Over a year and a half after the launch of ChatGPT, generative AI excitement remains fever-pitched. The consensus seems to be that this technology will utterly transform the way we work, create, and interact. The possibilities appear endless. But behind the scenes, most executives accountable for realizing economic gains from GenAI have not yet declared victory, and some are starting to sweat.
It’s one thing to be impressed by a chatbot’s ability to eloquently discuss any subject imaginable. It’s quite another to add a radically new capability to core business processes and workflows.
Today’s AI systems are still largely isolated from the beating heart of most organizations — enterprise resource planning, customer relationship management, human capital, and other digital nervous systems that keep the modern economy humming. Purchase orders, contracts, inventory management, employee benefits — many of these systems are behind firewalls, customized, and/or home-grown.
Simply put, a large majority of our country's GDP is tied up in systems of record and other mission critical systems that have been largely ignored by the AI community. To date, most AI companies have focused on building models, research, and sharing acronyms — GPT-X, context windows, RAG, graphRAG, perplexity scores — rather than building the interfaces (integrations and application embeddings) or the management systems (data access control, LLMOPs, and analytics). While the advancements in models and approaches have been incredible, the “making it actually work in enterprise” has been less than stellar.
Therefore, by the end of 2024, we’re likely to experience a foreseeable and pronounced backlash in the GenAI community focused on enterprise. CIOs have been given mandates and targets to implement GenAI and realize ROI without the tooling to do so in the places where it matters.
The unglamorous work that will determine whether the generative AI revolution lives up to its potential will be in building the bridges between these cutting-edge technologies and the legacy systems that have long served as the backbone of business operations. Not just a shiny new interface layer, but a transformative augmentation of the core systems and processes that drive real value.
This is no small challenge. Many organizations are saddled with decades-old tech stacks, held together by brittle integrations and mountains of custom code. The idea of plugging a state-of-the-art language model into this tangled web can seem daunting at best, impossible at worst. It will also demand a re-thinking of long-standing organizational boundaries and silos. In a world where AI models can route a customer complaint to the right resolution team, flag a supply chain issue for immediate attention, or surface a hidden opportunity in financial data, the lines between IT, operations, marketing, and other functions will blur. Cross-functional collaboration and agile ways of working will become table stakes.
None of this will happen overnight. The process of AI-enabled business transformation will be measured in years, not months. It will require not just technical integration, but also deep changes to processes, skills, and mindsets. There will be missteps and setbacks along the way — as with any complex, systemically important technology, responsible development and governance will be critical. But for organizations that navigate this journey successfully, the rewards could be transformative. By weaving AI deeply into the fabric of their operations, companies will not only see efficiency gains and cost savings, but also unlock entirely new ways of creating value for customers and stakeholders. They’ll be able to respond to changing market conditions with agility, surface insights that would have remained hidden in the past, and automate and augment tasks in ways that free up human ingenuity for higher-order challenges.
This is the real promise of generative AI in the enterprise — not as a flashy toy or a bolt-on curiosity, but as a fundamental rewiring of how business gets done. And while the sci-fi-esque demos might get all the attention, the true heroes of this revolution will be the ones doing the hard, unglamorous work of integration and operationalization.
They’ll be the ones building the middleware that allows data and insights to flow freely between AI models and legacy systems. They’ll be the ones re-engineering processes and upskilling workers to take full advantage of these new tools. And they’ll be the ones who steadily, perhaps even stealthily, weave generative AI into the critical systems and workflows that power their organizations.
So by all means, marvel at the output of cutting-edge AI systems. But don’t lose sight of the bigger picture. The missing link between generative AI’s incredible potential and its real-world impact is the hard work of connecting it to the systems and processes that drive business forward. Therein lies the true transformative opportunity.
