Manufacturing the Future

"No one in hardware wants to be sold more software. They just want their parts to show up so they can build their hardware faster."- Shan Mohta, Co-Founder and CEO at Vendra (YC S24)Shan Mohta is Co-Founder and CEO at Vendra, a Y Combinator-backed manufacturing marketplace that uses AI to connect engineering teams with the right US-based manufacturers for custom parts across capabilities including CNC machining, injection molding, and 3D printing. Before founding Vendra, Shan spent six years as a product design engineer at Microsoft, Apple, and Skydio, working on products including HoloLens 2, Vision Pro, and autonomous drones. He joined Lisa Bounds and Ellen Cox on Manufacturing the Future to discuss why existing US manufacturing capacity is underutilized, and how better data and supplier orchestration can compress the timeline from design to finished part.In This Episode:Shan shares how his frustration sourcing complex parts at Skydio led him and his cofounder to launch Vendra after being accepted into Y Combinator in Summer 2024. The company pivoted throughnine product iterations in six months before landing on what became its central thesis, that hardware teams want outcomes, not another software tool. Vendra now builds a data layer across US manufacturers, tracking equipment, capacity, certifications, and tribal knowledge to route each part to the best-fit shop at any given moment. Shan explains why the US already has enough factory capacity and the real gap is the data infrastructure to use it effectively, drawing a direct comparison to China's hyperlocal, state-sponsored manufacturing networks. He breaks down the added complexity of aerospace and defense production, where compliance requirements, multi-step finishing processes, and cross-supplier orchestration turn a single part order into a weeks-long coordination effort that Vendra aims to automate into a turnkey solution.Topics:Pivoting from SaaS to an outcomes-based manufacturing marketplaceBuilding a supplier data layer across US manufacturing capacityWhy existing US factory capacity remains underutilizedAerospace and defense compliance and multi-step production complexityCompressing the design-to-finished-part timeline through better routingHow China's manufacturing edge comes from network coordinationDistinguishing real market signals from friendly encouragement as a founderWhy AI solutions without execution experience fall short in manufacturingDownload, Listen, and SubscribeApple | Spotify | YouTube

"They have a data problem that's wearing an AI costume. Because if you don't have accurate data, it doesn't matter what you do, you just do it faster." – Stephen HightowerA sharp wake-up call from Stephen Hightower, Chief Technology Officer at Harmar, who put a moratorium on development until his team fixed the data underneath. The problem he's solving isn't just data quality, it's the operational waste that builds up when three integrated enterprise systems don't agree with each other, manual workarounds become tribal knowledge, and spreadsheets quietly replace the systems people no longer trust. When your data is broken, AI doesn't fix it. It just does the wrong thing faster.In This Episode: Stephen walks through how he's rebuilding data confidence at Harmar, a manufacturer of mobility and accessibility solutions. He describes launching a Master Data Initiative using CRUD and RACI matrices to assign clear ownership for every critical data element across engineering, ERP, and customer systems. He explains why cycle time, quality, and unit cost are the only three metrics that matter in manufacturing, and how tracking integration error rates across systems exposes where data breaks down. Stephen also shares how his Lean Six Sigma background, starting in the late 90s at Lockheed Martin, shaped his approach to treating broken tech stacks as waste to be eliminated through root cause analysis and corrective action. He describes putting monitoring tools in place to gain insight the ERP system wasn't providing, deploying a digital worker to handle customer care calls so his team can move up the value chain, and using the financial close cycle time as a diagnostic for how well a business is really running. He's also candid about AI's limitations: he uses Claude for data analysis and engineering cycle time problems, but stresses that every AI output requires human verification because it will, in his words, "lie to you all day long."Topics:Why most AI failures start with a data problem wearing an AI costumeRunning a master data initiative using CRUD and RACI matrices for clear ownershipTracking integration error rates across enterprise systems to expose wasteApplying Lean Six Sigma to broken technology stacksFinding the hidden spreadsheets that signal system waste and broken trustUsing monitoring tools to gain real operational insight the ERP wasn't providingDeploying a digital worker for customer care operationsWhy financial close cycle time reveals operational healthThe "go find the spreadsheets" test for any manufacturing businessA conversation with Stephen Hightower of Harmar about why most manufacturers have a data problem wearing an AI costume, and how master data management and Lean Six Sigma applied to technology fix it from the inside.Download, Listen, and SubscribeApple | Spotify | YouTubeOr search "Manufacturing the Future" wherever you listen to podcasts!

"What used to take me two hours takes me about 15 seconds now." – Tyler MadsenA striking before and after from Tyler Madsen, Director at Madsen's Millwork & Custom Cabinets, who rebuilt their estimating workflow by feeding drawings and specifications into an AI bot that strips out irrelevant data and returns only what's needed to quote a job. The broader problem he and IT & Asset Manager Jason Bassett are solving isn't just speed, it's the operational drag that builds up when a manufacturing shop is still running on paper drawings, manual data sorting, and an IT person who becomes the bottleneck every time someone needs an answer. When information lives on paper, you get version control failures, field installers working off outdated drawings, and a team spending its time managing data instead of acting on it.In This Episode:Tyler and Jason walk through how Madsen's Millwork has digitized its day to day operations, from estimating to the shop floor to field installation. Tyler describes feeding specs and drawings into an AI bot that strips out irrelevant data for quoting. Jason covers the physical changes on the shop floor: big screen displays at every workstation now pull live drawings, replacing paper that created double data sets and made version control nearly impossible. He also gives field installers the same live job data on site, so a mid job change reaches them in real time, and describes building knowledge banks inside Epicor Prism pre-loaded with the questions his team most commonly brought to IT, so employees self-serve instantly instead of waiting on him. They're also candid about what isn't solved yet: getting skilled trades workers to trust and adopt AI day to day remains an open challenge, and they also flag hallucinations as a real operational risk that requires experienced human oversight to catch.TopicsCutting spec review from two hours to 15 seconds using an AI botBuilding Epicor Prism knowledge banks to eliminate the IT bottleneckAI owns data retrieval; humans own judgment, their workflow decision frameworkThe mutual checking model: humans verify AI output, AI challenges human decisionsReplacing paper drawings with live digital displays at every shop workstationField installers accessing real-time job drawings remotely during installationUsing client AI renderings as the engineering starting point for custom projectsWhy getting skilled trades workers to trust AI remains their biggest unsolved challengeHallucination risk in manufacturing operations and how to stay vigilantDownload, Listen, and SubscribeEpicor | ⁠Apple⁠ | ⁠Spotify⁠ | ⁠YouTube⁠Or search “Manufacturing the Future” wherever you listen to podcasts!

