Imagine two assistants waiting for the same meeting to start. One mentions that she no longer pastes an entire email thread into AI for a summary. She gives it only the most recent replies and gets better results. The other assistant tries it that afternoon. By Friday, she has changed how she works.

Across the room sits a third assistant who has done the job longer than either of them. She was not part of that conversation. The suggestion was exchanged casually, before the meeting began, and never reached her.

Not every lesson about AI comes from formal training or an instruction manual. Sometimes, a useful discovery is shared in an ordinary conversation between colleagues. One person tries something different, mentions what worked, and another takes that insight into their own work. The knowledge moves sideways, from one professional to another.

These exchanges may never appear in a training record or professional development plan. Yet a passing observation can introduce someone to an approach they had not considered. As AI becomes part of everyday work, these informal opportunities to learn from one another deserve our attention.

The Learning Nobody Schedules

Researchers at Columbia University interviewed ten experienced Microsoft 365 Copilot users. None named formal training as their main way of learning. Eight relied on trial and error. Six relied on trading tips with colleagues (Sahni and Chilton, 2025). The study was small but its findings illustrate how experimentation and informal conversations can contribute to learning alongside formal training.

This type of learning takes several forms. A colleague mentions what she changed in a prompt. Someone explains why a response that looked right was misleading. Another professional describes what happened when she tried an approach and what she would do differently next time.

Sometimes a colleague simply asks the question you had not thought to ask. Each exchange can offer something different. A tip introduces a method and a correction identifies something to watch for. A firsthand account provides context from experience, while a question may introduce another way to approach a problem. None of these requires a scheduled training session, although each can contribute to someone's understanding of how to use AI.

When AI Becomes the One You Ask

AI also gives us another place to take questions we might previously have brought to another person. In a 2026 MyIQ survey of more than 22,000 employees, reported by FM Magazine, 74% of regular AI users said they now direct questions to AI that they once asked colleagues. Across all workers surveyed, 59% said they ask colleagues for second opinions less often than before. The survey authors caution that these findings do not establish that AI use directly weakens workplace relationships.

In a separate BCG study of 70 senior leaders, approximately one-third identified declining mentoring, coaching, and knowledge sharing as a concern. These findings raise a question about how changing habits of seeking answers might affect opportunities for people to learn from one another. They do not establish that informal learning is disappearing.

There are practical advantages to asking AI. It is readily available. You can explore unfamiliar subjects, ask follow-up questions, and work through an idea at your own pace. Still, an AI response and an experienced colleague's perspective may offer different things.

AI may explain how to complete a task. A colleague may describe what happened when she tried it, what she would do differently, or what she learned through experience. One offers an answer while the other may offer an answer with a history attached.

Different Ways of Learning, Different Opportunities to Share

Not every professional has the same opportunities for informal exchange. Consider the third assistant in the opening scene. She was not part of a conversation that might have introduced her to another way of using AI. Her absence tells us nothing about her interest, capability, or how she is developing her own skills.

For administrative professionals who are the only person in their role, opportunities for these conversations may look different. There may be no administrative colleague nearby to exchange ideas with during the workday. That is a difference in working conditions, not professional capability. Others may prefer to learn independently.

Much of my own AI education has come through private experimentation. I have explored tools, tested ideas, made mistakes, and followed questions wherever they led me. That independent learning eventually contributed to the work and frameworks I develop through InclusAI.

I value the freedom to explore on my own. I also recognize that another professional's experience can introduce something I might never have thought to test. Independent learning and learning through exchange are not opposing approaches. Each offers different opportunities to develop understanding.

There is another consideration within this discussion: hesitation. This is how I define Quiet Technophobia®:

Not fear. It is hesitation beneath competence. The silent withdrawal of experienced professionals from AI adoption when the conditions for engagement were never built for them.

Sometimes that hesitation may appear before the first prompt, when an experienced professional has access to a tool but not the conditions that would make engaging with it feel workable. Choosing to learn privately is not Quiet Technophobia®. Neither is every reluctance to ask a colleague a question. That reluctance can have many causes. You can be accomplished in your profession and still be a beginner with a particular technology; the two can exist together.

There is also a difference between learning privately and sharing what you have learned. AI use is not always visible in finished work. That alone tells us little about how someone arrived at the result or why they chose not to discuss their process but the learning itself may be worth sharing. What worked, what needed correcting, and what someone discovered through experimentation could offer another professional a useful perspective.

