Pinned Tweet
The big hurdles:
- Motivated folks building sol'ns for problems they don't fully understand
- Lean teams who want a perfection before applying tech
- Data teams claiming your data is garbage when the critical data isn't even being collected
Meet Practical AI for Manufacturing:
Honor the Skill. Build the future.
"We're don't plan on checkin' out until we're called out."
@KeystoneSynergy's Larry Crist and Craig Ray.
How can you not love manufacturing...
Find Murray at Booth S20046, behind @bigassfans. Ask him anything. He’s been waiting all day.
Design, onboarding, operations, maintenance, logistics, field service - all rely on deep tacit knowledge.
Capture, synthesize and leverage decades of knowledge. No scrubbing required.
See us at @FabtechExpo S20046
Visit the Murray team at booth S20046 @FabtechExpo to find out more...
Multi-gen manufacturers build the future and @FerretInc honors that legacy with exceptional quality and service.
Excited to hear from our customer, and a great @FabtechExpo panel this month in Las Vegas.
Visit the Murray team at booth S20046, right down the aisle from our good friends at @SMEMediaNews, in the AI for Manufacturing Section, and make sure to catch the panel on Thursday 10/22, at 12:30pm, hosted by Steve Plumb - you won't want to miss it.
MurrayMentor retweeted
This is worth a hard look - something is definitely wrong: 8 years, 223 sites, 23 that scaled to even one more site.
Congratulations to those 223 exemplars. I've been there herding the cats at one of the world's largest industrials. It's not only hard work to ideate and deploy, but very hard to scale technology in manufacturing.
But a mature leader needs to look at these numbers for what they are - abysmal.
📆 across 8 years
🚩a meager 223 manufacturing sites logged as lighthouses
🚨and a pathetic 23 scaled past site one
While so many influencers talk about lights-out AI-driven factories and Industry6.0, the real world is still stuck in the Industry3.9 Rut. And the influencers and solution providers and physical-AIers and only-one-name-spacers and one-more-universal-APIers have not made a dent. What is 223÷7,500,000 ? What it's not is a good batting average.
The problem is right in front of us. And it will clock out for the last time this coming Friday.
Deep tacit knowledge is:
✅ 80% of what it takes to keep a plant running
✅ both everywhere and uncollected
🚨 disappearing at an alarming rate
💰 ready (minutes) to be harvested, synthesized and leveraged across your teams, machines, lines, sites, regions and languages
We have talked about the brain drain for decades now, it is happening as we speak, and it is universal for manufacturing.
The advent of scalable AI for manufacturing has changed the math, but most are focused elsewhere; on more sensors, more explicit data collection, the requisite data scrubbing and yet one more architecture layer for a dashboard dressed up as decision rights.
I get the distraction. Really, I do. Remember, I lived it.
Validating my rant is that the Lean practitioners at the world's best Lean Exemplars are doing their own version of expectation theater. Stabilize-before-automation folks team with reimagine-your-workforce consultants and analysts. Kaizen dashboards tell the tale.
Two sets of advisors pulling in opposing directions, vertically, at the y-axis, ie, 223÷7,500,000.
In the meantime, the tug of war along the X-axis is between the disappearing knowledge and Autonomous Collaboration; the ability to harvest, synthesize and leverage that critical data set across teams, machines, lines, sites, regions, languages, ...
What happens when you link those two sets of opposing forces? A curve starts to form, with an eventual up-and-to-the-right trend as the robots do eventually claim more ground in manufacturing.
This is not one or the other, it is both / and; a collaboration of the Digital<>Human<>Physical.
Autonomous Collaboration:
- reduces risk
- secures IP
- deploys amazingly fast (1 hour)
- scales virally
- eliminates cognitive load
Murray Mentor was built to fill the critical data gap in manufacturing, the 80% that is either totally ignored, or so siloed it's rendered useless.
PS: in our short time in market, we've already deployed multi-site operations.
"A $ B worth of Mets... and none of 'em could pick it up", or how AI in manufacturing is going for most manufacturing teams...
Continuing our baseball theme:
- incredibly well-trained, highly paid experts converge on a problem space
- any one able to "make the play"
- while none deliver the goods
At IMTS - International Manufacturing Technology Show this week, we took the SME Smart Manufacturing stage with not one, not two, but four leaders using Murray as an AI tool to help transform their manufacturing business.
