What We Can Learn From These 10 Companies That Are Implementing AI/ML in Unique and Exciting Ways
According to a 2019 Gartner survey of three thousand CIOs across a variety of industries, the number of enterprises that utilize AI technology has grown roughly 230% over the past four years. This is a particularly startling number for companies who have yet to integrate this transformative technology, since its growth has placed the estimated percentage of organizations that employ some form of AI at 37%. Some executives, leading companies within the other 63%, may brush these stats off, believing that they don’t stand to benefit enough from AI right now to justify investing in significant digital transformation.
These companies, however, may find themselves playing catch up with competitors who have found creative ways to leverage this technology within their specific industries. Just take a look at these 10 trailblazing companies who are teaching us the value of modern technical innovation through the fascinating ways that they use AI!
Blue River Tech: Shrinking Chemical Use in Agriculture
It has been nearly 60 years since the publication of Rachel Carlson’s Silent Spring, which decried the indiscriminate use of pesticides and herbicides in commercial agriculture. Though we have seen different chemical and policy changes attempt to address the issues raised in the text, none may be as revolutionary as Blue River Tech’s AI driven farming equipment.
Founded in 2011, the company has since been incorporated by leading agricultural machinery company, John Deere, in an effort to grow its flagship “See & Spray” product. This device uses deep learning algorithms, similar to those used in facial recognition software, to discern weeds from crops, and apply appropriate amounts of chemical herbicides. Not only does this decrease the need for farm managers to employ, and train multiple workers to spot and correctly treat invasive, and highly threatening plants, but it can, according to Business Development VP, Ben Chostner, “save up to 90% of the amount of herbicide that you would spray if you were spraying the entire field”.
Cinalytic: Modeling the Financial Success of Big Budget Movies
As the ways that audiences consume film advertisements continues to fragment by the day, the cost for marketing films-- one of the largest areas of investment for production companies-- is growing exponentially.
However, this troubling news comes with a silver lining. According to 2018 data compiled by IBIS Worldwide, the compound annual growth rate for the American film industry is roughly 2%, which outpaces the nation’s overall economic growth. This saddles production companies with the task of finding new ways to optimize ROI at a time where there has never been so much money to make, but where it also requires such a financial investment to make a movie.
This is where Cinalytics steps in. The LA based startup’s AI powered platform considers fifteen unique attributes that can predict the potential success of a film. Using historical data about the performance of thousands of films, the program is able to identify relationships between the financial success of films, and factors like casting, genre, and ratings, to forecast the revenue that producers might be able to expect from any given project. Users can then make changes, such as swapping out one actor for another, or changing the scale of the film’s release, to see how different pre-production decisions may impact their performance metrics.
Not looking to remove the human element from filmmaking, CEO Tobias Queisser believes his company’s tool can “supplement the creative process”, and help producers think through artistic decisions with the added confidence provided by predictive data.
Brain Power: Augmenting Reality for People on the Autism Spectrum
According to leading educational non-profit Autism Speaks, 1 in every 59 children born today will fall somewhere on the autism spectrum. Although symptoms are diverse, many individuals with this diagnosis struggle with typical social conventions in a manner that impacts their daily lives. As autism is typically diagnosed in early childhood, many families are tasked with the responsibility of helping their children navigate a world that better accommodates neurotypical behaviors.
Brain Power is using transformative technologies, including AI and AR (augmented reality) to provide autistic children, and adults, with learning experiences that help them with daily tasks and life skills. Having consulted a diverse array of families, healthcare providers, and counselors, and ran clinical trials, the company has created software that allows users to build skills such as identifying emotions based on facial cues, and maintaining eye contact.
Utilizing Google Glass, the software’s zero UI interface reacts to the user’s eye motion while simultaneously reading their environment to detect not only the presence of other people, but also another person’s facial expressions. In one demo, featured on a PBS News Hour, Laura Krieger, the mother of an eight year old child with autism, plays a game with her son where she emotes surprise, which Brain Power’s software reads and then prompts her son to identify among two possible options. This, however, is just one of the wide array of skill building tools and games that Brain Power has delivered through its cutting edge software suite.
Stanford Computational Policy Lab: Standardizing Judicial Practices
Bias mitigation is a significant concern in modern policing. When many think about AI implementation in police and judicial processes, they shudder at the potential risks. Because AI/ML relies on prior data, some suggest that implementing AI software will increase bias given the historical over-policing of certain groups in America. Though they haven’t created a predictive software, Stanford’s Computational Policy Lab has pursued a unique way to combat racial prejudice in the judicial, and very well may be laying the groundwork for creating more equitable crime predicting AI in the future.
Stanford’s Computational Policy Lab recently partnered with the San Francisco Police Department to develop a tool that strips arrest records of information that might evoke such conscious or unconscious biases when being presented to the District Attorney’s Office. The software uses name-entity recognition technology to identify and remove not only an arrestee’s race, but also descriptive factors that might allude to a suspect’s demographic information from unstructured text. These include physical descriptions like eye, and hair color, names, locations, and neighborhoods where the subject lives or was arrested. The software will also remove information alluding to the identity of involved officers, including names, and badge numbers to even further prevent the DA’s office from making unfair inferences. The goal of this is to reduce or eliminate the probability that racial bias might impact what charges are levied against a suspect.
