HBR Guide to Generative AI for Managers

With the advent of generative AI models, managers can talk with machines in natural language, asking them to collaborate on tasks and engage in conversations.

With the advent of generative AI models, managers can talk with machines in natural language, asking them to collaborate on tasks and engage in conversations. Evolution of machines. Historically, machines served as tools, automating tasks and streamlining processes, but creativity and strategic thinking remained human domains. The introduction of AI virtual assistants marked a transitional phase, allowing simple natural language interactions, though often feeling robotic and limited. This era set the stage for a more profound shift. The game-changer. OpenAI's ChatGPT, launched in November 2022, revolutionized human-AI interaction by offering a truly conversational interface. This made AI accessible to everyone, regardless of coding skills, enabling fluid, human-like dialogues. Its rapid adoption underscored a fundamental change: machines were no longer just enablers but potential partners. New paradigm. Generative AI ushers in an era where machines become collaborators, capable of understanding and generating diverse content (text, images, audio, code). This symbiotic relationship means AI can now assist managers across a wide spectrum of tasks, from routine operations to complex problem-solving, fundamentally transforming how work is done.

Co-Thinking represents a process done together, as the “co” implies, where human and AI intertwine. Co-Pilot for execution. In Co-Pilot mode, generative AI acts as an efficient assistant, handling administrative, communication, and operational tasks. Managers provide clear prompts, and AI executes, drafting emails, summarizing documents, or creating slides. The manager's role is to provide initial direction, then review and validate the AI's output, ensuring accuracy and quality. Prompting is key: Be precise, contextual, specify format, give examples, set boundaries, and don't hesitate to ask AI for help in crafting prompts. Not autopilot: Always verify AI-generated content to avoid fabrications or hallucinations. Co-Thinker for strategy. In Co-Thinker mode, AI becomes a thought partner, engaging in deep, reflective dialogue to aid strategic thinking and problem-solving. This involves a back-and-forth exchange where both human and AI contribute ideas, challenge assumptions, and build upon each other's inputs. Preparation is crucial: Assign AI a specific role (e.g., mentor, devil's advocate), define the setting (one-to-one, one-to-many), outline the dialogue flow, and then create a structured prompt. Focus on reflection: Unlike Co-Pilot's speed, Co-Thinker emphasizes thoughtful conversation, encouraging pauses and deeper analysis. Blended interaction. These two modes are not mutually exclusive; they can be used sequentially. For instance, a manager might use AI as a Co-Thinker to brainstorm a strategy, then switch to Co-Pilot to draft a memo summarizing the discussion. This dynamic interplay optimizes both efficiency and strategic depth.