Sharing is not always without consequences. In The AI Disclosure Trap, I examined research suggesting that disclosing AI use can affect how competence or trust is perceived, even when the work itself is unchanged.

A 2026 study of 19 knowledge workers offers another perspective. Researchers found that learning from colleagues helped participants develop their AI skills. They also observed a tension in the workplace practices they studied: removing signs of AI use could be treated as evidence of expertise, while those practices could also limit opportunities for sharing what people were learning (Xia and colleagues, 2026). The study did not focus on administrative professionals. Its findings should not be treated as an explanation for why someone chooses to keep their AI use private.

What interests me is the relationship between individual learning and shared learning. A professional may discover a useful approach independently. Another may learn something through a conversation. An exchange creates an opportunity for both to consider what the other has discovered. Neither person needs to have all the answers to contribute.

What Another Person Can Add

This is how I define Living Knowledge:

The accumulated intelligence that lives exclusively within a human professional, built through experience, judgment, and relationship, that grows and self-corrects in real time and cannot be fully extracted, replicated, or frozen by any documentation process or AI system.

It is what you carry from situations you have navigated, mistakes you have corrected, and decisions that taught you something no manual could fully explain. Your own experience can help you recognize problems, revise assumptions, and make better decisions. Another professional's experience may introduce circumstances you have not encountered.

Think back to the email summary. One assistant discovered that using fewer replies improved her results. Another tried the approach and found it useful. But deciding which replies matter still requires attention to the content, the purpose of the summary, and what the recipient needs to know.

A different professional might approach that judgment through years of experience managing complex correspondence. Her contribution may not be a better prompt. It may be knowing what information should never have been left out of the summary in the first place. This is where Living Knowledge becomes relevant to the exchange.

A colleague may introduce a more effective way to use AI. You may recognize a limitation in the result that her familiarity with the tool did not prepare her to see. Neither contribution makes the other less valuable. The opportunity is to bring different experiences and kinds of knowledge into the same conversation.

What We Risk Missing

Learning from other professionals can introduce us to experiences we have not had ourselves. Someone may have discovered a better way to approach a task, recognized a limitation in an AI response, or learned something through trial and error that could be useful to us. In another conversation, we may be the person offering that perspective. These exchanges do not replace independent experimentation or formal training. They provide another opportunity to build on what we know through the experiences of others.

For administrative professionals who work independently or are the only person in their role, those conversations may happen through professional networks, online communities, or relationships outside their immediate workplace. The important distinction is not whether someone learns alone or with others. It is recognizing what another person's experience might contribute to our own understanding.

The scene before the meeting is a small one but it illustrates something worth considering. As AI becomes part of everyday work, some of what we learn about using it well may never appear in a training record, standard operating procedure (SOP), or formal training material. It can move through passing remarks, corrections, firsthand accounts, and questions someone else thought to ask. These exchanges create opportunities for what one professional knows to inform how another uses a tool.

AI can help us explore questions and develop our understanding. A conversation with another professional can introduce experiences and perspectives we might not have considered on our own. Each offers something different. The next time someone shares what they have learned about AI, be curious about their discovery. Then consider what your own experience might add.

Sources

Sahni, R., and Chilton, L. B. (2025). Beyond Training: How Workers Discover Value in Enterprise AI. Columbia University, arXiv, version 2, October 6, 2025. https://arxiv.org/abs/2502.13281

Xia, Q. N., Constantinides, M., Sarkar, A., Brumby, D., and Cox, A. (2026). "If You're Very Clever, No One Knows You've Used It": The Social Dynamics of Developing Generative AI Literacy in the Workplace. CHIWORK 2026, arXiv. https://arxiv.org/abs/2602.01386

Goel, S., Martin, D., and Kaffe, C. (June 17, 2026). When Everyone Uses AI, Companies Risk Losing Critical Skills. Boston Consulting Group. https://www.bcg.com/publications/2026/when-everyone-uses-ai-companies-risk-critical-skills

Employees increasingly consult AI, not coworkers. FM Magazine, August 6, 2026, reporting a MyIQ survey. https://www.fm-magazine.com/news/2026/aug/employees-increasingly-consult-ai-not-coworkers/