Scott Dennis

Why Forensic Costing Beats Traditional Software in 2026 Tired of paying premium for black-box software that spits out "should-costs"... but can't explain why your real EV inverter or medical PCB assembly is 25%+ off? In 2026, the cost engineering landscape is flooded with AI-powered tools promising speed and automation. They're great for high-volume, standardized parts — fast estimates from 3D files, bulk analysis, regional data libraries. But here's the reality check for complex electronics, electrification programs, and custom/low-volume manufacturing: Forensic Costing (what we do at Intellicosting) wins when software falls short: Adaptability to novelty → New designs, redesigns, or bleeding-edge EV components? Software needs perfect CAD + baseline data. We start from ground zero — physical teardown, process mapping, supplier reality checks. No assumptions, no "garbage in, garbage out." Uncovering hidden drivers → Tools simulate known processes well, but miss real-world variables: tooling wear, regional labor quirks, yield losses in high-mix assembly, or supplier negotiation leverage. Our forensic deep dives routinely reveal 15–35% savings that simulations overlook. Human insight + accuracy. Forensic teardown + expert analysis is precise. We combine both where it counts — delivering actionable breakdowns you can take straight to suppliers (and win negotiations).

4

Many robots are built as a solution looking for a problem. Tatum Robotics started with a real problem and built the solution for it. In this episode, we chat with Samantha Johnson, the founder and CEO of Tatum Robotics, who recognized an urgent need in the DeafBlind community and specifically developed an assistive robot for communication. Added bonus — Jaimi Lard of Tatum Q&A joins the podcast for a live demo and talks about how the solution is changing lives. Real talk — Most of us got into robotics for one reason: We deeply believe that they can make a positive impact in the world. This is one of those stories and I’m honored that Samantha and Jaimi took the time to share it with me. Check out all of our content at https://lnkd.in/egFFKnyy.

103

3 Comments

Which of these is hits the hardest? ��� FFT-CPL makes linear polarization leakage mathematically impossible. 🔹 FFT-CPL separates CPL and LPL by construction, not correction. 🔹 FFT-CPL removes the need to prove your CPL isn’t LD. #FFTCPL #MakingSureArtifactsCannotExist #circularlypolarizedluminescence On-Line Instrument Systems, Inc. (OLIS)

3

In 2023, a founder came to us with a $1.5M budget to build their first IoT device. They’d already burned $400K on prototypes and were running out of runway. By the time their pilot shipped, they’d spent $1.48M. Without a single recall. Most founders don’t get that lucky. They burn through the full budget before the first unit even passes QA. Here’s what we did differently - and what every founder needs to get right in pre-production. 1. We locked the design before tooling. Most teams keep changing designs after quotes come in. That’s how you lose months. We ran DFM reviews, tolerance checks, and stress tests before cutting steel. Every part fit on the first try. Without delays. 2. We built the cost model before the BOM. You can’t fix margin later. We mapped all the real costs first - yield loss, test rejects, logistics. Then built the BOM around a clean 40% margin at 10K units. When part prices went up 7%, the model still held. 3. We treated each prototype like a stress test. Every EVT/DVT build had to prove one thing: can it scale without breaking the budget? When a connector failed vibration testing, we didn’t just swap it. We calculated what field returns would cost. That single fix saved roughly $280K in warranty exposure. 4. We didn’t celebrate the first working unit. Getting one sample to power on doesn’t mean you’re ready. We only celebrate when ten units come off the line exactly the same. That’s when a prototype becomes a product. 5. We gave every decision a clear owner. Too many builds stall because everything has to go through the founder. We set direct communication between their engineers and our factory leads. Every decision — mechanical, firmware, compliance — had a name next to it. No bottlenecks. 6. We slowed down where it counts: tooling. They wanted to skip the soft-tool stage to save three weeks. We refused. That run exposed a 0.3mm warp in the housing that would’ve jammed assembly. The “delay” saved a full re-cut and $120K in rework. 7. We tracked every review with numbers, not opinions. Each weekly update showed yield, tolerance variance, and test pass rate. When one subassembly dropped below 96% yield, we stopped the line, found the bad jig, and fixed it within 48 hours. That level of visibility builds trust fast. If you’ve built hardware, you know how rare this is. This is what happens when your manufacturing partner treats pre-production like a system, not a gamble.

23

6 Comments

Additive manufacturers who have sold polymer part printing services based on being faster, cheaper or more responsive than injection molding should take notice: Injection molding is fighting back!   This week over on Manufacturing Connected, we have a really fascinating interview with Lou Young Jr, cofounder and CEO of Atomic Industries.   This Silicon Valley startup is using AI to create injection molds that operate 50% more efficiently, are made 50% faster and cost 50% less.   You read that right. With extreme automation in mold design, a small team of engineers can not only do the work of hundreds, but pass immense savings on to customers.   Under this model, the AI-designed and optimized mold tooling becomes an operating expense rather than a capital expenditure. Atomic says its tooling costs can approach zero with its vertically integrated operation.   That’s a huge paradigm shift in making molded parts (and in the future, die cast and sheet metal parts) more affordable and accessible.   But don’t take it from me. Check out Christina Fuges' conversation with Lou for more on Atomic Industries' aims and vision: https://lnkd.in/e56ShqNs I also recommend Christina's story for MoldMaking Technology that gets deeper into the technology behind all of this: https://lnkd.in/eSz8HhS9 OEMs, will this kind of technology advance change how you think about sourcing parts? Additive manufacturers, do you see it altering how you position your services?

88

19 Comments