Hey, welcome to the briefing. What I'm going to cover with you today is this. AI is changing my job. What am I supposed to become? This is direct action mailbag 0009. And the question behind it is one I think a lot more people are caring than they are saying out loud. The listener asked, My company keeps pushing AI into more of our work, but nobody can clearly explain what our jobs are supposed to look like afterward. Some of the things I spent years getting good at are becoming faster or easier with AI. I do not want to resist something I probably need to learn, but I also do not want to wake up two years from now and realize the part of my job that made me valuable is gone. How do I figure out what I should learn, what I should protect, and where my value moves as AI changes the work? New technology changing the professional landscape is not new. Tools change, workflows change, skills gain value, other skills lose value, and entire pieces of jobs get compressed, moved, redesigned, or eventually disappear. There has always been uncertainty when that happens. AI is different in its speed and in how many kinds of work it can touch at the same time. But the underlying professional problem is still familiar. If your value is tied entirely to the way one task happens to be performed today, then every major technology change is going to feel like a threat to who you are professionally. And that is really where I want to focus today. I am not going to help you hunt for some mythical AI-proof job because I do not think that is the right objective. I want to help you understand where you fit regardless of what the next technology is, what part of the result you actually create, what changes because you are there, and which pieces of that contribution remain important even when the tools around you change. That means separating the task from the capability, the output from the judgment, and the technology from the person who still has to understand the environment, make the decision, coordinate the work, recognize when something does not fit, and carry responsibility for what happens next. Because if you can understand that, then the question changes. You are no longer spending your career trying to outrun every new piece of technology or guessing which profession somebody on the internet has declared safe this week. You are learning how to read where value is moving, where your capability fits inside that movement and what you need to develop so you can continue creating value as the work changes around you. That is a good question. And uh before I do anything else with it, I want to slow the read down because the environment around this question is moving a lot faster than most people can responsibly process. New tools show up, capabilities change, somebody posts another video showing AI doing in 30 seconds what used to take somebody three hours, and suddenly you are expected to make a five-year career judgment before lunch. An executive says the organization is going AI first. A consultant says half the workforce is about to disappear, somebody else says AI is overhyped and nothing important is really changing, your company buys licenses and your manager tells you to start using them. Nobody can quite tell you what success is supposed to look like six months from now, and you are standing in the middle of that trying to make a decision about your career. There is a part of this that a lot of career advice skips. If you spent five, ten, or twenty years becoming excellent at something and a tool suddenly performs the visible part of that work in seconds, that can distort how you read your own value. You are not only asking whether the task changed. You may be asking whether the years you invested still matter, whether the thing people relied on you for still matters, and whether you are already behind without realizing it. I am not going to answer that with embrace change or learn to prompt. Those phrases are too small for the question you are actually asking. The technology is changing quickly, but your career decision does not have to move at the same speed. You do not need to decide this week what you are supposed to become for the next ten years. You do not need to choose between worshiping AI and refusing to use it, and you do not need to become an AI expert because somebody on LinkedIn told you everybody who does not prompt well will be unemployable by Thursday. Apparently, career planning now expires weekly, and you do not need to pretend none of this matters, because you have twenty years of experience and your organization has always needed somebody who knows what you know. Both reactions are weak reads. One is panic and the other is denial, and neither one tells you what is actually happening inside your work. Recent workforce data gives us a useful picture of why people are feeling this pressure. Gallup reported in July that 47% of U.S. employees said their organizations had already integrated AI tools. 52% were using AI in their role at least occasionally, 30% were using it frequently, and 15% were using it daily. The adoption curve is moving. But organizational clarity is not keeping pace. Only 25% of employees told Gallup their organization had communicated a clear plan for how AI would be integrated into existing work, and nearly half of people already using AI said they had received no training on how to use it in their job. Now put yourself inside that combination. Your employer is introducing the technology and your workload is beginning to change. You are being told that using AI matters, but the organization has not clearly explained what parts of your role are changing, what remains yours, what standards still apply, which capabilities are going to become more valuable, or how you are supposed to develop them. That is how uncertainty gets manufactured. And that is why I want this line held firmly throughout the entire conversation. If you are implementing AI without explaining what changes, what remains human-owned, what skills people need next, and how employees are expected to develop them, you are not managing transformation. You are manufacturing uncertainty. That is not an argument against AI. It is an argument for leadership. Because right now there are employees trying to reverse and they are trying to build a map of their professional future from a software license, a handful of leadership talking points, and whatever demonstration they watched last night. The rollout had a roadmap, the people apparently did not. That is not enough information to make a serious career decision, and the anxiety underneath that is real. Gallup found that 18% of U.S. employees believed AI or automation could eliminate their job within five years, and inside organizations that had already implemented AI that rose to 23%. So when somebody tells me, I watched AI do part of what I spent years learning how to do, and now I am wondering what happens to me, I am not going to answer that person with some nonsense about embracing change. They are asking a legitimate question. The answer is not to calm them down by promising their job is safe because I cannot make that promise, and their employer may not be able to make that promise either. Some jobs will change substantially, some roles will become smaller, some tasks will disappear, and some positions may be eliminated, other jobs will expand, new roles will appear, and work will be redistributed. Some organizations will use AI to improve employees' capability, while others will use it primarily to reduce labor. Some leaders will manage the transition well and some are going to make a complete mess of it. The honest answer begins with uncertainty, but uncertainty does not mean helplessness. That is where I want to work with this listener. Your first responsibility is not predicting the future. Your first responsibility is improving the read. The direct answer is this do not try to determine your future value by asking whether AI can perform your current job. Break your job apart and look at the actual work. Identify what AI is changing, what it is accelerating, what it is making cheaper, what it is making easier, what still requires judgment, what depends on context, where exceptions matter, where coordination changes the result, where another human being depends on your interpretation, and where somebody still has to own the consequence. Then