5 AI Innovations

More companies are figuring out how to put AI and data to work for them or their clients. Here are five leading the way across different industries.

This article is part of our collection about the 2019 state of IT. Click here for more.

A CIO has always kept busy running the IT and computer systems that support an enterprise. Now with artificial intelligence (AI) turning every company into a data company, a CIO’s job is more central—and more complicated—than ever before.

A recent Gartner survey found that some of the world’s biggest companies plan to double the number of AI projects within the next year, with four AI efforts already in place, on average. Companies plan to spend more to make this happen, in the neighborhood of $77 billion in 2022, up from $24 billion in 2018, according to IDC. These projects include automating tasks, improving existing operations, and finding new insights. Companies are relying on CIOs to find out how to integrate AI into existing systems, what datasets they have that might be useful, and what problems a company has that AI can tackle.

More companies are figuring out how to put AI and data to work for them or their clients. Here are five leading the way across different industries.

Easing Tavel: Singapore’s Changi Airport

Singapore’s Changi Airport has won over 600 awards, most of which highlight amenities like its butterfly garden, an indoor waterfall, and free massage chairs scattered across the terminals. It’s also busy: Passenger traffic has gone from 37 million in 2009 to 66 million in 2018. A flight takes off about every 80 seconds, making it the 19th busiest airport in the world. [GK1]

To keep customers happy and terminals humming along, the Changi Airport Group, the airport’s operator, is investing in a program called SMART Airport Vision—its holistic approach that leverages sensors, data analytics, and AI. SMART Airport Vision launched in 2017 and is already making the airport a data-driven enterprise with numerous deployments in place.

Computer vision-powered cameras are a major part of the effort. These cameras are working around the airport to identify whether an abandoned piece of luggage is a threat, or if its owner stepped away for a moment, and notifying the correct team about it. Computer vision is also inspecting runways for foreign objects, such as gloves or keys dropped by baggage handlers, that cause damage to a plane or the runway. One terminal is testing out facial recognition systems that let passengers check in, drop off their bags, and get through immigration. It can even track down people running late for flights.

Besides cameras, an algorithm using historical data and current information, like weather, is trained to predict arrival times for long-haul flights with nearly a 95% accuracy. These planes have the largest amount of passengers and luggage; AI lets the airport deploy workers more efficiently and be prepared for an inflow of exiting passengers.

AI doesn’t stop at the airport doors. Changi Airport Group uses customer service chatbots to answer basic questions. Not only does that cut down on wait time for users, but their questions provide insight into the market demand for new service or product offerings.

Yet even with all the new tools, Changi Airport Group’s CIO Steve Lee is cognizant that technology alone is not enough to deliver world-class levels of service.

“Humans need to work symbiotically with AI-enabled technologies and machines to deliver the best outcome, and this supersedes what could otherwise be provided with only technology or only people,” Lee told CMO.com.

Faster Finances: AppZen

Big companies might have 80,000 invoices a month, which means looking at every line or double-checking all charges is nearly impossible. Even companies with 100 employees might get 1,500 monthly invoices, overwhelming a small team. Machines, however, are apt at this sort of close reading and can spot problems that even the most diligent auditor sometimes misses.

Fintech startup AppZen uses AI to automate auditing in real time, getting employees paid faster as well as catching fraud before any money is exchanged. Since launching in 2016, AppZen has signed up more than 1,000 enterprises for its services. AppZen takes care of their day-to-day invoicing and automate easy payments while freeing them to concentrate on a much smaller, high-risk list of potentially problematic spending. That might be duplicate invoices, fraudulent documents, or even payments that could potentially have the company out of line with the law if not done correctly. AppZen also can protect companies from paying fraudulent invoices.

“Humans cannot always tell the difference,” said Jamie Barnett, the CMO of AppZen. “But AI that is doing multiple models may be able to pick something up and flag it at as high-risk.”

AppZen also improves adherence to internal regulatory compliance rules. For example, using AppZen companies have found employees expensing things like guns, bullets, spa trips at a hotel, and cuff links. For some businesses, those might be appropriate purchases, but others would want to know about it.

