I often hear people saying that the goal of technology is to match human capability and efficiency. The real goal should be to outperform humans. Humans might be impressed with a facsimile, but they’re not likely to rely on a replicated version of something they can already do themselves.

Sure, self-driving cars are cool, but very few people use them, in large part because they feel like they can do a better job. It would hardly be a victory to create automated vehicles that match the current accident and fatality rates of human drivers. For the moment, people seem most interested in automated assists like automatic braking, but long term we’ll require cars that are better at driving than we are, that push those rates down toward zero.

The same can be said of natural language processing (NLU). Some might think we want our machines to interact with us like humans do, but we don’t. No emotional complexities, no psychological nuance. When it comes to automation, we want efficiency, not emotional complexity and psychological nuance. Perform this task for me without excuses, give me this information without caveats, help me prepare for my day without giving opinions I didn’t ask for.

The goal with hyperautomated tech is to make our lives more efficient and productive. Too often, humans get in the way of both. These better-than-human experiences will also be the gateway to widespread adoption. Once people start feeling confidence in (and excitement about) working with automated goods and services, their proliferation will be seismic.