There has been a lot of advancement in artificial intelligence over the past few years. Much focus has been on bots, which many people view as the future of front-line service interaction. But I believe the future is not one in which artificial intelligence replaces humans on the front lines, but where people and AI work together in exquisite harmony to provide an unparalleled customer experience.

There have been significant stepping stones in AI interface and service design, each with its own inevitable hype. These have included SMS, old-school chat-based interfaces, personal assistants, automation, and the new hybrid we call chatbots. We can see the next ones on the horizon as well when we squint at Alexa and other developments — voice and text leading to natural language processing and automation. But some of the hardest work is actually happening silently in the background, and it’s connecting the dots and will provide the most value to the user.

This work includes a human-centric approach, deeper integration of multiple service layers, and new ways of interacting with our interfaces in more pleasing, human ways. The key here is applying human intuition and judgment where there are currently gaps, so we don’t get unreliable results. We cannot just leave things to automation, but rather must employ the elegant interplay between humans and machines to create new experiences.

The world is awash with data right now, and it’s causing overload. But slowly, in the background, AI applications and hardware will begin to learn context and personalization from our own routines.

This kind of empathic AI engineering does three key things: It considers the humans using and making the technology, it adds context, and it empowers people to connect with as few rough edges and friction as possible. And when this solution is delivered with speed and elegance, it will open up new possibilities of convenience and anticipatory service.

Empathetic AI will, among other things, help make sense of routines we can’t even perceive at first glance. For example, you don’t know if you have a habit of ordering Thai food in 50-day increments; that you’re more likely to pick up your mail on Tuesday than any other day; or even that you need antihistamines at certain times of the year that don’t correlate with typical allergy cycles.

We are still in the installation phase, and the engineering is not going to be flawless yet. It is difficult work to lay the AI infrastructure — like paving the literal highways throughout the United States to connect the offline and online worlds. But it is a noble challenge, and one that will set the stage for completely new consumer, commerce, and urban living experiences.

There are some things, like understanding and adapting to someone’s changing life circumstances, that a computer cannot do as well as a human can, and that will hold true for at least the next 20 to 30 years. We have to remember to bring back the human elements even as we find value chains, businesses, and industries that are changing completely with technology. Therefore, the next generation of AI is really about the marriage of human intuition and technology that is going to make a business or service that gets to know you over time.

Said another way, the objective is to leapfrog the chatbot trend and make something consumers will be truly delighted by. At Alfred, we argue that service of the future isn’t a bleeding edge futurist vision. Rather, it will simply hum with (silent) new efficiency. And it should be as frictionless as possible.

Howie Altman is the senior vice president of engineering at Hello Alfred, a technology and hospitality platform focused on evolving the most important space in people’s lives: the home.

Above: The Machine Intelligence Landscape. This article is part of our Artificial Intelligence series. You can download a high-resolution version of the landscape featuring 288 companies here.