Though San Francisco commissioned the non-profit to develop this software, Stanford has delivered this software to the city at no charge, and intends to release the software widely for any city to use.
IBM and McCormick & Company: Enhancing Human Creativity
Remember in 2011, when IBM’s Watson annihilated champion contestants Ken Jennings and Brad Rutter on a special episode of Jeopardy? Well Watson has since stepped away from the podium and into the kitchen.
Through a partnership with McCormick & Company, IBM has deployed “Chef Watson” to generate AI conceived spices that will be sold under the iconic seasoning and condiment company’s branding. The company hopes to eliminate the need for consulting often dozens of developers to create a product by leveraging their decades of recipes in order to invent standardized metrics for understanding flavor as data. This removes the issue of developer preference and bias by assigning objective metrics to ingredients that then can be combined to create highly original recipes that are less conventionally referential, and thus less easily replicated.
This technology will not totally remove the human element of product development, but may lead to a 70% reduction in labor according to Hamed Faridi, the company’s Chief Science Officer. Since not all elements of the flavors can be easily and totally objectifiable, there still remains a need for human testing and augmentation. However, as McCormick believes, using IBM’s technology will allow even less experienced product developers to work as efficiently as developers with twenty years of experience. And the company believes its first AI generated product will be released by the end of 2019.
Nauto: Uniting with Human Drivers to Create Safer Roads
Though Stanford researchers have suggested that there will be 10 million self driving cars on the road by the year 2020, concerns still abound regarding the safety of full automated vehicles. Even though a McKinsey study has shown that 90% of driving related deaths could have been avoided had the drivers been using AI powered vehicles, there is something unnerving, for some, about placing their lives in the hands of a computer, and at this time, the technology is prohibitively expensive for most. And this concern multiplies for fleet companies, whose vehicles are both extremely expensive to own as it is, and capable of doing considerably more damage in an accident scenario.
Nauto, the commercial fleet safety company, has created Prevent, an AI-powered device that employs deep learning algorithms to identify when drivers are distracted based upon their facial expressions, and eye lines, as well as their spatial relations to other cars, and whether they commit traffic violations. Then, the system is able to notify the driver, as well as the driver’s employer, in real time, about the nature of their risky behavior. However, CEO Stefan Heck stressed, in an interview with Tech Crunch, that the technology’s intention is, “to be really focused on keeping the driver safe without being intrusive… We want to help human drivers, not just rat them out to their boss.”
As of the publication of this article, Nauto reports that it has detected nearly 49 million high risk driving events, and have saved commercial fleet companies an estimated 976 million dollars.
Inspirata: Driving Precision Medicine
For individuals who develop cancer, early detection greatly improves their chances of being able to successfully treat the illness both in the short and long-term. However, accurate and timely detection is not always easy when the signs that pathologists need to identify in order to diagnose and treat cancers consists of such a wide breadth of information.
Inspirata, an oncological informatics company, is helping us realize a future where our doctors can utilize the pinpoint accuracy of artificial intelligence to pull critical data from not only our physical sample slides, but also our radiology narratives, as well as unstructured clinical text. These softwares rely on deep learning algorithms that collect user input and are able store that data in to make inferences about future input that it encounters. For example, if a pathologist identifies a cluster of cancerous cells in a sample slide and reports this finding to the software, the software can use that information to help future pathologists identify similar cases. This, of course, also relies on the software’s astounding capacity to map the structure of human cells and quantify that data in a manner that can be understood and reapplied in the future.
Similarly, Inspirata has continued to expand its suite of workflow tools to include natural language processing that extracts data from millions of clinical texts deriving from a network of over 400 global healthcare providers. In this sense, doctors are not only able to benefit from the forward progression of medical technology, but this implementation of machine learning, in the words of Inspirata Founder and Executive VP, Dr. Mark Lloyd, “gives pathologists the power to not only contribute to precision medicine, but to drive precision medicine.”
LeadGenius: Automating the B2B Sales Pipeline
In modern sales, there are generally two broadly defined strategies: old school, reliant on cold calls, handshakes, and remembering the names of your clients’ kids, and new school, where representatives connect over Linkedin, and hook potential clients by pushing material relevant to their product or service. But is there something even newer than “new school”?
Not too unlike the way that Inspirata generates and leverages data by relying on doctors to teach its AI technology to identify diagnostic data, LeadGenius benefits from the talent and know-how of over 500 researchers from 40 countries to train its software to do much of the legwork involved in sales.
LeadGenius uses crawlers to search the web for potential clients, and is able to identify and isolate roughly thirty data points about a company, including its name, industry, revenue, and the technologies it uses, in order to determine the business’ suitability as a customer. Since much of this information is not contextualized or well structured, LeadGenius’ software must not only rely on language processing, but must also be trained by seasoned salespeople who know how to find, and identify this crucial data. The software is even capable of emailing prospective clients using natural language as if developed by a salesperson themselves.
