Call to cash automation… thought about it?

[thunderskunk]

I made a post asking about high volume work a while back. Someone made a comment, I think it was OX, said something like, “If those jobs were out there, easy, and paid good, we’d all be doing them.” It made a big impression; I think about it when I get crazy ideas that the sales people say solve our every problem. If the money was where their mouths were, we’d all be doing it.

That said, has anyone ever entertained front to back automation? Customer places an order online, the work is completed with no human intervention, packed, shipped, and the customer receives the product. I’ve had a few unattended posts lately that clue in to what we’re doing. I’m not saying it would actually work, but I also gotta believe someone else has done it. Especially something like 3-D printing. I’m not a fan, but if your printer is automated and everything the customer wants is in the webform, all you should need to do is feed the machines and conduct maintenance.

Probably the best example I can think of is petroleum processing: a DCS orchestrates the whole operation out of necessity. If it can do that, why couldn’t it work for a machine shop? Especially if you have a narrow scope of products you make, but even if not, CAM is post-and-go these days. Is it simple, absolutely not. I’m just wondering if anyones thought about it or seen examples of someone trying it.

 

[sfriedberg]

SendCutSend, ProtoLabs, and similar operations are getting awfully close to that. They still have humans, but all (or nearly all) the quoting, CAM, stock/material management, and most of the shipping operations are automated. I don't think they have replaced humans at the press brakes yet, but I bet the backstops and ram depths on the brakes are all auto-programmed for each part that gets bent. Everything driven from the CAD drawings supplied by the users.

 

[DDoug]

Gee...I wonder why more automation...:https://www.wjr.com/2022/10/11/gordon-has-covid-created-a-society-of-entitled-deadbeats/

 

[thunderskunk]

Gee...I wonder why more automation...:https://www.wjr.com/2022/10/11/gordon-has-covid-created-a-society-of-entitled-deadbeats/

That article has absolutely nothing to do with automation. COVID sucked, COVID policies sucked even more, businesses suffered. Got it. Not disagreeing, just wondering what your point is.

 

[Ries]

I think the reality is that automation still is expensive, and the efficiencies of scale really only kick in when we are talking about VERY large volumes.

Consider the costs of the more self sufficient machines, which still need some degree of human feeding-
People often ask here, can I make money with a $250,000 laser cutter, or a half million dollar 5 axis machine, and the answer is usually- only if you already have customers in the very high profit sectors.

If you are doing aerospace, or medical work, where the parts prices are quite high, you can justify and pay for machines with more lights out features, but, then, adding robotics to load, you still need a fair amount of human presenced to integrate with inspection, finishing, coatings, and packing.

Sure, you can build a production line thats completely automated, but usually that means its built to make a narrow family of similar products.

Amazon has found its still much cheaper to hire 100% new employees every year to pick and pack, rather than automate. (actually, I think their turnover is worse than that. Soon, we will reach the point where every single US citizen has worked for Amazon and quit with injuries)
The degree of complexity of even a simple box filling and taping system capable of dealing with the approximately 100 different sizes of boxes Amazon routinely uses would be in the millions of dollars per, easily.
I just dont see how you could make any profit with that degree of automation, unless your minimum sale was ten grand per item.

 

[standardparts]

Obviously a complete elimination of humans is going to be a tough one---someone has the be the big boss I guess.
The main mission for over 100 years has been to automate as much as possible to eliminate human factors.
It is not unusual to run 'lights out' on one or more shifts at this time. To make the jump to accomplish online order taking-- to manufacture-- to shipping is in the future for sure and the reason advances in AI are strategically important. Slowly but surely it will happen.
But when I look at vids of manufacturing plants I'm stunned on how few people are visible in the process compared to years ago. The number of workers required to manufacture goods is dwindling but that is a story to be discussed on another day.

 

[DDoug]

That article has absolutely nothing to do with automation. COVID sucked, COVID policies sucked even more, businesses suffered. Got it. Not disagreeing, just wondering what your point is.

W-O-W...the point ?
Too many lazy people, so employers are automating.

