Smart Investors Support Enterprise AI that Creates Jobs and Helps Employees 

Entrepreneurs and investors regularly dismiss claims that artificial intelligence will replace people in the workplace. AI needs humans to function, they say, and AI doesn’t want to take the place of workers.

This response rings hollow to critics, however, because the assumption remains that humans become an afterthought once leaders broadly deploy AI within an organization. And nobody, critics argue, should lead or invest in a business that puts people second.

But here's the thing -- AI actually grows stronger and more valuable through collaboration with people who make AI succeed in the first place, which then accelerates that mutual success. It is humans who teach AI how to function in workflows. The more lessons that AI can learn from humans while helping people achieve their business goals, the more an organization, its workers and investors benefit on a compounding and ongoing basis.

The Real Problem

Predictions of the rise of the machines have already begun to feel like crying wolf, however, because AI needs more, not fewer, people in order to improve and succeed.

The real problem is that there are not enough data scientists to go around to use the most powerful machine learning algorithms effectively. Some companies are helping businesses overcome that obstacle by assembling domain experts with data scientists and AI experts to build useful industry applications and processes. Vertical SaaS companies as well as firms composed of many SaaS businesses focused on separate industries are especially paving the way. But progress in realizing AI’s potential is still slower than it should be.

The smart money is being invested into AI that is easier to use and cuts down on the time it takes for developers and employees to use it for critical tasks and workflows.  That’s where the massive value is waiting to be unlocked – in the plethora of AI platforms and machine learning algorithms that professionals other than data scientists can readily understand.

Let’s look at healthcare as an example.

Case Study: Northwell Health

Enterprise AI has already demonstrated its value to large healthcare systems that possess the data scientists and other resources to implement it.

Consider Northwell Health in the New York City metro area. COVID-19 hit Northwell, the state’s largest health system, hard.

Because its enterprise AI and automation infrastructure tools and systems had been put in place as much as 15 years earlier, the tools helped its pandemic response significantly. Those AI capabilities, however, required a long and intricate process to come to fruition.

The process for Northwell’s cardiac imaging team at Lenox Hill Hospital developed long before the pandemic when it became clear that increasing workloads required streamlined and improved imaging automation and AI technologies and procedures. The health system’s scale and a growing volume of imaging meant they had to explore automation to keep up with demand. Over time, the success of AI and automation in coping with those needs then expanded to cardiac CTA imaging, calcium scoring,  mitral valve clipping, aortic surgery, congenital heart disease and cardiac MRI evaluation.

When the COVID-19 pandemic surged by mid-2020, Northwell was able to bring forward a patient database of vitals, laboratory results and medications, as well as continuous data capture to cope with the surge in new data from the pandemic. Medical teams could use regularly updated demographic information to track early clinical outcomes of more than 6,000 hospitalized COVID-19 patients in the region. The collaboration between people and technology yielded important victories against the disease. Clinicians at Northwell Health were among the first to recognize a correlation between the virus and heart disease complications, for example.

Democratizing AI?

Northwell’s experience has been replicated by other large institutions like Intermountain Healthcare, Mount Sinai and Mercy Health. But enterprise AI will not become a reality until community hospitals with fewer than 400 beds can put it to work, too.

There are outliers, though, where smaller community hospitals are taking up the AI torch, including in the Sun Belt.  A Florida community hospital is using AI tolls and topological data analysis, or TDA, to develop new care paths for high-cost, high-mortality conditions. The first condition was pneumonia. After teaching as many medical staffers as possible to use AI and TDA, the hospital established a new care path that segmented patients according to the care that was most efficient and which reduced re-admissions. The re-admission rate for pneumonia, which was almost three percent before the AI and TDA implementation, dropped to 0.4 percent after the new processes were implemented. The hospital moved on using the same systems to reduce sepsis cases and continued to expand to other medical issues from there.

What Does Success Look Like?

Successful enterprise AI platforms exhibit three characteristics. They are explainable to whoever uses them. They demonstrate exactly how they benefit organizations and users. And they are interactive, meaning users can adjust the technology to their needs, while being accessible to a broad range of users.

If an industry as complex as healthcare can illustrate why enterprise AI can work without the deployment and involvement of an army data scientists, imagine what virtuous cycles it can produce in other spaces. Data and discoveries lead to conclusions and insights. That delivers more data and discoveries. Leaders who assemble teams of people who can harness AI to foster this process can achieve staggering value.

About the Author

Pradyut Shah is a senior investment partner at SymphonyAI, overseeing all investment and acquisition activity at the firm. He currently serves as a director on the boards of Symphony IndustrialAI, Symphony MediaAI and TeraRecon.