If you'd come here two days ago, I could have saved myself the irritation of trying to get a general applicable definition out of everyone I talked to. I literally started an INTp post to understand Ti . I don't use it, but now I'm somehow an expert.@Octowave you're confusing Te/Ti and Ne/Ni, which is probably not your fault considering there is a lot of misinformation.
Te believes a decision should be made with efficiency and quickly looks for external data to confirm the efficient conclusion. For example, 90% of men who are accused of hitting women are guilty. When a man is accused, the efficient answer is to conclude he is guilty. Te tries to raise it's own standards, to where it isn't that simplified. So 50% are obviously guilty, 40% require 3 days of research, and 10% are innocent. Because of these new standards, when a man is accused, from now on, the protocol is to do 3 days of research. Te notices that this system can be improved, however, so etc....
Many Te users are highly intelligent and creative. The core focus of their improvements, however, is about improving an automated standard. When a decision needs to be made, Te prefers to look to experts, prior decisions, and current systems to efficiently make a decision and move on. Te can be very beneficial in large systems where we don't have time to research every nuance.
Ti sees each case as unique, requiring its own set of research criteria. Instead of 90% are guilty, Ti challenges that and wonders what data backs that up. While Te adds 3 days mandatory research, Ti says there isn't a limit to what can be considered, as Ti prefers a thorough analysis, of that current system. After several weeks of research, Ti concludes that 50% of the accused are likely guilty, based on the evidence. However, with the other 50%, there isn't enough evidence to conclude that any of them are either guilty or not guilty. Ti proposes that an investment team is created to look at each individual case to see what amount of research is needed.
Both Te and Ti users can relate to both of those up there, but preferences come through. As an employee, I'm fully capable of working in a system with automated decisions in place. I like to work efficiently, but that is more so because my personal thoughts prefer it. Te users tend to be more clustered, in terms of conclusions, while Ti is all over the place.
Getting tired of typing, so quick Ni vs Ne. Ni is deductive. It takes potential conclusions and selects the one that is most clear. Ni is much more of a "a-ha" moment than Ne, as Ni waits for the revelation to occur. Ne is inductive. I would describe it more as a "ripple effect". Ne considers A which connects to B, which connects to C, all the way to Q. From there, Ne sees a parallel between D, H, and M and concludes those 3 are related. Ni sees that parallel and concludes "D is what I'm looking for". The fun thing about Ni vs Ne accuracy is neither are guaranteed to ever be correct. Ni sees D and gets fixated on D. Ne, however, proposes D, H, and M as the likely answers. Ne waits for feedback, and then concludes H is the answer. Te and Ti are then used to uncover information to prove either D or H.