DUH

 

[Scottl]

While I think automation has made great improvements I think it is still far from the time where it can run 100% without human intervention. While some customers might have a perfect experience using such a start to finish operation I suspect it would only take a few major screwups to sour people on the idea and those could be expensive mistakes to correct, on your nickel.

I've long said that people and automation should be partners, not competitors.

 

[boosted]

There have been literally dozens of companies that have tried this with machined parts. Some (like Protolabs) are actually pretty successful at it. Some (like Hadrian Manfucaturing) are still in the exploratory phase. Most have given up, and either pivoted the business or disappeared.

I think the secret to making the Protolabs model work is that they have a whole bunch of rules which dramatically reduce the scale of the problem.

If you want to automate just a regular job shop all the way through, the ask is way too complex. The scope needs to be highly focused.

 

[thunderskunk]

W-O-W...the point ?
Too many lazy people, so employers are automating.

DUH

Ok… sure. That article was about a landlord who was going broke because of government “save me” policies that cater to votes, not problem solving. Is it because people are lazy? I guess. I haven’t seen any policies that say we have to pay employees to not work for months/years because of COVID. This is odd, as I think we agree on views in this scenario, I just don’t think it applies here. Why does everything need to be about COVID? Our drive for automation is because of upcoming retirements, not laziness.

I think the reality is that automation still is expensive, and the efficiencies of scale really only kick in when we are talking about VERY large volumes.

…
I just dont see how you could make any profit with that degree of automation, unless your minimum sale was ten grand per item.

It’s a valid point, and we’re running those figures. We have projects pitched for every step, and thus far everything has an ROI. Not always a quick ROI. The physical steps have been easy justifications: $300k machine center replaces 4 old ones, requires 1 operator instead of 4, $80k per op per year, ROI reached in a bit over a year, more when you consider other benefits. Automate a manual labor task like parts loading with a $200k automated cell, replaces 2 ops, reduces process risk, maybe ROI in months not years. But then again, the more variability and volume said system has to handle, the more it’ll cost. I’ve seen automated stock management before, and packing/shipping is pretty common. If you can automate material handling from step 1 to 2 and so on, it doesn’t seem impossible.

The thing that spurred this was working on specifications for automated assembly. The age old debate: what’s the best PLC to use and why? If we were trying to go as cheap as possible and build it ourselves, I’d go with Automation Direct’s hardware as it’s way cheaper. When you think holistically about work centers communicating with each other and whatever plant management software/hardware you go with, something like Siemens is more attractive. Which is where it doubles back: how far can/do we take this? Is the goal 100% automation? Maybe, and maybe not. To your point, will it ever be worth it?

 

[rcoope]

W-O-W...the point ?
Too many lazy people, so employers are automating.

DUH

I was very much struck, reading Fredrick W Taylor's original writing on Scientific Management and what became Taylorism, how obsessed he seems to have been with lazy people in 1881. Now it's true that he resigned from studies at Harvard due to temporary vision problems, and then inexplicably returned to work as an apprentice machinist who worked his way up to manager by 25 and then studied engineering, so he may have come by his exasperation honestly, but we can say that the desire to automate to get around the prevalence of what are perceived to be lazy people, is very old.

 

[rcoope]

It is remarkable how much easier automation gets when you go from 3D to 2D, and by that I mean sheet or bar. Laser and other sheet cutters, and bar feeding swiss turns, lathes and mill turns, all seem to be a remarkably amenable to automation of the sort where one owner operator can run multiple machines. See Danny Rudolph for example, (eg on Instagram), one man with six or so Citizens, the guy in the new NYCNC IMTS video who runs two Willemin Macodel mill turns with bar feeders by himself.
If you watch Saunder's new video, the repeating theme from all the mill manufactures is they are building machines for longer and more flexible runs, with more pallets, more robot friendly work holding, modular tool magazines, and a lot more in-process measurement. So the industry definitely thinks its coming more and more, but how the economics works for job shops, is beyond this institutional prototyping guy's knowledge.
An observation from my own experience is that end to end automation is hard with pure prototyping, where the shop and designer probably need to go through some iterations to make the thing easier to manufacture, but if you could automate from the customer's second batch of parts and then handle rising orders as the product matures in the same pipeline, you might well be winning. You have to believe the customer is going to succeed of course...