"Everything is basically a fact-finding mission. You're trying to figure out what you can do, how fast you can do it, and how efficiently you can do it." – Mike WargockiManufacturing leaders who scale fast tend to break their own operations without realizing it, not through a single bad decision, but through a string of reasonable ones: reinvesting in an existing plant one quarter too late, forcing identical equipment into buildings that were never built for it, or promoting an early team member into a role they were never suited for. The trouble is that most of these metrics and mistakes don't show up on a standard dashboard. Mike Wargocki has spent his career finding the operational truths that hold across food, biotech, and custom consumer manufacturing, and figuring out which numbers actually predict performance versus which ones just look good in a report.Much of this conversation draws on Mike's time at Framebridge, a company rethinking how custom framing is designed, produced, and delivered. Framebridge built a direct-to-consumer model that turned a traditionally complex, expensive process into something far more accessible, and has produced over two million custom frames as a result. The company has continued scaling quickly, bringing new manufacturing facilities online across the US to support both e-commerce and retail growth.In This Episode:Mike Wargocki, VP of US Operations at DINGS' Motion USA, walks through the operational lessons he built over a career spanning research chemistry, food, biotech, and custom consumer manufacturing. He explains why OEE stops being a useful metric once production gets custom, and why he shut down a limestone cave facility outside St. Louis despite its near-free refrigeration, once quality started slipping. He breaks down the real cost of forcing identical equipment across a multi-site network, how to time a new facility build against a real growth plateau instead of a projected one, and how his team planned peak season staffing at Framebridge without blowing the annual budget. He also shares where he sees manufacturing leaders waste money without noticing, why early employees aren't always the right fit for the next stage of growth, and how his team has used AI tools to help non-native English speakers write difficult workplace communications.Topics discussed:Why OEE fails as a metric in custom manufacturingThe real cost of forcing equipment uniformity across facilitiesTiming new facility builds against a growth plateauBalancing peak season staffing without blowing the annual budgetClosing a newly built facility over a hidden quality trade-offApplying a chemist's fact-finding approach to operations leadershipRight-sizing early team members as a company scalesUsing AI to help non-native speakers write difficult communicationsDownload, Listen, and SubscribeApple | Spotify | YouTubeOr search “Manufacturing the Future” wherever you listen to podcasts!

"We're common where possible and unique where necessary."A deceptively simple principle from Daniel Topp, VP of IT at TASI Measurement, that gets very complicated when you're managing ERP across 16 autonomous business units and integrating newly acquired companies on a rolling basis. The real question is how you operationalize it so it doesn't collapse under the weight of competing business unit priorities and leadership requests for customization.TASI Measurement is a global industrial measurement holding company with more than 1,000 employees, headquartered in Largo, Florida, operating through a highly decentralized structure where each business unit runs independently under the broader group.In This Episode:Daniel walks through the specific systems he's built to keep ERP standardization from becoming a constant negotiation. That includes a formal customization approval process requiring sign-off from key stakeholders before any deviation from off-the-shelf is permitted, and a 30-60-90 day acquisition integration playbook that separates non-negotiables like IT security tools from ERP decisions, which get evaluated through a risk and value heat map. He explains why data migration is where repeat ERP transformations actually improve, and why building a dedicated headquarters-level migration team, rather than relying on business unit staff, is what makes the process repeatable and scalable. He also makes the case that locking in structural decisions like chart of accounts early in a project is the difference between finishing on time and losing months at the end. On AI, his position is direct: having it available inside your ERP without a structured rollout plan is a liability, not an advantage.Topics:Why ERP projects fail at training and change management, not implementationFormal customization approval process requiring stakeholder sign-off30-60-90 day acquisition integration playbook and what's non-negotiable from day oneRisk and value heat map for sequencing ERP integration priorities across acquisitionsGold standards center of excellence and how it serves newly acquired businessesWhy locking in decisions like chart of accounts early can make or break a project timelineBuilding a dedicated, headquarters-level data migration team as a repeatable capabilityUnstructured AI rollout inside ERP as an organizational liabilityTranslating IT efficiency into quantified dollar savings to shift IT from cost center to strategic partnerMeta Description:A conversation with Daniel Topp, VP of IT at TASI Measurement, about how he built repeatable systems for managing ERP standardization and acquisition integration across 16 autonomous business units, and what it actually takes to position IT as a strategic partner in a decentralized global organization.Download, Listen, and SubscribeApple | Spotify | YouTubeOr search “Manufacturing the Future” wherever you listen to podcasts!