Has enterprise artificial intelligence (AI) lived up to the hype generated at a decade’s worth of industry conferences? Or is it coming up short? Maybe putting the word “enterprise” in front of AI just adds up to a marketing spin. It depends on how individual businesses deploy AI.
When companies adopt AI wisely, they do more than shift repeatable tasks and processes from humans to more efficient computers. They bring humans and machines together to build more intelligent workflow — transformational workflows.
What Makes Enterprise AI Different From Any Other AI?
The enterprise AI-focused operating company SymphonyAI has been earning headlines for its strategy. Its portfolio companies have been making inroads in the industry verticals they each address, including Symphony IndustrialAI. With the recent acquisition of Savigent, Symphony AyasdiAI in banking, and Symphony MediaAI in the business of subscription and media distribution revenue, including gaming.
In data ops for private capital, Harmonate has been leading a quiet revolution in how private equity and funds-of-funds middle and back offices operate with machine learning.
Humans and machines together can achieve more, in a more repeatable and reliable fashion, and with better insight. But apart from some funds and companies, is that actually happening throughout the economy?
Where is the money going?
No, and yes. Money is being poured into AI, and it’s making a difference. It’s just that the difference being made is not necessarily visible. This lack of visibility fuels skeptics. And the progress is not fast, given that the availability of huge amounts of data is both a blessing and a curse. Copious data delivers the raw material AI needs. But AI is still learning how to cope with the complexity and needs help from human domain experts.
The smart companies are the ones that are not tinkering and failing to make big moves. And the smart companies also aren’t trying to leap too far ahead with moonshots that skip steps.
What the smart companies are doing is putting together point solutions into products that solve real business enterprise solutions. They are developing the right loop between domain experts and machines. The result is real AI product suites that capture the knowledge capital of enterprises and can transform industries.
We all know AI investments have been increasing in recent years. Skeptics would say the trend derives from big promises and false expectations. But I’m compelled to think many companies are deploying AI more wisely than we understand. They are discovering value and growing the potential of AI.
It’s just happening in quieter corners of business enterprises. It’s happening in places where domain experts and the right technologists are solving small problems, then connecting those breakthroughs to others, until there’s an inflection point. There’s a germination period underway right now.
We are moving from a diffuse cloud of point solutions to product suites in industry verticals powered by business leaders who’ve embraced the new reality of their markets.
When do I get my flying car?
AI skeptics, however, persist in believing that artificial intelligence advances are like flying cars – a sci-fi fantasy that has failed to materialize despite years of hopes and promises. It’s true that optimistic predictions have sometimes outstripped the reality of AI.
By one estimate, AI has been through seven false starts since the 1950s. Impressive multimillion-dollar AI efforts have faltered. Some ostensible “AI startups” aren’t even really using AI but rather are selling automation with elements of machine learning. This poor performance and confusion fuels skepticism, inhibits innovation, wastes money and reduces returns.
Most investor enthusiasm for AI is based on sound logic, however. AI tools have evolved from defeating humans at chess. Machines are good at recognizing patterns, a powerful and important cognitive function.
And, in fact, processing patterns are humanity’s intellectual edge over other species. It also accounts for many daily business tasks that AI-driven machines can now frequently do better than humans across a range of sectors. The results are driving enhanced AI chips that reduce costs and dramatically improve performance.
But those chips are also being driven by the fact that repeatable tasks can be deceiving. When multiple choices of what to do lead to many more multiples of options. Even AI can start to lose track of where it’s going. Experience with humans, and more chip power can bridge that gap.
More to work with
There is a lot more data to process today, too, which means more potential value. Thanks to the internet, social media, connected devices and the Internet of Things, total extant data exceeds 40 zetabytes, a ten-fold increase since 2013.
There are now “40 times more bytes than there are stars in the observable universe,” according to the World Economic Forum. Cloud computing has facilitated elastic consumption of storage and network demands to handle that data. Digital transformations have resulted.
A growing number of companies are recognizing the benefits. AI adoption tripled in the 12 months leading up to March 2019, perhaps “the fastest paradigm shift in technology history” according to a major study. PWC forecasts that AI could add $15.7 trillion to the global economy by 2030.
AI is not a fad. It is a key differentiator. Like the internet, it has the potential to completely transform the economy. Companies that deploy it effectively will make changes.
How to Transform a Business with Enterprise AI
Of course, companies can possess all the ingredients necessary to conduct top-performing AI analysis but still fail to achieve results, particularly if they lack a robust understanding of their industry’s business processes. Human perspective and insight are more art than science. Inspiring the former while developing the latter is the challenge we all face in the new AI age we’re now in the middle of.
Companies sometimes tinker, improving obsolete systems rather than rethinking and reinventing their operations to capitalize on enterprise AI.
Tinkering is good. But tinkering too long leads to a flawed approach that may help a company reduce its costs or streamline processes in the short run. But such gains are unlikely to justify the investment needed to gain significant market share.
Worse yet, the company will have missed an opportunity to achieve a transformational advantage, one that competitors may be exploiting.
Adding to the problems with tinkering are startups seeking to harness AI for individual point solutions. Their value proposition is harder to figure out. The potential for differentiation is typically diminished, and their survivability is less certain. A task and a point solution are not a business enterprise.
The middle way
Companies don’t face a choice of incremental change or narrow focus, however. Instead, established and new ventures need to harness enterprise AI’s capacity to capture and profit from the knowledge capital in their given sectors.
In 1998, Paul Strassmann argued that the proper function of the software is to serve as the business’s “prefrontal cortex,” storing and exploiting the working knowledge that has traditionally remained stuck in employees’ heads. When applied correctly, enterprise AI is the ideal technology for this work.
The goal of enterprise AI is not only to empower humans but also to program and institutionalize stronger, smarter, more efficient organizations.
Enterprise AI can expedite those changes because, unlike traditional software, which follows the static instructions of a programmer, AI can evolve to capture a wider variety of tasks and learns through practice.
Furthermore, enterprise AI is undaunted by the many terabytes of data that companies gather. It quickly observes complex and obscure patterns that humans miss.
That’s why forward-looking companies are using it to build next-generation platforms – systems of actionable intelligence that capture siloed data from existing systems of record. The enterprise AI solution makes this data available in a holistic way, through a set of AI models, applications and solutions.
These platforms also acquire and integrate data from external sources, providing intelligence for further revenue growth.
Businesses will need a vision for “AI-ification” if they want to rethink their operations, transform their technology stacks, overhaul existing solutions and win in the future. And we’re fast approaching the point where it’s not a question of wanting to rethink, but needing to rethink.