If you only paid attention to the hype about AI, you’d think our world was one algorithm away from either a frightening robot takeover or a fantastical machine-enabled utopia. But a closer look at the reality of how AI is evolving on the frontlines of business reveals a story that’s not nearly as far-fetched — but nevertheless driving a sizable impact.
Consider Moderna and Pfizer, two healthcare companies that employed AI technologies to rapidly test and develop COVID-19 vaccines months after the first reported cases. Intelligent computer models rapidly evaluated the effectiveness of vaccines-in-development, assessing immune responses across tens of thousands of virus subcomponents with a depth, speed, and accuracy that humans could never achieve on their own. AI cut the vaccine development timeline to mere months vs. decades.
The pharma companies mentioned above are global organizations with extensive resources for testing and development. How can companies that don’t have access to that depth of funding or talent begin to lay the foundation to take advantage of AI benefits?
For many leaders, distinguishing hype from hard facts is a chief hurdle stifling their AI journeys. Here are three key lessons to bring clarity to the AI conversation and illustrate how this very real, very practical technology can drive significant business value in the near-term.
Forget the moonshots. Start with the mundane.
In 2013, the MD Anderson Cancer Center at the University of Texas launched an ambitious AI initiative. The idea was to use IBM’s Watson cognitive computing system, aiming to speed the process of diagnosing patients and matching them with clinical trials. But the project never delivered on its promise. Four years and $62 million later, there was no record of the system used on a single patient.
Rest assured, there’s a silver lining. The cancer center’s IT team was quietly tinkering with the technology during that four-year period and delivering a host of efficiencies and improvements to the organization. Engineers instructed AI to recommend hotels and restaurants to patients, reach out to guests with personalized bill support, and improve the tech team’s workflow. These outcomes may not have been as sexy as the medical breakthroughs hoped for at the start of the project. But the results were undeniable: Guests were happier. Employees were more efficient. And financial gains realized.
The takeaway: Don’t get too swept up in grandiose planning. Start by looking at the mundane ways that AI can improve your day-to-day business operations.
AI is only a tool. Humanity is required.
Part of the mythology around AI stems from the notion that these technologies can “think” for themselves. After all, they can reportedly write articles, design clothes, and compose rock songs. These examples make it easy to understand the common fear that AI will replace humans in the workplace. In addition, recent research from the World Economic Forum estimates that AI will create 60+ million more jobs than it eliminates. That all said, it’s important to understand that these technologies are just tools — they still require a tremendous amount of human input to function.
More specifically, it comes down to data — and how well a workforce can collect and leverage it. For instance, Netflix’s dynamic optimizer generated lots of excitement about how it uses machine learning to enhance streaming quality on a scene-by-scene basis. But the tech doesn’t operate on its own — the data is gathered from human viewers tasked with manually assessing hundreds of thousands of shots. Similarly, there’s lots of talk about how conversational AI can potentially improve customer service experiences by engaging intelligently with customers via Natural Language Processing. But according to TrueLark, an AI-powered customer support platform, training that technology requires a tremendous amount of human input — hundreds of thousands of conversations.
The takeaway: AI is not a standalone solution, nor a silver bullet. Humans need to “partner” with the technology to learn from one another, extract value, and better serve the market.
Know what you want AI to do — and be specific.
It’s easy to imagine AI bots as helpers akin to Rosie from the Jetsons: Intelligent, catch-all assistants standing by to anticipate and fulfill a wide range of work-related tasks and needs. But the reality is that each AI application is actually highly-specific — the technology is not yet capable of general intelligence.
Just take it from MIT Sloan Professor Thomas W. Malone, director of the MIT Center for Collective Intelligence. Referencing IBM Watson’s notoriety for beating the world’s brightest Jeopardy champions, he says, “You think, wow, that machine must be really smart. But the truth is that the [version of the] program that beat the best players in Jeopardy couldn’t even play tic-tac-toe, much less chess.”
The reality is, most AIs can do one thing and one thing well. A given AI can:
- Reach deep into your systems and identify where you’ve been overpaying vendors.
- Automate marketing campaigns that correspond to how individual customers engage with your website or social media.
- Generate a more accurate actuarial model for insurers in real-time.
- Synthesize patient data from a range of platforms to create a single up-to-date health record.
These example tasks are hugely impactful for their respective business functions. But they’re nowhere near the general intelligence required for creativity, strategy, empathy, and so on. That’s still the exclusive domain of humans — and likely will be for the foreseeable future.
The takeaway: Don’t think too broadly about how AI can benefit your business. Start by drilling down into highly-specific business problems to identify the highest opportunities for ROI.
AI in Practice: A Transformation By Degrees
Continuous change is essential for remaining competitive in the digital environment. But when it comes to AI, what’s realistic is a transformation by degrees. Little by little, an organization that steadily integrates AI capabilities into its workforce and processes will see compounding rewards over time.
So don’t get too swept up in the hype — but do get started. When it comes to AI and business impact, slow-and-steady wins the race.
Want to build a data-driven academy? Get in touch to learn more.