Summary of Key Points
This article discusses a new phenomenon in the workplace: AI collaborating with each other (referred to by the author as "a2a," or AI to AI) to complete tasks. For example, employees use AI to draft proposals, and managers use AI to review them. Both parties rely on AI for the most mentally demanding initial steps, while humans are only responsible for copying, pasting, adjusting numbers, or forwarding the documents. Although this increases efficiency, it also brings problems—the content may be superficial and lack genuine thought and responsibility for the outcomes. What becomes increasingly scarce is the willingness of people to take real responsibility for their words and proposals.
Detailed Analysis
1. How Proposals Can Be Submitted Without Much Real Effort
In the past, drafting a proposal required starting from scratch: researching information, building a framework, and considering relevant cases, which often led to delays. Now it’s different:
- Employees first spend half an hour selecting an AI tool: one AI creates the framework, another adds examples, and a third adjusts the tone to make it less impersonal;
- They then assemble the content and modify budget figures (e.g., changing 500,000 to 480,000) before sending it to the manager;
- The manager doesn’t have time to review it in detail and simply asks the AI to identify any flaws or budget risks;
- The employee sends the feedback back to the AI for revision, and this process is repeated until the proposal is “completed.”
No one thinks deeply throughout the process, but the workflow seems efficient, and the proposal may appear professional (using advanced terms like “global mental touchpoints”).
2. a2a Is More Like an Elevator Than True Collaboration
Although it appears that employees and managers are working together, it’s more like being on an elevator:
- The elevator moves you forward without effort, but it doesn’t decide which floor you need to go to (e.g., the AI can create a complete proposal, but it may not meet the company’s actual needs or highlight critical issues);
- Employees save the hassle of working overtime on drafts, and managers save the patience of reviewing every detail. Both parties benefit from “faster delivery” and “more standardized documents”;
- The problem is that AI only makes the content look competent—e.g., it may mention “increasing user engagement,” but it doesn’t explain how to achieve this.
3. In the Future, Competitiveness Will Lie in “Resolution” Rather Than Speed
In the past, the focus was on who could write faster or create more attractive PPTs. Now, AI is better at these tasks:
- AI can produce large quantities of seemingly complete and professional content, but whether this content is accurate (e.g., are the examples real? Are the risks significant?) remains uncertain;
- The true difference lies in “resolution”—some people can only see vague outlines, while others can spot details (e.g., they realize the AI has just piled together correct words without any substance);
- The former can merely complete the task, while the latter can be responsible for the outcome (e.g., identifying the right influencers or necessary budget adjustments).
4. Managers Are Also Becoming “Tools” of AI
a2a isn’t just about employees; managers also rely on AI:
- In the past, managers relied on experience to evaluate proposals (e.g., “Something doesn’t feel right”), but now AI helps them list 10 issues and generate detailed feedback;
- As a result, employees use AI to write and managers use AI to review—both have outsourced the most mentally demanding parts of their work, with humans acting as mere intermediaries;
- On the surface, it seems like people are communicating, but in reality, two AI systems are discussing the proposals first, with humans merely relaying the results.
5. The Absurdity of the Situation
Everyone finds this arrangement convenient: AI solves the most pressing workplace issues (saving time, eliminating unnecessary tasks, and ensuring polished communications). For example, managers use AI to write praise letters, and employees use AI to respond with thank-you emails—both are flawless, but no one expresses genuine thoughts;
- Gradually, AI will become the standard in the workplace: weekly reports, proposals, and emails will all be generated by AI, and meeting minutes will be organized by AI;
- In the end, what becomes scarce is not the ability to use AI, but the willingness of people to take responsibility for the content—e.g., to say, “This proposal from AI has issues that need to be addressed” or “That example is false and cannot be used.”
This article serves as a reminder that while AI can save us time, it cannot think for us. What truly matters in the workplace is people’s commitment to the outcomes of their work.