The key takeaways 🍿
🚨 Deep Tacit Knowledge (DTK) is disappearing at an alarming rate
🔥 Autonomous Collaboration uses AI, automatically, to Capture, Synthesize and Leverage Deep Tacit Knowledge (DTK)
✅ there already exists a huge well of data (DTK) for AI to leverage immediately
✅ DTK = 80% of what it takes to keep a manufacturing plant running
💰 DTK is only valuable if it can be captured, synthesized and leveraged
From training to on-boarding to easier consumption of engineering and innovation at the front line to reliable execution - the leaders on stage told the attendees how they exploit Murray and Autonomous Collaboration to drive next level performance.
Here's the most 🚨 part
- Murray works out of the box
- Murray is deployed in hours
- Murray scales in days to weeks
- Murray requires no special infrastructure
- Murray creates not additional operator work or cognitive load
Honor the skill. Build the future.
76
Honor the skill. Build the future.
Replying to @MurrayMentor
@MurrayMentor has named its new CEO.
Industry veteran Paul Boris will lead the company as it continues to bridge the manufacturing knowledge gap with AI-powered technology.
Boris will participate in a Sept. 15 panel discussion at #IMTS in Chicago. Titled "From Machines to Minds: How AI is Reshaping Manufacturing Talent," the panel will focus on how AI-powered software products can help develop and upskill the manufacturing workforce.
Full announcement here: advancedmanufacturing.org/ne…
#IMTS2026 #MurrayMentor #AI #technology #solutions
47
MurrayMentor retweeted
Replying to @MurrayMentor
@MurrayMentor has named its new CEO.
Industry veteran Paul Boris will lead the company as it continues to bridge the manufacturing knowledge gap with AI-powered technology.
Boris will participate in a Sept. 15 panel discussion at #IMTS in Chicago. Titled "From Machines to Minds: How AI is Reshaping Manufacturing Talent," the panel will focus on how AI-powered software products can help develop and upskill the manufacturing workforce.
Full announcement here: advancedmanufacturing.org/ne…
#IMTS2026 #MurrayMentor #AI #technology #solutions
Friday 5am, two machines down, order must ship today, veteran mechanic Edmundo gets handed the wrench, and (cue the fireworks)...
Autonomous Collaboration allows manufacturers to:
- Skip the wasteful data-scrubbing
- Deploy AI in your plant(s) in hours, scale it in weeks
- Harvest the deep-tacit knowledge that no other system collects
- Leverage that knowledge across shifts, lines, plants, languages, ... instantly
Honor the skill. Build the future.
🤷♂️ If your senior operators and technicians need to step in and work magic when the game is on the line - while this makes for great Phillies' baseball - it is no way to run a manufacturing plant.
...and no amount of explicit or synthetic data, naming conventions, data connectivity and scrubbing will save the day, that's why they call it gambling.
According to my good friend Gemini:
✅ 80% of what keeps a plant alive is uncollected, deep tacit knowledge
LNS Research warns:
😬 Average time in position has collapsed 94% from 108 months to 7 months
St Louis Federal Reserve raised the alarm:
📉 Manufacturing productivity has cratered in the last 15 years
Deloitte says it's a time-bomb:
💣 41% of current skilled maintenance workforce eligible to retire by 2028
SHRM says mass replacement is expensive and ineffective:
💰 $40K+ is the average cost to replace one skilled maintenance technician
Autonomous Collaboration is not like any other approach, and you can see why in just 15 minutes. Reach out, we'd be happy to show you, and skip the PPT (do solution providers still lean on PPT to hide their deficiencies?).
1
55
Failure to tag up, four outs, downtime on an $80M 'manufacturing process' and why "Standard"OPs are failing you.
Even the announcers were confused...
It's all in the "Standard"OP, but apparently no one on the entire field knew what that SOP said, or how to find it when they needed it most, resulting in extended downtime on an impressively expensive machine.
In reality, SOPs are management's way of pushing blame to the front line when things go wrong: 'it was in Par 37.A.iii on page 71, how did you possibly miss this one, Timmy. It's your fault'.
🚨 The Best Doers are the Worst Documenters
✅ your work instruction / connected worker apps are failing you and the knowledge it takes to run your plant is escaping, the data is unambiguous
✅ the lack of collaboration across tacit knowledge teams is very costly
🔥 deploy AI in hours, improve collaboration in weeks, reap the benefits in reduced downtime, drive performance and revenue higher, faster.
