I have a very general question for those with turn-key machine tool solutions in their shop, did the sales pitch mirror the outcome? Did quality, cycle time, through-put, operator or automation duties, match what was quoted from the sales pitch, and was it contractual to hit the numbers? I had and experience years ago, capability of dimensions and cycle times were not met, and the sale was more important than facts. Been there, done that when the finger pointing begins, it's the cutting tool manufacturer or other finger pointing. Personally I found that I had to tweak much of the turn-key to be production ready, and processing time was understated from the get go. To be fair, I work in the casting world, so near net shape and uniformity of material is not a constant like machining from a steel blank. What has the rest of the shop world found from their turn-key?
This is kind of like asking "Are used car dealerships honest?" It depends entirely on the dealership, your experience, and what you think 'honest' means to a company. Can you tell the truth about something you are ignorant about? Can you be honest, and wrong? Can something out of your control happen that makes you appear dishonest?
Automation can be tricky, but the independent pieces of a 'turnkey' solution should be relatively easy to verify.
Will the machine meet cycle time and accuracy standards? Simple: have the manufacturer do a trial production run for you. Either be there in person to verify, or have them take video of 24-48 hours of production of your exact part.
Will the complete cell meet cycle time and accuracy standards? A little more complicated, because of all the moving pieces, but still doable. Whatever device cycle time dominates the cycle (most likely the machine tool) obviously needs to meet cycle time requirements, but if the automation can't keep up with it, it doesn't matter how fast the machine is.
Robot cycles are also relatively easy to verify - any integrator worth the name has simulation software to generate robot paths offline, which also generate accurate cycle times. Inexperienced or sloppy engineering means not allowing enough time for sensor reading/communication or actuator movement, or both, and can really add up to trouble in a complicated system. Ask for a video of the robot simulation (assuming it's a robot) and watch it carefully. A two-up gripper changing out parts in a spindle takes longer than you think once you add everything up.
If you're looking at more of a custom automation solution with multiple processes, ask for their timing chart, including device callouts, and start trying to verify those. Be on the lookout for motions that just don't seem right. i.e. a pneumatic motion going faster than 12 inches per second is a concern, and real-world switching times (which include the time it takes for devices to communicate) of less than .5s would also draw questions from me. Conveyors have 'settle time' when bringing parts to a stop, because they tend to bounce off of stops, or need to come to a rest position. Don't assume a perfect 'fill rate' from conveyors on shorter cycle times either. And on and on and on.
Making tweaks after installation and site runoff is pretty common, although it should be getting less so all the time. Optimizing a process is almost always the responsibility of the end user/buyer, unless otherwise indicated at the start of the job. Agreeing on what details are what can be very contentious, so it's usually better to just set the bar for all of the metrics the builder will have to meet, subject to the process you give them being good. If the process is bad, the finger pointing can get pretty ugly, so it's better to have an established manufacturing process to work from, or be prepared to eat any losses if your process doesn't work.
If you're looking for a second set of eyes on sales proposals, concepts, or anything else with the acquisition process, feel free to reach out through a message on here.
Good luck!