build from evidence. Your job title is a container, and your value is the difference you reliably create inside the work. That means we have to get much more specific than I am a project manager, I am an analyst, I am a supervisor, I am a recruiter, I am an engineer, I am in finance, or I work in operations. Those labels are too big. If AI changes three activities inside the role, the title does not tell us whether your value disappeared. It tells us almost nothing, and that is one of the reasons some of the current discussion is so sloppy. People keep talking about occupations as if an occupation is one indivisible activity. It is not. That is also why the distinction between task automation and job displacement matters so much. SHRM's 2026 research estimated that about 20% of U.S. wage and salary employment is already at least half automated at the task level. But after SHRM accounted for non-technical barriers to actual displacement, about 5.1% of U.S. wage and salary employment currently fell into its high displacement risk category. That does not mean nobody is going to lose a job. It means the sentence AI can perform part of this work is not identical to this job disappears. Those are different claims. And if you are the person trying to decide whether your career is falling apart, you need those claims separated. If AI can complete five tasks inside your role, the first conclusion is not that your career is over. The first conclusion is that five tasks changed. Now inspect what surrounds them. What was the purpose of those tasks? What happened before them? What happened after them? What decisions depended on them? Who checked the result and who interpreted it? What happened when the normal answer did not fit? And who coordinated the next move? Who talked to the customer, the employee, the patient, the supplier, the executive, or the regulator when the situation required explanation? Who determined that the information was incomplete? Who knew which variable mattered? Who stopped the process when something looked wrong? And who accepted responsibility when the decision created a consequence? That is where the job starts become invisible. I want you to slow down here because this is exactly where people start turning the conversation into some generic strengths exercise. They get worried about AI, somebody tells them to identify their transferable skills, and now they are staring at a sheet of paper writing things like communication, teamwork, leadership, creativity, strategic thinking, and problem solving. That is not useless, but it is not enough, and it will not answer your questions on value. Those things are useful for filling out a resume. They are not the totality of your capability. There is a reason this hits me the way it does. In US special operations, the first soft truth is humans are more important than hardware. US SOCOM's explanation is essentially that the right people, properly trained and working together make the critical difference, while even the best equipment cannot compensate for not having the right human capability behind it. I spent a large part of my professional life inside an environment built around that principle, and I think it translates directly into what we are talking about here. AI is hardware, very sophisticated hardware, absolutely, and hardware that can already outperform us in specific kinds of work. But do not make the mistake of measuring the totality of your human capability against the one task the hardware can now perform faster than you. You are not just the report you produce, the spreadsheet you build, the code you write, the schedule you create, or the document you review, those are outputs. Your capability includes what you notice, what you understand, what you learn, what you question, what you connect, how you adapt when the situation changes, how you work with other people, what judgment you develop through experience, and what responsibility you can carry when the answer is not sitting cleanly in front of you. And there is something larger underneath that. We can continue building a world where technology performs more of the work humans once performed, we probably will, but the objective cannot simply become removing human capability from everything until nobody is left to determine what any of it is for. Organizations exist because human beings have objectives, customers have needs, communities need services, teams need direction, problems need judgment, and somebody ultimately has to decide what outcome is worth pursuing in the first place. Technology can radically change how we accomplish those objectives. It does not automatically define why the objective matters. That is why I do not want you defining yourself by the task. A task is easy to see and easy to measure, which makes it very tempting to confuse the task with your value. But a task is only one expression of your capability under the current operating condition. Never let the easiest part of you to measure become the definition of everything you are capable of contributing. There is also a harder part of this conversation that I think people sometimes avoid. Your employer did not bring you into the organization because the organization had a moral obligation to preserve a particular collection of tasks for you forever. They had work that needed to be done, a problem that needed to be solved, a customer that needed to be served, or an objective that needed to move. You brought capability to that requirement, and in exchange the organization compensated you for the value you helped create. That does not mean employees are disposable, and it does not excuse weak leadership. If an organization changes the work, leadership still has responsibilities around clarity, expectations, development, resources, support, and honest communication. But uh there is an important distinction here. The organization can legitimately decide that a task should be performed differently when a better tool, process, or technology becomes available. Your professional identity cannot depend on leadership protecting the old method simply because that method is where you became comfortable demonstrating your value. And that is exactly why I do not want you defining yourself by the task. Tasks are seductive because they are observable. They are easy to count, easy to put on a job description, easy to measure, and easy to point at and say, that is what I do. But what you do and what you are capable of contributing are not the same thing. A task is one expression of capability under one operating condition. Capability shows up across the totality of the work. It shows up in what you notice, what you understand, what you can learn, what you can diagnose, what you can improve, how you respond when the standard path stops working, how you connect information, how you coordinate people, how reliably you produce an outcome, and what level of consequence you can responsibly own. Some of that is harder to put into one clean metric, harder to measure does not mean less valuable. So do not make the mistake of saying, oh, AI can perform this task, therefore AI has replaced my value. It may have replaced the task, it may have reduced the value of that task, it may even be changing the economic reason your current role exists. You have to be willing to see that honestly. But then ask the deeper question, what capability was underneath the task, and where can that capability create value now? Those are labels. I want evidence. Go back into your actual work, not your resume, and not your job description. Your work. Look at the last ten working days and find one normal piece of work that went well. Find one situation where something changed, failed, or became more complicated than expected, and then find one situation where another person depended on your judgment, information, coordination, approval, or decision. Now reconstruct those three events. What was happening? What were you trying to accomplish? What information did you have? And what did you notice? What did you know that somebody less experienced may not have known? What did you verify? What did you question? And what did you decide? What consequence did you prevent? What consequence did you accept? And what happened differently because you were there? More precisely, what changed because you were there? That distinction matters. We are