Honor the skill. Build the future.
tl;dr thanks to my buddy Grok
- Flaw in the sensor data (scoreboard ) coupled with distracted operators (runners never tagged up) led to the preferred outcome (run scores) despite no one on the field but the batter (planner) knowing what to do (sac fly), including the Quality, EH&S and department (umpires) or HR (coaches).
- The dashboard (game summary) will show a good run (scored) but no one will remember the deep tacit knowledge required (and then forgotten) to get there.
What happened:
- runners on 1st and 3rd
- scoreboard (sensor data) says 2 outs when there's really only 1
- short sac fly (solid plan) which would easily score the runner from 3rd
- confused runners (distracted operators) never tag up, thinking the inning is over
- the maintenance team (defense) know it's 2 outs, went for the third out at 1st
- none of the other operators, quality (umpires), etc remember that 'not tagging up' is NOT automatic out
- defense never logs the out at 3rd, and leaves the field, thinking work is done
- it THEN becomes a 'time play' (anomaly without operator context)
An $80M asset (salaries of all involved) didn't read or know the SOP, and all looked confused causing excessive downtime.
What the SOP that no one could find says
- defense should have thrown to 3rd for a 'fourth out' (weird huh?)
- defense challenges BEFORE they leave field (team didn't know the SOP)
- the runner would have been called out at 3rd
- no run would have scored
Even the Quality and EH&S department leader (crew chief) could not properly explain why it happened the way it did, ie, not tagging up is NOT an automatic out.
Manufacturing life imitates a game - except with different urgency.
Some hard facts from Claude on manufacturing and Deep Tacit Knowledge (DTK) and turning do-ers into documenters...
❓is explicit data or deep tacit knowledge the majority of information it takes to run a plant, and what percentage is it ?
✅ "The direct answer: ~80% tacit, ~20% explicit is the number that gets cited most consistently across knowledge-management literature"
❓of the 80% what portion of that can be collected by forcing the senior experienced person to dictate, type or video tape a procedure to convert it to explicit data?
✅ "Short answer: not much of it — call it 10-20% of that 80%, i.e. roughly 8-16% of total plant knowledge"
❓what is the documented best way to convert an senior experts deep tacit knowledge into explicit knowledge?
✅ we can show you in 15 minutes
Deploy in hours. Value in days. Scale in weeks.
💰 Lean. Light. Fast.
Honor the skill. Build the future.
Boss Man keeps this old man around after retiring because he is the only one with tribal knowledge to run this part 🤦
30
"If they can break the intellect-chain, they beat us." Honor the skill. Build the future.
@DestinSandlin says it all in this @smartereveryday video:
"...I would find out who the wizard is, or the person at your company who knows the secret sauce, and give them an apprentice, and keep that knowledge continuity going..."
BUT...
Resist the temptation to turn manufacturing do-ers into documenters. It won't work.
Gemini told us:
🔥"the harsh reality of self-documentation ...a novice can successfully apply only about 2% to 12% of an expert's total operational wisdom to solve a specific, unexpected problem on the plant floor."
Interested in learning more, just DM and we'll explain it in 15 minutes.
Critical knowledge is escaping manufacturing, and the typical crop of solutions aren't up to the task.
It's not all bleak, though.
WHY TRIBAL KNOWLEDGE IS SO CRITICAL...
The drive to embed technologies is intended to accelerate productivity, but often impedes progress, especially as senior resources leave our operations with the requisite knowledge to troubleshoot it all still trapped in their heads.
Is this transmission dead at 193K miles? Seems likely. As a result:
- processes are disrupted while new equipment is onboarded
- new equipment takes time to ramp up delaying orders
- there is a large cost (often unexpected) associated with terminal failure
The video is pretty fascinating:
✅ Some of the things we see as mundane are often wildly complex
✅ Many of the failures that appear terminal are not
💰Deep tacit knowledge keeps operations running effectively (just ask Ford)
Dead at 193K miles seems logical.
Doing a "break-in" drive at 193K miles does not.
The "Standard"oP for the break-in is well documented, but:
- why check fluid first, vs tearing it apart?
- what triggered the idea to initiate a procedure used on new transmissions?
- what signals would have indicated a different approach?
DON'T WE HAVE A TON OF DATA THAT WILL ADDRESS THIS...
We asked Gemini:
❓what amount of knowledge required to run a typical mid market manufacturing plant is considered tribal and uncollected ?
