Artificial intelligence pops up as a buzzword every few years, but it has never moved beyond novelty status. This time, though, it is here to stay, and startups are poised to drive the AI economy forward.

Indeed, we are beginning to see glimpses of it.

Newcomer ROSS Intelligence, for example, has gained law firm clients by developing a fully automated AI “lawyer” capable of supporting the legal research needs of large offices. Developed on IBM’s Watson, ROSS is well on its way to becoming a fixture in the legal industry, automating tasks that could take days or weeks for humans to complete.

Popular business messaging app Slack — another startup — is working on incorporating AI to act as an intelligent personal assistant capable of talking back and answering questions that have been asked before, saving companies time.

Now is the time for AI to truly take off. So what has changed?

The infrastructure is ready for startups to run with AI

Limited computing power hindered AI research in the past, but our current infrastructure and ecosystem can now support the processing power necessary for AI to “think.” Processing and memory capabilities, cloud computing, fiber-optic high-speed internet access, widespread Wi-Fi and a connected IoT world all combine to create the perfect environment for AI to exist — and have staying power.

Twenty years ago, only enterprise companies with research and development teams could work on AI; even then, most of the work was primarily theoretical. But now, every entrepreneur has access to quick connections, fast devices and the technological infrastructure built by large corporations.

The rise of Facebook and other social media platforms has contributed, as well. Never before have we been able to harvest so much real-time and historical data about how people interact. In the previous millennium, access to this kind of detailed data was unthinkable, much less the idea of it existing largely in the public realm. Because data as a resource is more freely available than ever before, the ability to create an AI ecosystem is obtainable by anybody willing to work toward it. The barriers are down.

Like the original internet bubble of the 1990s (or even the current shift to mobile), AI represents a disruptive new lane in which small startups can take the lead and drive innovation.

Recent forays into AI

Enterprises like Microsoft, IBM and Alphabet are among the corporations that have been actively working on AI for some time. The machine-learning applications these enterprises have launched to the public have had varying degrees of success.

Watson’s ability to answer natural-language questions makes it an ideal resource for professionals in complicated industries, such as medical and financial. Google’s search has long employed complicated algorithms akin to AI, and parent company Alphabet is already using its research and development to reinvent the web. Even Facebook is heavily invested in AI.

While these giants make mainstream media swoon with their AI platforms, it’s indie developers utilizing these platforms, like those at ROSS and Slack, who create the real innovation. According to Bloomberg, more than $300 million was invested in AI startups in 2014. By 2020, AI is expected to become a $20 billion business, including cross-platform search, voice assistance and proactive support. Gartner predicts 85 percent of customer interactions will be managed by machines.

Where startups can take AI

When we start seeing the disruptive technology come out of the AI industry, it will emerge from the smaller players using the tried-and-true platforms in new, focused ways. But while startups will overtake enterprises to capture the AI market, that doesn’t mean the larger corporations don’t have a role to play.

They will act as infrastructure providers, similar to the way cable companies have served as the foundation for startups to build new and innovative services after the emergence of the internet. Success will be with the startups that leverage AI technology for specific verticals, such as agriculture, manufacturing or insurance.

Startups may be able to provide established insurance companies like State Farm, Allstate and Farmers with technology that will allow them to become more proactive in their policy planning. Consider a new artificially intelligent insurance underwriter that can better forecast natural disasters and accidents, and adjust premiums. AI accounting and other financial services are sure to follow, as the technology can be used to optimize spending.

The predictive decision-making capabilities move beyond just being a cool, novel technology. Even the food supply chain could be managed by AI. Consider that up to 40 percent of the food in the U.S. is wasted. Startups could create an end-to-end farming solution with AI predictions that would be able to move to Africa what used to be waste, for example. It will have a huge impact in terms of how we tackle global hunger and famine issues.

Startups are already at the cusp of impacting agriculture. Consider the LettuceBot, a startup’s AI creation that pinpoints weeds. Instead of spraying an entire field with weed killer, the machine identifies weeds and kills them with precision. This not only takes the place of time-consuming manual weeding (reducing overall operation costs), it also increases yield. Farmers using this technology no longer have to over plant to compensate for lettuce lost by weeds, either.

Whether serving as a research assistant supporting a large office, acting as a voice-activated resource in complicated medical procedures or proactively maintaining itself, AI is quickly becoming a reality. Like previous iterations of the internet and computing technology, the AI revolution will reward new players who learn to use it to their advantage. AI will play a role as a fundamental predictive enabler to help us solve large-scale problems, and startups are poised to lead the way.