not looking for a flattering description of yourself. We are looking for observable contribution. Maybe you are in logistics and a shipment looks ready, but you notice the paperwork and available inventory do not agree. AI may be able to summarize the shipment, compare documents, or flag an inconsistency, but somebody still has to determine whether that inconsistency matters, whether the freight can move, who owns the discrepancy and what risk the organization accepts if it moves anyway. That is where the contribution starts to become visible. Maybe you are in healthcare administration, an AI can draft a patient communication or summarize an intake record. Somebody still has to recognize when the normal communication is not appropriate, when a handoff is incomplete, when escalation is required, and when the information available is insufficient to support the next step. The production may be faster, but the consequence still has to be read. Maybe you are a project manager and AI can produce the status report in 30 seconds. Fine. If your professional value was typing the status report, then yes, that portion of your value is under pressure, but maybe your real contribution was knowing that three tasks marked green were actually dependent on a decision nobody had made. Maybe you were the person who saw the sequence problem before the schedule failed. Maybe you are in recruiting and AI can generate job descriptions, summarize applications, or help organize candidate information. That does not automatically answer who understands the hiring manager's actual need, who recognizes that the stated requirement does not match the work, who conducts the conversation that reveals whether the candidate can operate inside the environment, or who accepts responsibility for the final hiring recommendation. Again, production and judgment are not automatically the same thing. Maybe you are an analyst and AI can draft the report, summarize the data, and produce charts faster than you can. Good. Then your question becomes whether you understand the data deeply enough to challenge a result that looks plausible and is wrong. Recognize the missing source. Identify when correlation is being presented like causation. Explain what the information does not establish, and translate the analysis into a decision somebody can actually use. That is the level of inspection I want. AI is changing my job, is true, but it is still too broad to act on. The question becomes useful when you can say AI has substantially changed these four parts of my work. These three still require my judgment. These two areas are becoming less valuable. These other capabilities are becoming more important because the production work is faster. Now we can make a decision. There is another reason to do this from real experience instead of prediction. AI demonstrations are designed to show capability under a particular condition, and they do not automatically tell you what happens inside your organization. The tool may perform beautifully with one kind of information and poorly with another. Your organization may allow one use and prohibit another. A task may be technically automatable but not economically worth automating. A customer may require human review, a regulatory requirement may preserve human authority, or a security rule may restrict what data can enter the system. The work lives inside a system. That is why I keep coming back to the read. Do not make a five year career judgment from a five minute AI demonstration. The demonstration tells you what happened in that demonstration. Your job is to determine what that means inside the work you actually perform. Now, um let's go a little deeper into what I mean when I say your value may move. There are different kinds of work occurring inside almost every professional role, and AI does not affect all of them the same way. There is production, evaluation, judgment, coordination, and ownership. Production is the thing being created. The email, the report, the presentation, the schedule, the analysis, the code, the summary, the forecast, the first draft, the classification, or the research package. AI is already making a lot of production work faster, and because that is the visible change, people often assume it is the entire change. It is not. Evaluation asks whether the output is good enough for the actual situation. Is the answer correct? Is the source credible? Is something missing? Does the recommendation violate a known constraint? Did the model misunderstand the customer requirement? And does the summary accurately represent the underlying material? Did the code introduce a security issue? Did the analysis remove context that changes the conclusion? Did the employee communication create a risk the generator did not understand? Those are evaluation questions, and the existence of an output does not answer them. Judgment begins when there is no clean answer. Two priorities conflict, the procedure does not fit, the information is incomplete, the customer asks for something the operation technically can do but probably should not do, or the output looks good while experience tells you something is wrong. That is judgment. Coordination is getting the right people, information, sequence, and authority aligned so something can actually happen. AI can support coordination by summarizing meetings, building schedules, tracking assignments, drafting updates, identifying dependencies, and reminding people about deadlines. But an organization still has to decide who owns the work, who resolves the disagreement, who changes the priority, and who tells one department that its preferred outcome cannot happen. Then there is ownership, and what I am trying to get at is the part that matters here is that somebody still has to carry the consequence. If an AI generated analysis is wrong and the company loses a customer, nobody's going to hold a meeting with the model and put it on a performance improvement plan. The algorithm will not be attending the corrective action meeting. The organization has to determine which human role owns the decision, review, approval, or execution. That ownership may move, it may be reduced, it may become more centralized, or it may become more distributed, but it does not disappear just because the production became automated. And this is one of the strongest questions you can ask about your work. Where is ownership moving? Not just what can AI do. Ask who owns the output when AI does more of it, who reviews it, who has authority to reject it, who is expected to detect failure, who is responsible for the exception, and who explains the result. If the answer to all of those questions used to be you and your organization is now moving all of them somewhere else, then yes, your role may be under meaningful pressure. That is evidence. If AI is producing more of the work, but you are still being asked to frame the question, evaluate the output, resolve the exception, coordinate the decision, and own the consequence, then your role is changing in a different direction. That is also evidence. Microsoft's recent research on AI users gives us another useful signal here. 66% said AI allowed them to spend more time on higher value work, and 86% said they treated AI output as a starting point rather than the final answer and remained responsible for the thinking. Quality control and critical thinking also emerged as capabilities becoming more important. That does not mean those are magically protected forever. It means the current evidence says the value around AI-enabled work is not just in generating more output. It is in determining what should be generated, whether it can be trusted, what it means, and what should happen next. And uh I want to challenge another common response to this uncertainty. Do not immediately go shopping for courses. I am not anti-training, obviously, but people get anxious and start collecting certifications because a certification feels like movement. You can spend a year learning tools without becoming more valuable in your actual work. Nothing says career control like 17 certificates, and no clue what changed on Tuesday. The question is not what AI course should I take. The question is, what capability does my changing work require that I do not currently possess at the level I need? That may be AI fluency, understanding how to validate AI supported