As generative AI technology races forward, mastering systems thinking has become indispensable. Large language models (LLMs) like ChatGPT boast impressive capabilities, yet their current applications often fall short of delivering the results users want with a single, well-crafted prompt. While some envision autonomous AI agents in the near future, to most users, LLMs in their current form are little more than thought partners at best, and at worst, lyrical masterminds spinning haikus about folding laundry.
Effective Prompting and Retrieval Augmented Generation (RAG)
Systems thinking is the essential skill to bridge the gap to more advanced AI agents. While LLMs can produce seemingly coherent and structured outputs, to truly trust and utilize these outputs, a requisite level of expertise about the subject is necessary. Since LLMs cannot predict exactly what you're thinking, they often lack the context needed to provide accurate answers. Effective prompting addresses this by surfacing relevant content and subsequently querying the LLM.
Prompting systems show strong similarities to RAG—a leading AI engineering technique that involves loading an LLM’s context window with key information by retrieving it from relevant documents. In these use cases, structuring prompts effectively to guide the model’s responses and ensure relevance across subsequent interactions is key to achieving optimal outcomes. For developers, this systematic method of thinking bridges the gap from early LLMs to what will likely be early agents, essentially just stacks of various AI tools.
Systems Thinking for Business
Systems thinking is essential not only for engineers but also for business leaders. Just as AI agents are built systematically, organizations rely on the design of their systems and processes—a manifestation of a company’s collective historical skills and expertise—to drive decisions and actions. AI advances signal new opportunities to redesign organizations, and pioneering employees are critical to driving this change. With the abundance of AI tools available, making the right choices can be challenging, as procurement processes at organizations can be long and drawn out. For many enterprises, this can take more than 12 months. For better, faster results, individuals should be encouraged to identify and solve their challenges with AI tools. Successful pursuits will lead to organizational redesign of systems to improve efficiency.
When employees grasp the current strengths of AI, they can pinpoint opportunities to streamline their own tasks and processes using AI tools. Teams equipped with systems thinkers who understand both system design and AI's capabilities are poised to excel. These will be the 10x employees—those who amplify their impact and drive substantial value through innovative problem-solving and strategic application of AI.