😳 "industry benchmarks reveal that an astonishing 70% to 90% of operational knowledge is completely undocumented, uncollected, and treated as tribal knowledge."
❓what percentage of the deep tacit knowledge from experts can be collected by making doers become documenters and making them capture video or write SoPs?
🤔 "writing standard operating procedures (SOPs) or recording videos typically captures only 10% to 30% of an expert’s truly deep, tacit knowledge."
❓what percentage of that can be used by novices for troubleshooting?
☹️ only 20% to 40% is actually usable by a novice for troubleshooting.
Shockingly, Gemini told us:
🔥"the harsh reality of self-documentation becomes clear: a novice can successfully apply only about 2% to 12% of an expert's total operational wisdom to solve a specific, unexpected problem on the plant floor."
Read that again. 2% to 12%
MURRAY IS CHANGING THOSE NUMBERS DRAMATICALLY:
✅ an AI-Mentor for skilled trades
✅ zero cognitive load
✅ deploy AI in your plant in hours
✅ scale it and deliver value in weeks
💰💰 Capture that critical knowledge before it escapes
Find out why companies like Indiana Furniture, Flexible Concepts, SteelWarehouse, Pinnacle Polymers and others turned to Murray to tackle the most critical problem facing manufacturers.
Honor the skill. Build the future.
How I Fixed a Bad Shift Without Replacing the Transmission (Part 2/2)
Resetting the adaptive learn cycle is only half of it. Warm it to 180, then drive all 10 gears six times in a stop and go sequence.
#DieselTrucks #WhitefaceFord #SuperDuty
MurrayMentor retweeted
Today, @Ford announced we’re joining forces with our founding partners at @BlackRock, @Google, and @Carhartt to form a new organization, the Alliance for America’s Skilled Trades.
Skilled trades are at the heart of what we call the Essential Economy, the 95 million Americans who build, move, and fix the things our country depends on every day. These vital trades and industries are the backbone of the American economy, but millions of these jobs are expected to go unfilled in the years ahead.
At Ford, we recognize the skilled trades workforce shortage is a national crisis but also a generational opportunity. Why this alliance? Well, like any big, thorny problem, no one person or organization is going to solve the skilled trades shortage alone. And so when I started talking to Larry Fink, Bayo Ogunlesi, Ruth Porat, and Linda Hubbard about these issues, we knew that we could do more together.
We’ve got three main goals: build a jobs pipeline for skilled trades, highlight what's working across the country including the launch of a Skilled Trades Report, and grow this movement with more partners.
This is just the start. We look forward to welcoming others on board who are equally passionate about solving this for America.
Read our op-ed today in @FortuneMagazine to learn more about our collective work: fortune.com/2026/07/21/black…
#EssentialEconomy #SkilledTrades
I found video from our Plant Manager's call with the Corporate AI consultant and the plant CI Leader.
I hope he doesn't mind me posting it.
The explicit data has remained unchanged for a hundred years, the deep knowledge to leverage it all changes mid-pitch.
The ball has been the same for 75 years, the bat for 100 and the human for thousands.
Yet 75M people still sit for hours to watch what might happen when they all meet at the plate 160K times a year. And it is a source of endless entertainment because of the unpredictability - unpredictable despite everyone using the same "assets".
Operators have the same scenario with the assets on every shift.
Except the unpredictability is not entertaining, it can be costly and wasteful.
Imagine Justin Verlander never having a coach, and never sharing his technique with the next generation except through a manual, after-the-fact analytics or videos.
This is largely how manufacturing functions. Until Murray.
Your manufacturing coaches are being paid to play, not coach. And when their expertise escapes, it forces the next up to discover the winning techniques all over again, on their own. What if it were easy to collect all the expert knowledge and put it at your operators fingertips - or in their ear - right when they need it.
🔥the loss of deep tacit knowledge is incredibly costly, ask Ford
🤖no amount of explicit or synthetic data can substitute for expert know-how
Yet 95% of AI project fail, so what's the option?
💰Murray captures, synthesizes and leverages all that knowledge
✅ deploy in hours
✅ scale in weeks
✅ zero cognitive load
Reach out to Christian to see it live and hear why world-class operations like Indiana Furniture, Flexible Concepts, Inc., Pinnacle Polymers and many more trust Murray.
Honor the skill. Build the future.