outputs, data interpretation, process design, technical depth, stakeholder management, exception handling, quality control, or decision making. It might be learning enough about a neighboring function that you can operate across the boundary, or becoming the person who understands where AI should not be used. Those are very different development paths. Your actual work should tell you which one matters. I have learned to respect this kind of transition because the tool usually changes faster than the organization changes the role around it. People start treating that gap like a personal failure when the operating picture moved before the structure caught up. So choose one workflow. One, use the tools your organization allows, establish what the workflow required before AI, and then run the work with AI support while paying attention to what actually changes. What became faster, what became easier, and what became worse? Where did you spend less time? Where did you spend more? What had to be corrected? What did the AI miss? And what did you miss because you trusted it? What information did you have to provide before the tool became useful and what decision still required you? Then ask the uncomfortable question too. What part of your expertise turned out to be less special than you thought? That can sting, especially if you spent years becoming extremely fast at producing something AI can now produce almost instantly. That does not erase the work you did, and it does not mean those years were wasted. But protecting the task because it once created your value is not a career strategy. The job is to learn what that experience gave you beyond the production. Did those years teach you what good looks like and what bad looks like? Did they teach you where the mistakes hide, how the customer behaves, how the system actually functions, where the exceptions occur, what consequences matter, and which relationships hold the work together. If yes, then the experience still has value, but you may need to stop proving your worth by performing the old task manually and start proving it through the higher order decisions that the experience now allows you to make. There is another side to this too. Do not outsource so much of your work that you destroy the capability required to evaluate the output. This matters especially when someone is still developing. If you are early in your career and AI produces the first draft, the analysis, the summary, the code, the recommendation, and the presentation every time, you may produce more work, while learning less about why the work is good or bad. That is an operating risk. The answer is not refusing the technology. The answer is deliberate use, because there are moments when using AI is the right decision, because speed and productivity matter, and there are moments when doing the work yourself is development. Those are not contradictions. Advanced AI users in Microsoft's research were more likely to intentionally stop and decide whether a particular activity should be done by the human or by AI. They were also more likely to deliberately complete some work without AI in order to preserve their own capability. That is a mature read. Use the tool where it improves the outcome, but protect the capability you still need in order to know whether the outcome is good. If you cannot evaluate the answer without asking the same system that created it whether it is right, you may have created a dependency you do not understand. That is not augmentation anymore. That is intellectual subcontracting without a quality department. And leaders need to pay attention to that, because this is not only an individual employee problem. If you automate the developmental work that used to teach junior people how the system functions, then you have to replace that learning path. Otherwise, five years from now you may have fewer people who know how to perform the basic task and fewer people who understand the deeper judgment that the task used to teach. That is not an argument for making people do inefficient work forever. It is a development problem. If the machine takes the repetitions, leadership has to decide how the human builds the judgment. U.S. Army bomb disposal lived a version of this problem for roughly two decades during the war on terror. For a long time, the operational environment itself became part of the developmental engine. Bomb technicians were not only learning procedures in controlled training environments where the variables were limited and the consequences were managed. They were repeatedly being forced to assess, adapt, and make decisions against an enemy who could take components from an electronics store, a hardware store, or common commercial products and turn them into a lethal weapon. When the other side can walk into Home Depot with bad intentions and suddenly give your entire profession homework, the learning environment gets very real very quickly. That environment built pattern recognition, technical judgment, consequence awareness, and advanced problem solving, because the problem kept changing, and the consequence of getting the read wrong was very real. You could not simply memorize yesterday's answer because somebody was already modifying the problem. That is one of the things people outside EOD sometimes miss. The technical procedure mattered, but the capability underneath it was being able to look at something you had never seen before, understand what mattered, control yourself, and make the best decision possible with the information and time available. Then the operating environment changed. Deployment slowed, experienced bomb technicians began leaving the services, and suddenly there was a question the institution could not avoid. Where does all of that judgment go when the people who developed it walk out the door? If the answer is it lives in their heads, then you never actually institutionalize the capability. You had highly capable people carrying institutional knowledge personally, and retirement just became a data loss event with a plaque and a sheet cake. That sounds funny because it is ridiculous, but organizations do this all the time. They spend 20 years developing somebody, let that person become the walking exception manual for everything nobody bothered to document, and then act surprised when the knowledge disappears with the CAC card. You cannot put institutional memory in a shadow box and call it knowledge management. I was eventually selected to work on exactly that kind of problem after the Army's Deputy Chief of Staff for Operations, plans, and training. A three-star lieutenant general and one of the Army's senior leaders responsible for Army wide operations, training, readiness, and future planning determined the problem was significant enough to require institutional action. This was not a local commander or department leader deciding a course needed an update. The concern had reached the level of the Army staff because the service had spent years developing advanced bomb disposal judgment through wartime operational experience, and leadership could see what would happen if that experience began leaving the force without a deliberate way to preserve and reproduce it. And uh before I make myself sound more impressive than I am, I was not selected because somebody suddenly discovered I was the intellectual final boss of bomb disposal. I have done enough dumb things in my life to remain comfortably disqualified from that title. EOD just has a useful way of teaching you that some dumb decisions come with an extremely aggressive performance review. What I did have was years inside the problem. I had seen what repeated operational exposure was doing to the way bomb technicians assessed unfamiliar problems, controlled pressure, recognized patterns, and made decisions when there was no convenient answer sitting in a manual. So the value I brought was not that I was some genius who could magically replace combat experience. It was that I understood enough of what the operational environment had been forcing people to learn that I could help turn those lessons into something the institution could deliberately develop instead of hoping the next generation somehow absorb them through osmosis. The issue was not simply whether we could keep teaching bomb technicians the technical procedures. We had to figure out how to preserve