The AI revolution is in full force, with businesses everywhere looking to Generative AI to transform their organizations. This next step in how humans interact with computers promises to upend virtually every industry on the planet.
The history of technology is marked by before-and-after moments – advances that spurred new industries, enabled the transformation of others, changed how people interact with technology, and reshaped society. The public rollout of the Netscape web browser in 1994 was one such moment, the introduction of the Apple iPhone just 16 years ago another. We’re now living through an equally – if not more – significant moment with the emergence of generative AI. While the consumer applications of generative AI have dazzled us all, with ChatGPT and other services churning out remarkably polished essays and high-quality images in a flash, the implications for the enterprise are even greater. This is why business leaders across industries are examining where and how to incorporate generative AI throughout their organizations. While many large enterprises, such as Koch Industries and the U.S. Air Force, have long been digitally transforming with enterprise AI application software, the emergence of generative AI promises to accelerate such efforts at a rapid pace. The opportunities are enormous; the risks of watching from the sidelines just too great.
As the name suggests, generative AI enables computers to generate content based on a set of AI and Machine Learning (ML) algorithms applied to vast data sources. The user experience, akin to having a conversation, marks a fundamental shift in the human-computer interaction model. Generative AI enables software to “learn” the fundamental patterns of a corpus of images, text, and audio files, and then rapidly produce comprehensive, thorough results (although at this point inaccuracies aren’t uncommon.) These AI models use a variety of techniques, such as transformer models, generative adversarial networks (GANs), and variational auto-encoders. (See definitions).
A user types in a prompt, the software spits out what you asked for. Beyond ChatGPT, there’s now Stability.ai, Google’s Bard, and Microsoft’s generative-AI equipped Bing, with the tagline, “Ask real questions. Get complete answers.” To generate content—churning out an essay about Plato, for instance, or about the history of AI technology—generative AI services rely entirely on publicly available data. And that’s what stands between current generative AI offerings, and what’s needed to serve the enterprise.
Access to the appropriate enterprise data is the essential ingredient for generative AI for business. A financial services company, for example, doesn’t need a generative AI system that learns only from public data. It needs a system built for its domain (read The Significance of Domain Models) that generates comprehensive insights from its proprietary data. That might include deposit trends, information about its loans, and so on. Ditto for a manufacturing company, or healthcare institution. Generative AI for the enterprises also incorporates public data via Large Language Models (LLMs). A financial services firm would want real-time interest rate data in the mix, for instance. But it’s the domain-specific, enterprise data that promises to make generative AI so transformative for businesses.
The difference doesn’t stop there. Generative AI makes it possible for far more people across an organization to take advantage of the predictive insights generated by the underlying enterprise AI applications. Put simply: Generative AI available to the public creates what you ask for based on current (meaning past) data available across the web. Generative AI in the enterprise, by contrast, can produce vital insights based on what’s going to happen for a specific business, such as telling you which parts of your manufacturing facility will need maintenance when or which customers are likely to close their accounts.