Three ways AI is already at work, and what comes next
AI has moved from potential promise to real impact. It’s already being used in powerful ways in the enterprise. As I’ll outline below, there are deeply valuable ways it’s being used today and emerging improvements to the underlying technology that will further accelerate it’s impact on the enterprise.
AI is coaching humans to work smarter
AI-powered systems are being adopted into the enterprise, and machines and humans are learning how to best work together — letting machines do what machines do best and humans do what they best — work will potentially become easier, faster, and more accurate. As MIT’s Erik Brynjolfsson has said, we need to learn to race with the machines — not against them.
For example, Gong.io is an artificial intelligence sales platform that records and analyzes sales conversations to determine what words, phrases, and parts of a conversation lead to the best sales outcomes. Gong creates a constant feedback loop for sales reps and managers about what’s working in real-time to help them do their job better.
New AI-powered tools such as Gong help people tackle the basics more effectively, so they can focus more energy on doing what people do far better than machines: building emotional connections.
AI is helping software improve itself
Even after decades of computing progress and experience, we’re still terrible at developing software. It is usually full of bugs, frequently has security holes, and doesn’t always do the job it’s supposed to do.
With the advent of AI, software will ultimately be able to develop itself, creating more capable and scalable systems. This next-generation of applications will have fewer bugs, fewer security holes, and (best of all) continually get better at doing the jobs they were designed for the more they are used.
For a real-life example, HeadSpin.io is a global testing platform for mobile apps. HeadSpin uses AI and computer vision to detect issues with performance and reliability, analyze end user experience, and measure performance across different areas of the world — ensuring that the user’s experience is as fast and delightful as possible. The platform integrates directly into developer’s existing tools, so they can immediately be productive and automate their testing tasks.
I believe AI will streamline the development of most enterprise software. Perhaps we will see the rise of fully automated Software-Development-As-a-Service (SDaaS). Instead of buying an app off the shelf, a customer picks from a menu of what they need, and receives a custom-developed SaaS product in hours, coded up by an AI-powered engine.
Verticalization of AI is bringing new insights to established industries
Artificial intelligence is transforming long-established industries like mining, insurance, and healthcare, as players in these verticals begin to realize the full potential of AI in ways that matter to their industry. Industries from mining and quarrying, construction, insurance, and property management are already leveraging platform to digitize the physical world and apply AI to find deep business insights.
A great example of this is Airware, which provides enterprise drone analytics across a range of industries. Airware’s platform uses AI to analyze the imagery data gathered from drones and perform precise analytics that are shared and used across the enterprise. This aerial data plays an important part in digitization by creating data driven models of what’s happening in the physical world to optimize the flow of materials across large and complex work sites and saving millions of dollars in costs and increasing output.
How AI will work better tomorrow
Looking around corners with artificial “intuition”
Historically, AI has been good at recognizing patterns and making appropriate recommendations. But things are changing quickly. AI has recently begun to carry its learning farther — almost as if the technology has developed a sense of intuition. For example, self-driving technology can allow cars to see their current surroundings and react almost instantaneously to them, such as when a child runs into the street.
Today, new techniques in AI are enabling systems to see slightly into the future, such as identifying the presence of the child and anticipating whether or not he might run into the street. As this type of advanced learning makes its way into the enterprise, the types of insights provided by AI systems will get better, more accurate, and more prescriptive.
Getting smarter faster with transfer learning
People who already know more than one language usually can pick up new languages without difficulty — even if the languages don’t share words or even a common alphabet. The basic principle at work illustrates the AI technique called “transfer learning” — the idea that a computer can improve learning capacity in one area through previous exposure to a different one. Until recently, AI has demanded that computers must learn largely from scratch via exposure to large data sets associated with similar tasks. Therefore, the quality of training data a computer receives determines the effectiveness of the AI. However, AI systems are now beginning to use transfer learning to move lessons that they learn from one task and reuse them in performing other dissimilar tasks.
Transfer learning promises to speed up the pace of improvement in AI systems. As this advanced form of machine learning comes to the enterprise, businesses won’t always have to start from scratch when implementing AI and can more quickly adapt existing AI systems to new areas.
Increasing Value Today and Tomorrow
There is tremendous opportunity for enterprises to add value through AI today from focusing humans on what they do best to developing better software to digitizing the physical world. Near-term advances will only accelerate this value, deepening and spreading insights across the enterprise.
*DISCLAIMER: The portfolio companies identified and described herein do not represent all of the portfolio companies purchased, sold or recommended for funds advised by NextWorld Capital. Certain portfolio companies may be kept confidential for various reasons, including contractual or subject to a non-disclosure agreement. The reader should not assume that an investment in the portfolio companies identified was or will be profitable.
Not all acquisitions or IPOs are profitable; the positions can be acquired at a price that is greater or less than the price at which NextWorld Capital purchased its interest for client accounts. The information is being shown to reflect the firm’s ability to select investments and not to reflect any positive investment experience.