the developmental value that years of operational repetition have been providing and recreate as much of that judgment-building environment as possible through redesign training, scenarios, materials, processes, and deliberate repetitions. You cannot manufacture combat and you should not try. The Army has produced enough accidental realism over the years without putting that requirement into the lesson plan. But you can study what the operating environment was forcing people to learn. You can identify the decisions, ambiguity, pressure, pattern recognition, consequences, and technical problems that were building capability, and then deliberately create training that forces people to exercise those same mental muscles before somebody has to learn them for the first time when the consequence is real. That was the important part. We were not trying to recreate the war. We were trying to preserve the development the war had been providing. And I will say openly that some services institutionalize that lesson better than others. Some people will probably argue with me about that, and that is fine. I have seen the results, and some of the resistance is still wrapped in institutional ego. Military organizations, like every large organization, can occasionally develop the remarkable ability to confuse, we have always done it this way with peer-reviewed evidence. But the larger lesson is not about which service gets to win that argument. The lesson is that when an environment stops providing the repetitions that built expertise, leadership has to build another mechanism for developing that expertise. Otherwise, the organization becomes dependent on experience, it no longer knows how to reproduce, and eventually somebody discovers that the entire capability development strategy was basically hopefully an experienced person is standing nearby. That is the connection to AI. If AI removes the repetitive work that used to teach a junior employee how the system behaves, you do not solve the problem by forcing them to keep doing inefficient work forever. But you also cannot celebrate the efficiency gain, delete half the developmental repetitions, and assume judgment will simply appear later because everybody completed the new software training. The machine took the repetition. Fine. Now leadership has another responsibility. Determine what that repetition was teaching before you throw it away. If the old work was quietly building pattern recognition, technical depth, judgment, exception handling, consequence awareness, or an understanding of how the system actually behaves, then those capabilities still need a developmental path. That is the point. Technology can remove the task without removing the need for the capability the task was developing. Leadership has to be smart enough to know the difference. And before I leave this part of the story, I want to acknowledge a few Army EOD leaders who helped make that work possible. They know who they are. I still talk to some of them, and there are a few I have not spoken with in years. But they were sitting in key positions when this mattered. They were experienced enough to understand what we were trying to protect, secure enough in their own knowledge to let other people challenge the existing model, and confident enough to accept the professional risk that comes with changing something the institution already believes it understands. They gave us room to test ideas, argue through problems, rebuild training, and push against some very comfortable assumptions because they understood the cost of doing nothing. And to my Navy and Marine EOD brothers and sisters who stood with me, thank you too. You did not have to support what we were trying to do. You had your own organizations, your own requirements, and your own institutional pressures to deal with. But when it mattered, you helped create a united front around the larger capability problem. That support gave the effort credibility beyond one service and made it much harder for the conversation to be dismissed as simply an army issue. I probably never told any of you this clearly enough at the time, so I will say it now. Thank you. Your support during those years mattered more than I ever told you. You did not just help a project move forward. You helped preserve hard-earned capability for bomb technicians who would never know your names, and you helped make sure lessons paid for in real operational experience did not disappear simply because the operating environment changed. That is leadership, and I have never forgotten it. Now, ah, I want to go back to the listener finally, that one who is worried about what they are becoming. There is a temptation to turn this into an identity crisis because you used to be the person who knew how to do this, the person everyone asked, or the person who spent years building a skill that suddenly feels more common. That can hit hard. But there is a difference between losing exclusivity and losing value. If AI makes it easier for ten people to generate the type of output only you could generate before, then yes, the scarcity around that production changed. That matters, and denying it will not protect you. Now ask what becomes scarce next. Can those ten people evaluate the result, distinguish strong analysis from plausible nonsense, integrate the output into the actual operation, explain the trade-off, identify the recommendation that conflicts with another requirement, or recognize the customer signal hidden inside the data? Can they defend the decision to an executive, change course when reality does not match the model, train somebody else, design the workflow, decide when AI should be used, and identify when a result requires escalation. That is where value often moves when production becomes cheaper. Most people look only at what AI is removing. Look at what the removal makes more important. If first drafts become cheap, evaluation becomes more important. If summaries become cheap, source judgment becomes more important. If analysis becomes faster, problem framing becomes more important. And if routine communication is automated, difficult human communication may become more important. If more employees can create sophisticated outputs, quality control may become more important. If code generation becomes faster, architecture, testing, integration, security, and technical judgment may become more important. If the machine can generate 20 options instead of three, option generation may become less valuable while option selection becomes more valuable. And if execution speed increases, poor direction becomes more expensive. That is the read. Do not chase whatever skill is trending. Find the capability immediately upstream or downstream of the work AI is changing and inspect whether that capability changes consequence. The World Economic Forum has estimated that 39% of workers' core skills will change by 2030. AI and data-related capabilities are among the fastest growing areas, but employers are not saying every human capability stops mattering. Analytical thinking, leadership, collaboration, resilience, and other capabilities continue to matter because technology changes the configuration of work, not just the tools inside it. But again, do not take a broad list of future skills and circle five words. Go back to your work. If your organization introduced AI and now your job requires you to evaluate more output, build evaluation capability. If more routine customer communication is automated, And you increasingly receive only the difficult cases, develop your ability to operate inside ambiguity, conflict, and exception. If AI is handling more administrative project work, and you are now expected to resolve dependencies faster, become better at reading sequence, ownership, and trade-offs. If AI is accelerating data analysis and leaders now expect faster decisions, improve your ability to frame the right question, and identify when the data does not support the conclusion being requested. That is how you build from evidence instead of fear. And I want you to keep a record of it, not a motivational journal, a record of what changed in the work, what tool was introduced, what activity became faster, what activity became less necessary, what capability became more important, where you created measurable difference, where you failed, what you learned, and what that