's constantly looking for these types of scenarios and, you know, flagging the system to switch to VLA mode, essentially. And then the robot can drive down and solve with that input…”
Engineering & simulationSiemens Fuse EDA AI agent10x reduction in turnaround time; 10x reduction in token costs
The agent autonomously orchestrated the silicon characterization workflow, invoking tools via skill files and MCP servers to generate liberty files, reducing manual babysitting of jobs and handling failures automatically.
“we're able to reduce the turnaround time by 10 X and reduce the token cost by 10 X. So dramatic improvement and just one example, right? So across the entire EDA workflow … you can imagine that same kind of workflow … applying applying across the board.”
Engineering & simulationfoundation models (Pi0, Groot)
The team implemented and benchmarked foundation models like Pi0 and Groot by fine-tuning them on production data, but found them unsuitable for current production use due to poor performance and large model sizes.
“we've, we've done implementations of pi zero. Um, we've done implementations of Groot, where we, we take those models, we actually take our production data and fine tune them and then run them in our office and benchmark them. Um, and the reality is both because they're not performant and because the model sizes are so large, they're not suitable for production applications right now.”