The Future of Work: To Prompt or Not to Prompt

“The future is already here — it's just not evenly distributed.” — William Gibson

Not so hot take here: AI in the workplace is no longer a novelty—it’s a necessity. According to 麦肯锡 , AI adoption has surged to 72%, up from 50% in recent years. It’s clear: AI’s influence on our professional lives will continue to grow and evolve. The question now is how our roles will transform alongside this technology and what the future of office work will look like.

I see the answer to this question unfolding in one of two scenarios.

Scenario 1: We all become prompt engineers, where our success—promotions, raises, performance reviews—depends on how well we craft questions, extract precise information from LLMs, and apply it effectively. In this scenario, prompt engineering becomes a core skill, shaping how we harness AI for our professional tasks.

Scenario 2: User-level prompting becomes obsolete. Instead, a machine layer sits between humans and AI, dynamically prompting AI to deliver insights without explicit instructions. Here, humans leverage LLMs natively, while AI supplements our expertise, freeing us to focus on creativity, strategy, and nuanced decision-making.

Scenario 1: I Got a Lot of Promptin’ to Do Before I Die

Imagine a workplace where every project begins not with a brainstorming session or meeting, but with a well-crafted question to an LLM. In this future, prompting is not ancillary but central to professional life, where success hinges on how effectively we interact with large language models. Prompt engineering becomes the new workplace literacy, transforming how professionals across fields tackle daily tasks. Imagine lawyers crafting precise prompts to build stronger cases or marketers refining queries for deeper audience insights. In each case, success hinges not only on their expertise but also on their skill in eliciting the right responses from AI.

There are several benefits—chief among them, the continued need for human expertise to extract meaningful output from LLMs. Prompt engineering provides control, making human input essential for quality results. While AI may “taketh away,” it also “giveth back,” creating specialized roles and growth paths, like prompt engineering certifications and training. It also brings efficiency and productivity, allowing workers to quickly access information, generate ideas, and solve problems.

However, where there are pros, there are also cons. LLMs are prone to “hallucinations”—responses that are incorrect, misleading, or entirely fabricated, despite appearing plausible. Therefore, prompt engineering requires careful wording, as minor mistakes can lead to biased or misleading outputs. This introduces inefficiencies, as humans must fact-check AI’s content. Additionally, there’s a steep learning curve, which may increase training costs and slow productivity.

Scenario 2: AI Ain’t Nuthin’ to Prompt With

In this future, AI understands our needs with minimal input. User-level prompting, as most are applying it today, is now a thing of the past, as AI now operates seamlessly, requiring only a brief from users. From there, a sophisticated machine layer manages prompting, tweaking, and filtering, delivering accurate insights at a scale far beyond human capability. AI becomes a silent collaborator, allowing us to focus on creative, strategic, and human-centric work without constant oversight.

Imagine lawyers giving a one-sentence brief and receiving a tailored analysis, or marketers setting campaign goals and letting AI automate content creation, testing, and deployment. This shift reduces the need for precise querying, enabling professionals to focus more on strategic decisions, as AI anticipates their needs and operates as a seamless partner.

The benefits here are notable—simplicity, improved adoption, and increased productivity, as AI autonomously handles adjustments needed to deliver solutions. By reducing repetitive tasks, this setup fosters creativity, allowing more time for tasks requiring human insight.

As in Scenario 1, this approach has risks. Key concerns include potential loss of human oversight, as quality becomes heavily dependent on the machine layer. Minimal input could lead to detachment from analytical or creative processes, as workers feel less involved in shaping outcomes. Additionally, potential job displacement could impact satisfaction and growth for roles traditionally focused on analysis.

A Hybrid Approach: The Likely Future

In my view, the future of work will likely combine elements of both scenarios. A small subset of professionals would become prompt experts, building templates into the backend of an adaptive machine layer. At Jacquard , we’re already working toward this vision with an architecture that places our software between the human input and LLMs like ChatGPT and Claude. Our team has built specialized prompts into our backend, where our software dynamically adjusts these prompts based on strict brand tone of voice and stylistic guidelines, and modifies results based on predicted performance. This setup allows for the production and personalization of lifecycle messaging optimized for impact. Marketers simply input a brief or example of the content they need, and our backend takes care of the rest—generating over 2,500 content variants and selecting the 10 most likely to perform, all within 30 seconds.

As we move forward, AI’s role in the workplace will deepen, transforming how we approach tasks and define success. Scenario 1 envisions a future where prompt engineering becomes a core skill, while Scenario 2 sees AI as a seamless partner. Both offer exciting possibilities and challenges, but the question remains: How do we ensure AI empowers rather than replaces human insight? Which scenario resonates with you, and what do you see as the future of AI in your own work?