tells you to build next. That is useful information. If you do this over time, you stop treating your career like one giant irreversible decision. You begin treating it like a changing operating picture. And that is exactly why dynamic assessment fits this mailbag so well. This is not a static situation. You can make an accurate read in August and have part of that read become outdated in November because the tool changed, the organization changed, the policy changed, the customer changed, or your own capability changed. Dynamic assessment matters because the first read is not sacred. You assess, act, watch, and update. That is different from panic because panic takes one signal and turns it into a permanent conclusion. It is also different from complacency because complacency assumes the first read will remain true. Neither one is useful here. You need a living read. Close-up analysis supports that because AI is changing my job is too large to solve. You break the role apart into its component work and inspect production, evaluation, judgment, coordination, ownership, standards, dependencies, exceptions, and consequences. That gives you something you can actually examine. ACE becomes useful when the assumptions start getting louder than the evidence. AI is going to eliminate my job. What evidence supports that? My company will never automate this. What evidence supports that? Nobody can do what I do. Are you sure? Anyone can do what I do now? Are you sure? I need to become an AI engineer. Why? My experience is obsolete. Which part? My manager said we are AI first, so layoffs must be coming. That may be possible, but it is not the same thing as established. Ace forces you to challenge the assumption without pretending the concern is ridiculous. Then pro asks what happens if you read is wrong? What happens if you refuse to adapt and the work changes? Or if you chase every AI trend and stop building professional depth? What happens if you outsource your judgment too early, stay inside a shrinking role and do nothing, or leave a viable career because you misread task automation as total displacement? What happens to you personally, what happens to your professional capability, and what happens to the organization that depends on the work? Those consequences belong in the decision. Now, um, there is another part of this advisory response that matters just as much as the employee side. Leaders, you cannot dump this entire problem on your workforce. You cannot announce an AI initiative, purchase the tools, tell everybody innovation is now expected, and then act confused when employees start wondering whether you are quietly designing them out of the organization. And this is where the decision gets so stupid it becomes visual. Leadership buys 2,000 AI licenses, sends one email titled The Future of Work, schedules a town hall with no questions answered, and then acts shocked when half the workforce starts studying the org chart like it is a weather radar for layoffs. The visual is ridiculous. The operating failure is not. Employees are trying to interpret workforce consequences because leadership introduced a major capability change without giving them a usable picture of what the change means. If you know where the organization is going, explain it. If you do not know, explain what is known and what is not. Tell people which workflows are being examined, what the technology is expected to assist, what you currently expect it to automate, and which decisions remain human-owned. Tell them who reviews AI-supported output, what quality standard still applies, what data cannot be used, and what policies govern the tool. Tell them what capability will matter more if the technology performs as expected and how the organization intends to help them build that capability. Tell them whether the change is a pilot, an operating test, or a permanent design decision. That is the information people need to operate. And when you do not know the answer, say that. I do not know yet, and here is what we are testing. Employees can operate with uncertainty when uncertainty is honestly bounded. What destroys trust is vague certainty. We are not replacing anyone. Do you know that? AI is just here to help. Is that actually the strategic intent? Everybody needs to become AI enabled. What does that mean for the employee on Tuesday morning? AI will make us more efficient. Efficient at what? What happens to the capacity created? What happens to the role? What happens to the workload? And what happens to development? What happens to quality and accountability? You do not manage a workforce through slogans. If productivity improves and the role changes, somebody has to redesign the work. That means leaders need to think beyond adoption metrics. How many licenses were activated, how many prompts were run, and how many people completed training? Fine. What work changed? What outcome improved? What failure appeared? What skills became more important? What responsibilities moved? What decision rights changed? And what controls now need to change? What capability is the workforce losing because the old workflow no longer exists? And what capability needs to replace it? That is the part of transformation leadership actually owns. Buying the technology is procurement. Redesigning the work around it is leadership. And there's an uncomfortable leadership question around early career development that organizations need to start asking now. A lot of professional judgment was historically built through repetition. The junior analyst spent hours in the data. The new manager built the schedule, the junior developer wrote routine code, the recruiter reviewed applications, and the project coordinator built the status package. Was every one of those activities the highest value use of time? No. But some of them taught the person how the system behaved. And if AI removes the repetition, the organization cannot simply assume the judgment will appear anyway. You have to redesign development. What does the junior person need to see? What mistakes do they need to learn to recognize? What decisions should they observe? And what work should they still perform manually sometimes? What review should they participate in, and how do they learn what good looks like and why the AI generated answer is wrong when it is wrong? That is not nostalgia. That is capability development. If you fail to solve it, you may create a strange organization where everybody can produce senior-looking work, and fewer people understand the work deeply enough to know whether it is correct. Now for the employee, I want to make this practical without turning it into some fake certainty exercise. There are several different conditions you may actually be in, and those conditions require different responses. The first is that AI is primarily adding capability to your current role. Your job is not disappearing, the workflow is improving, production is faster, and you can accomplish more. In that condition, your development problem is adoption and integration. Learn the tool, understand its limits, learn how to verify output and protect your judgment while using the technology to improve the work without lowering the standard. The second condition is that AI is compressing a meaningful part of your production work and moving your role toward evaluation, judgment, coordination, or ownership. That can feel threatening because the part you were best known for may become easier. Your development problem is not protecting the old production. It is moving toward the capability that is becoming more important because that production got cheaper. If you used to spend 60% of your time producing reports and AI can reduce that substantially, what does the organization need from the capacity that becomes available? More analysis, customer interaction, decision support, quality control, complex work, people leadership, or process design. If nobody has answered that, ask. The third condition is that AI is removing enough of the core role that the position itself may need to be redesigned. That requires a different level of attention because now you are not only improving a workflow, the shape of the role itself is changing. What adjacent work uses your knowledge? What internal roles are gaining demand? What capabilities transfer, and what additional skill closes the gap? What part of your