In a 2015 talk at a Silicon Valley Data Engineering Meetup, LeadGenius Co-Founder and Chief Scientist, Anand Kulkari made the startling claim that machines will replace salespeople in 10 years. But this doesn’t mean that salespeople will be out of work- what it means is that professionals will have to spend far less time searching out companies, analyzing their buying patterns, and initiating contact. This will allow representatives the ability to get in front of more potential customers and simply focus on closing.
Solutions 4 Health: Tailoring Support for Cigarette Cessation
According to the CDC, 68% of US smokers report wanting to quit. However, in the same year that these statistics were obtained, less than 8% of adult smokers reported being able to successfully give up smoking that same year. For many smokers, the support of tailored counseling and monitoring can be a significant motivator to maintain their cessation journey. That being said, lack of access to these services, which are often cost prohibitive or otherwise difficult to access, can leave many looking to lead healthier lives without any support.
Leading health technology company, Solutions 4 Health, is on a mission to address such healthcare inequalities with its line of various AI powered innovations. One of these technologies is “Quit with Bella”, an iOS mobile application that allows users to text and even speak with an intelligent chatbot, who offers responsive, personalized support to help them quit smoking.
Inturn: Strategizing Supply for Retailers
Overstock is one of the most crippling sources of waste for millions of global retailers, with the current market creating a 170 billion dollar annual problem. Many, if not all, major retailers depend on being able to sell their overstock to discount and resale companies to reduce the loss in revue. However this can be a weeks long process for retailers, which often involves massive excel spreadsheets, and moving their inventory from showroom to showroom. The time and difficulty of this practice not only results in increased labor costs, and decreased human productivity, but also places companies under immense pressure to move their products before products and fashions become out of date.
Inturn is leveraging the power of AI to help companies sell their overstock more quickly and strategically. Founded in 2013, the company offers a suite of different tools to automate what is often the impossible manual task of identifying which products are not selling at appropriate volumes, and expediting their liquidation to discount retailers. The software relies on analyzing large volumes of legacy data across multiple sources to cross reference the sales patterns of a retailer’s products based on multiple categorical factors such as size, color, style, fabric. Not only allowing companies to better strategize when they move inventory to the discount supply market, the software can also aggregate all of the information contained across multiple inventory databases, and present that information on a single platform for easy use by interested buyers. This allows companies to reach more potential resellers, secure better deals, and ultimately move their overstock more quickly, increasing margins by an average of 23%.
The purpose of this article is not just to highlight 10 companies that are utilizing AI to disrupt and advance their particular industries. The creative implementation of AI, Machine Learning, and other transformative technologies is on the rise, and with it, comes their demystification. The technological capabilities of AI are vast and have yet to be fully explored, however, our baseline understanding of its innumerable capacities, and the ways to unlock these functions is there.
In putting together this list, we attempted to find companies who are each implementing digital transformation in unique ways. However, when you explore the technologies that power their products and services, you realize the concepts behind them aren’t all that dissimilar. Both Blue River Technologies, and Nauto have leveraged video-based object-recognition to address devastating problems unique to their industries. Solutions 4 Health, and the Stanford Computational Policy Lab operate in wholly unrelated spaces with drastically different goals. However, both are using machines to isolate, identify, and extract data from uncontextualized human speech and writing. Even Cinalytic, the film pre-production advisor, is utilizing very similar data analytics concepts that McCormick has employed to help them develop the next big food seasoning!
If you or the company you work for has not yet investigated how AI and machine learning can help you better automate your workflow and services, while drastically reducing waste and maximizing profit, you are not alone. Remember that an estimated 63% of major enterprises have yet to successfully integrate digital transformation into their operations! That being said, the International Data Corporation (IDC) predicts that companies will spend roughly $1.25 trillion dollars by the end of 2019 on their transformative digital journeys.
It is clear that major corporations are recognizing the long-term investment value for AI and ML integration. As these companies and organizations continue to invest, they are likely to see significant improvements not only in the quality of their products, services, and work processes, but will be setting themselves up for long-term economic growth that we can not possibly imagine at this time. Continued investment also gives these companies a great foundation off of which they can grow their technologies through further technical integration as new problems arise, and new capabilities are imaginable.
Beginning your process of digital integration doesn’t have to start with a multi-million dollar overhaul of your technical stacks. It can start with small improvements to the functionality of your existing technologies, or creating simple software to better automate services you already provide. Where you do not want to find yourself, as a business, is being in the position where, in order to compete with companies in your own industry, you are forced to make significant investments over a short period of time, to meet the new standards created by these revolutionary technologies. There has never been a better time to start your journey. So many companies have done the legwork of not only creating the multiple, diverse, open source technologies that are leveraged to achieve these advancements, but they have also done the equally difficult work of imagining ways to effectively apply these technologies to challenges that impact their businesses and industries. As an unintegrated business, you have a unique opportunity to expand your technical program at a time where the science and concepts abound, but where the features offered by AI and ML are not yet the baseline standard for enterprises in the modern world. It now falls upon C-level executives to make the often difficult, but obvious choice to begin transforming their digital technologies now.
The future of your business depends on it....