“We aren’t the judges,” Barnett said. “We are here to help companies match their policies.”

Spotting Fashion Fakes: Entrupy

Counterfeits are a half-trillion-dollar business that supports organized crime and exploits workers in sweatshop-like conditions. It also diminishes a brand’s carefully crafted reputation, hurts sales, and forces companies to spend time and funds on fighting back. Entrupy is a startup using AI to detect counterfeit luxury items and get them removed from the supply chain faster.

To catch the fakes, Entrupy uses a set-up that can look at a diverse range of materials, from leather handbags to fragrances. It’s non-intrusive (and more importantly, non-damaging), using a microscopic camera attached to a smartphone to take a photo showing the “DNA” of the items. Then, an AI that has been trained on a dataset of authentic items dating back 80 years makes an assessment about whether the item is genuine.

“Every single thing in this universe has its own fingerprint,” said Entrupy’s CEO, Vidyuth Srinivasan.

Since the technology is viewing pieces at a microscopic level, it would be nearly impossible for a counterfeiter to replicate material closely enough to fool the system, Srinivasan said. There has only been a handful of false negatives, or missing a fake, due to user errors, like misusing the tool, he added.

Entrupy works with large-scale luxury retailers at the supply-chain level to combat return fraud, such as when someone tries to switch out an expensive luxury item for a counterfeit copy. The technology can tell whether an item been washed by the customer, suggesting it is not being returned unused. It is also working with government agencies, including Dubai Economy’s intellectual property department, to try to deter at the customs level before the items even make it into the marketplace.

Building Your Next Blockbuster: Cinelytic

Making a movie is considered an art form. But what if it’s also a data problem? Cinelytic is a Los Angeles-based startup taking historical data about movie performances and putting it into machine learning to find out what sorts of movie themes and potential stars could be combined to make the next blockbusters.

Hollywood has been run by taste and other subjective decision-making by a few stakeholders since its inception and is now one of the final frontiers for data analytics to prove it can add value—not that the industry is exactly shy about using technology.

“On a film set now, it’s robots, it’s drones, it’s super high-tech, but the business side hasn’t evolved in 20 years,” Cinelytic co-founder and CEO Tobias Queisser told The Verge .

With Cinelytic, moviemakers can consult data at every stage, from green-lighting a project to setting a budget to where to release it. An especially important tool forecasts how well a movie will do. Cinelytic looks at 15 attributes that they have found drive a film’s performance, including casting, whether it’s based on a book or a franchise, release size, and the rating. Having this data lets directors and producers plan the rollout or adjust budgets. If a casting director is considering multiple actors for one role, the software can look at what a star’s impact will be for release in different regions, media attention, box office vs. streaming, and across genres.

Better Patient Care: Nuance

Modern doctors have a lot of work beyond taking care of people. According to one study in the Annals of Internal Medicine, for every hour of face-to-face patient time, doctors spend another two hours on desk work, with data entry being the biggest time-consuming activity.

Doctors try to get some of that digital recording done while seeing the patient, but that has its own side effect: It might save time, but studies show that clinician’s gaze significantly affects how well an appointment is perceived by the patient. Nuance is a company using AI for its suite of software to try to reduce notetaking time, make patient data more accessible to clinicians, and maybe even improve the doctor office visit.

Nuance’s recently released Ambient Clinical Intelligence (ACI) is a tool with a hardware component that includes 16 microphones affixed around the room to record what both the patient and doctor say. ACI then transcribes it, removing the need for data entry in the room. ACI also uses voice biometrics for authentication, speaker identification, patient consent, and document signing.

Eventually, Nuance wants to add computer vision technology to spot and understand non-verbal cues, such as the number of sneezes that might be important for diagnosis, and automatically gather data such as body measurements, which would cut down on time spent behind a computer and also capture more information that a doctor would typically put in.

Besides catching important details and cutting down on paperwork, automating tasks might help prevent burnout for doctors, a condition that over 50% of doctors say they have experienced. Providing good patient care is the primary source of job satisfaction for 79% of physicians, according to The Physician Foundation. AI, surprisingly enough, could allow more time for the personal touch.

Would love an end quote