experience remains useful elsewhere? This is where your organization should be helping if it intends to retain you. The fourth condition is that the organization is actively moving toward displacement. Maybe leadership has announced a restructuring, work is being consolidated, specific positions are being removed, or the responsibilities you own are being transferred permanently to another system or team. At that point, the read changes again. Do not keep treating a displacement problem like a learning problem. Training is not a substitute for recognizing that the organization no longer intends to preserve the role. That is where you start building the next path deliberately. And notice what I am doing here. I am not telling you every AI change means displacement, and I am also not telling you that if you work hard enough, every role can be saved. Every employee needs to adapt. Well, not every employee in the same way. The point is that you need to identify which condition is actually forming around your work before you decide what adaptation even means. That is the difference between reacting to AI and assessing the situation. So what do you do next? Over the next two weeks, start with your real work and list what you actually did. Use your calendar, task history, projects, deliverables, customer interactions, and the problems that came across your desk. Do not use the job description unless you need it as a comparison. Then identify what AI currently touches, not what it theoretically could do. What is your company actually using, what are you actually using, and what changed? What got faster, what got easier, what disappeared, and what quality issue showed up. That is the operating evidence. Then find three real situations where your contribution materially changed the result. Document what happened, what you noticed, what you decided, what you verified, who depended on you, and what consequence followed. Then mark where your contribution lived. Was it production, evaluation, judgment, coordination, or ownership? Do not worry if it sits in more than one because most important work does. Then look at the tasks AI is reducing and ask what becomes more valuable because those tasks are easier. That is your first development signal. Choose one capability and test it inside real work. If source evaluation is becoming more important, build it. If exception handling is becoming more important, deliberately take on work where exceptions occur. If your role is moving toward customer judgment, get closer to the customer. If the future work requires better data interpretation, build that. If AI is making first drafts cheap and final decisions more important, improve your decision quality. Then reassess. Did the capability improve the did it actually improve the outcome or did it only make you busier? Did your manager value it? Did your team depend on it? And did the workflow change again? That is how the read stays alive. I would also have a direct conversation with your leader if the environment supports it, but do not walk in and ask, is AI going to take my job? That question is understandable, but your manager may not know, and even if they say no, you may not learn anything useful. Ask operational questions instead. Ask what parts of the workflow leadership expects AI to change over the next six to twelve months, which parts of your role they expect to become more important, which decisions will still require human ownership, and what capability would make you more useful as the work changes. Ask what work you should be learning now that you are not doing today, what quality or performance standard will matter when AI is part of the workflow, and whether this is a pilot or the beginning of permanent redesign. Those questions create information you can use. If your manager cannot answer, that is also information. But I would go one step further and inspect why they cannot answer. Do they actually understand what your job requires, or have they reduced it to the most visible tasks on the job description because those are easier to count? Do they understand where your judgment enters the work, what breaks when that judgment is missing, which relationships you are holding together, and how AI is actually changing the workflow around you? A manager does not need to be the technical expert on everything you do, but they need enough understanding of the work to represent it accurately when decisions are being made about technology, staffing, structure, and future capability. Because here is the uncomfortable part. Your manager may be one of the people explaining your value in a room you are not sitting in. If their entire understanding of your job is creates reports, schedules meetings, reviews applications, or answers customer questions, then AI removes part of that visible task and suddenly somebody decides your role has been reduced by 60% before lunch. Congratulations, your career was just reorganized by a PowerPoint written by people who have never actually watched you do the work. That does not automatically make your manager incompetent, but you need to evaluate their read. Do they ask enough questions to understand the work before they start redesigning it? Do they know where the actual value is created or are they managing from dashboards and task counts because those are clean, colorful, and never argue back? Can they explain how AI changes your role beyond saying it will make everybody more efficient? Because if that is the entire workforce strategy, the spreadsheet is now apparently head of talent development. And uh you need to be realistic about what that means for you. If your manager does not understand the capability underneath your task, they may not be your strongest advocate when larger workforce decisions are made. Not because they are necessarily against you. They may simply be trying to defend something they do not fully understand, which is a difficult position to win from. You cannot advocate accurately for work you have mentally reduced to three bullets and a headcount number. That means part of your responsibility is making your contribution visible without turning yourself into a walk-in campaign commercial. Show the outcomes you influence, the decisions you own, the failures you prevent, the exceptions you resolve, the relationships you coordinate, and the work that depends on your judgment. If leadership only understands your task, then when the task changes, they may believe your value disappeared with it. And that is how somebody who understands twelve interconnected parts of the operation gets reduced to the person who made the weekly report right before software starts making the weekly report. So judge the competence of the leadership around you too. Are they curious enough to understand the work? Are they secure enough to admit what they do not know? Do they bring the people closest to the work into the redesign? Or do they lock six leaders in a conference room, stare at a process map, and emerge three hours later having successfully automated something nobody was actually doing that way? That sounds ridiculous because organizations do exactly that. If after all of that, your manager still cannot explain where the role is going, what capability matters next, or how the work is changing, account for that uncertainty. Do not immediately interpret it as proof that your job is doomed, but do not pretend it means nothing either. An organization introducing significant technology without leaders who understand the work they are redesigning is creating a risk you need to see clearly. That may mean you build more portability into your career. Portability means your value is not trapped inside one company's exact process or dependent on one manager understanding it correctly. Can you explain the capability you create in language? Another organization understands, show outcomes, demonstrate judgment, transfer your experience into a neighboring function, and operate with the tools becoming normal in your field. That is not panic job searching. That is professional risk control. And I want to deal with one more phrase that gets thrown around in these conversations. Become irreplaceable. No. That is not the objective because nobody is irreplaceable. An organizations survive people leaving every day. And if one person actually is irreplaceable, congratulations. You probably found a documentation and succession problem wearing a name badge. The objective is to become capable of creating value in more than one configuration of the work. That is very different. If your only value exists because you are the only person who knows one manual process that technology can eliminate, you are exposed. If your value comes from understanding the customer, the system, the trade-offs, the standards, the consequences, and the decisions surrounding that process, you have more room to move. That is what I want for this listener. Not reassurance and not a promise that nothing changes. I want you to have options and a stronger ability to move with the change without letting every new tool redefine your professional identity. Your task list may change quickly, but your professional value should be read from something deeper than the task list. And leaders need to hear one last part of that too. If your employees are asking, what am I supposed to become? Do not automatically dismiss that as resistance. Sometimes resistance is resistance, and sometimes people simply do not want to learn a new tool. But sometimes the employee is asking a question leadership has failed to answer. What is my role becoming? What will you still need me to own? What capability do you expect from me? What happens to the capacity the technology creates? And how will I know whether I am succeeding? Those are reasonable questions. If leadership cannot answer them, do not call the workforce resistant because they are uncertain. You created an uncertain operating picture, so fix the picture. And do not hide behind the phrase, we are still figuring it out. You may still be figuring it out and that is fine, but then define what you are testing, what you know, what you do not know, who owns the decision, when the next decision will be made, and what evidence you are collecting. People can operate inside uncertainty when the boundaries are visible. What they struggle with is uncertainty disguised as strategy. If you do not know the destination yet, at least stop pretending the fog is a map. So uh to the listener who submitted this, I want to bring it all the way back to you. You are not late because you do not know exactly what your job becomes yet, and most organizations do not know exactly what their work becomes yet either. The technology is changing too quickly for the answer to remain static. Your responsibility is not to predict perfectly. Your responsibility is to read accurately enough to make the next good decision, and that starts by slowing the read and breaking the job apart. Look at your real work, identify what changed, identify where AI is reducing value, and identify where that same change is increasing the value of something adjacent. Find the moments Where your judgment changed consequence and where people still depend on your context, evaluation, coordination, or ownership. Then build the capability the evidence supports. Do not defend yesterday's task because it used to make you valuable, and do not abandon your profession because one part of it became easier. Do not let somebody else's prediction become your career plan, and do not surrender your thinking just because the tool can produce something faster than you can. Use it, test it, learn it, challenge it, understand where it fits, understand where it fails, and understand what it changes. Then update your read when reality changes again. That is dynamic assessment and that is CSA doing what it is supposed to do. It gives you a disciplined way to understand a changing situation before fear, excitement, or somebody else's certainty drives the decision. Close-up analysis helps you break the work apart, ACE helps you challenge what you are assuming, and Pro helps you see what a bad decision could cost. If the evidence eventually shows that your current role is no longer viable, PACE can help you stop treating one path as the only path, but do not start there. Start with the read, because you may discover that AI is not eliminating your value. It may be exposing where your value actually was. And if what you discover is uncomfortable, good. Better to know that the work you protected for years is becoming cheaper than to spend three more years protecting it because it used to matter. Better to know that your expertise is becoming more important because the operation needs somebody who can evaluate automated output than to waste time competing with the machine on production speed. If your organization is moving responsibilities away from you, know it. If you have been using experience as a reason not to learn the new tools, know it. If you have been using AI as a reason not to deepen your own capability, know that too. You do not need a flattering answer. You need the best read possible with the information and time available. And then you need to keep updating that read as the operating condition changes. That is how you move without panic, without denial, and without letting somebody else's slogan become your career strategy. The work may change faster than you would like, and some of what made you valuable yesterday may become cheaper tomorrow. That does not automatically mean you have no future in the work. It means you have to become more precise about where your value is moving and disciplined enough to follow the evidence. Instead of protecting an identity, the work has already changed. Try this on for a second. Imagine I am five ounces of silver. I use an analogy for this all the time, and stick with me for this one because these days I am rapidly becoming about 200 pounds of former action guy in questionable nutritional decisions. So the precious metal comparison is doing some heavy lifting already. In one environment, that silver may be extremely valuable. Somewhere else there may be so much silver available that nobody is particularly impressed that I showed up carrying five more ounces. That does not mean the silver suddenly became defective. It means the operating environment changed the demand for what I brought. The market does not owe my five ounces of silver an emotional support group because gold is having a better quarter. My job is not to spend the next six months screaming at the market that silver deserves more appreciation. And I do not need to spray paint myself gold, because gold happens to be trending this quarter. That is not development, that is professional arts and crafts. My job is to understand what silver is actually useful for, where those properties create value, where that capability is scarce, and what new applications become possible as the environment changes. And yes, I can develop, I can refine what I already have, shape it differently, combine it with other capabilities, and learn things I could not do before. Maybe I add gold, maybe I add platinum, maybe one day somebody figures out how to turn this 200 pound former action guy into a diamond. I would not structure a retirement plan around that last one, but you get the point. The objective is not to abandon everything I already know every time the market discovers a new shiny object. It is to understand the underlying capability well enough to ask a much better question. Where does what I already bring create value now and what do I need to add so it continues creating value tomorrow? That is the point I want you to carry out of this entire discussion. Do not define your professional worth by whatever task happens to be scarce today. Scarcity changes, technology changes, organizations change. Your responsibility is to understand the properties of what you bring, recognize where those capabilities matter, develop where the evidence says you need to develop, and put yourself where that combination can actually be used. You are not five ounces of silver forever. You are five ounces of silver today with the ability to refine it, combine it, shape it, and add capability. Just do not waste half your career pretending you are gold because somebody else decided silver was boring this week. When you are ready to go deeper with dynamic assessment, CSA, and the direct action tools that support discipline situation reading, go to www.direct action dash system.io slash course dash directory. Open the course directory, find the course connected to the tool, and start there. That is where the deeper application belongs. Thanks for listening to the briefing.