In brief

Ask for a short meeting about the work that remains. Bring a before-and-after workflow map, then agree on the outcome you own, your decision authority, required checks, priorities, and who handles exceptions. If the workflow is unsettled, request a temporary agreement and review date. If agreed employment terms may change, ask what formal process applies where you work.

How do I ask for role clarity after automation removes part of my job?

Ask for a short meeting about the work that remains, not a prediction about whether your job is safe. Say which step changed, then ask to agree the outcome you own, the decisions you may make, the checks you must complete, what takes priority when time is tight, and who handles exceptions. Bring a before-and-after map of the workflow so the conversation starts with observable duties. If the process is still being tested, ask for a temporary working agreement and a date to review it. If pay, hours, employment status, location, or another agreed term may change, ask what formal process applies where you work; that is different from clarifying day-to-day task ownership. The International Labour Organization’s “Workers’ exposure to AI: What indicators tell us—and what they don’t” explains that exposure measures describe potential task susceptibility, not an individual’s employment outcome or an employer’s decision. The UK Health and Safety Executive’s “Management Standards: Role” says responsibilities should be understood and compatible, with a way to raise uncertainty or conflict. The immediate aim is a usable arrangement for current work, not a job-loss forecast or a complete career plan.

Sources: Workers’ exposure to AI: What indicators tell us—and what they don’t; Management Standards: Role

What can a removed task tell you—and what can’t it?

A removed step tells you that the observed workflow has changed; by itself, it does not show that the whole job is automated, that an employer has adopted a tool across the organization, that demand for the role has shifted, or that your position will disappear. Capability, exposure, use, adoption, redesign, demand, and displacement describe different things; evidence for one does not answer all the others. The International Labour Organization’s “Workers’ exposure to AI: What indicators tell us—and what they don’t” explains why exposure scores cannot settle a worker’s employment question. The April 2026 brief compares indicators with different methods and definitions, many based on static task lists. They do not establish whether automation is economically feasible, how a firm will redesign a workflow, or how demand may adjust. The brief treats exposure as technological susceptibility, not an employment forecast. It cannot reveal one manager’s intentions or calculate an individual’s chance of losing a job. The ILO’s “Generative AI and jobs: A 2025 update” assesses global occupational exposure using a refined method drawing on human and AI input across nearly 30,000 tasks, grouped into four exposure gradients. It says most jobs are expected to be transformed rather than made redundant because human input remains necessary. That is an occupational-level assessment, not an observation of a particular workplace or a guarantee about any role. It is a reason to inspect how tasks fit together. “The impact of GenAI on jobs, productivity and work organization: a review of the empirical evidence,” an ILO brief from 2026, reviews experiments, firm-level data, platform studies, and worker surveys. Its accessible record confirms the review’s scope but does not expose detailed findings; it cannot support a precise claim about how common redesign or displacement is. A 2026 working paper, “From Exposure to Adoption: Generative AI in European Workplaces,” examines adoption and worker-reported task restructuring across 35 European countries. It reports adoption ranging from under 3% to 25%; occupational exposure predicted uptake, but not passively, and a shift-share analysis found no detectable average effect of adoption on reported restructuring. This does not prove that no worker’s tasks changed: it is a single-wave early-diffusion snapshot, and restructuring is worker-reported. It shows why exposure or adoption alone cannot establish how a particular bundle of duties has been redesigned. Replace a job-loss guess with observable questions: Which step stopped? What outcome must still be delivered? What quality check remains? Who handles exceptions? Is the change temporary, or has management announced a redesign? If they remain unclear, that uncertainty is the subject to take to your manager.

Sources: Workers’ exposure to AI: What indicators tell us—and what they don’t; Generative AI and jobs: A 2025 update; The impact of GenAI on jobs, productivity and work organization: a review of the empirical evidence; From Exposure to Adoption: Generative AI in European Workplaces

Where can work reappear after automation?

Work may move into review, correction, exception handling, coordination, explanation, and accountability after an automated step. Whether those duties have moved into your role is a question about the workflow, not something you can infer from a tool’s capabilities. A task may now produce a draft quickly, yet someone may still need to check its inputs, resolve cases the system cannot handle, and decide what happens next. The question is: what work remains around the output, and who owns it? A systematic review, “Generative AI & Changing Work: Systematic Review of Practitioner-led Work Transformations through the Lens of Job Crafting,” examined 23 studies in the ACM Digital Library focused on workers’ experiences with generative AI. Its abstract reports that professionals delegated routine tasks while taking on monitoring and refinement of outputs; it also describes changed collaborations and tasks fragmented into piecework. This shows that checking and coordination appear in some documented practitioner experiences. It does not estimate how often this occurs across workplaces, establish a net change in jobs, or show what your employer intends. The review points to questions worth asking, but cannot answer them for your role. Consider an example: a report-drafting step disappears from a process. The input might be a set of records and instructions; an automated step produces a draft; remaining steps could include validating sources, explaining an anomaly, obtaining approval, and answering a client’s question. This sequence is illustrative, not a claim about a specific workplace. Ask who checks the output, what evidence they use, what happens when it is wrong, who can approve a correction, and who owns the next decision. If a figure is unexplained, someone may need authority to investigate the underlying record and enough time to resolve it before the report is relied on. If these steps sit with different people, make the handoff visible instead of calling it all “review.” Compare time released with time added. A faster first draft does not show that the whole process takes less effort: checking, rework, and exceptions may use some of the time saved. The International Labour Organization’s 2026 brief, “The impact of GenAI on jobs, productivity and work organization: a review of the empirical evidence,” reviews experiments, firm-level data, platform studies, and worker surveys concerning tasks, employment patterns, productivity, and workplace dynamics. The abstract does not quantify the balance for your workflow, so do not turn a change in one step into a claim about total capacity. Bring one or two examples to the manager conversation: a case that passed smoothly and, if one occurred, an exception or correction. Ask whether the time released is available for other priorities, reserved for checks and exception handling, or expected to support additional output. Ask who owns each remaining handoff and what to do when a result cannot be verified. The answers can show whether the task was removed cleanly, responsibilities shifted, or the workflow still needs an agreed owner. Do not assume the employer has chosen an arrangement before discussing it.

Sources: Generative AI & Changing Work: Systematic Review of Practitioner-led Work Transformations through the Lens of Job Crafting; The impact of GenAI on jobs, productivity and work organization: a review of the empirical evidence

What does role clarity need to settle?

A useful role-clarity agreement settles more than which task disappeared. It identifies the result you remain responsible for, the work needed to produce it, the decisions you may make, the standard for acceptable work, and what happens when the process fails or priorities collide. The UK Health and Safety Executive’s Management Standards – Role guidance offers a practical benchmark: employees should understand their responsibilities; requirements should be clear and compatible as far as possible; and a local system should let employees raise uncertainty or conflict. This is workplace-management guidance, not a universal legal rule. Its value here is the simple test it suggests: can you explain what is expected, and can you raise a mismatch? To make that test concrete, map the changed work across eight connected points. First, name the outcome: what must be delivered, for whom, and by when? Second, name the owner of each step, including review and exception handling. Third, define the decision boundary: what can you approve or correct, and what needs authorization? Fourth, state the quality check, such as which evidence must be verified before an output is used. Fifth, identify where an unusual or unsafe result goes, and who decides what to do with it. Sixth, agree which work takes priority when time cannot cover every request. Seventh, clarify the support and access needed, including system permissions, instructions, or a contact for help. Finally, set a review date if the workflow is still being worked out. This is a map of the agreement, not a requirement for a new formal document. The gaps matter because accountability can move without the authority or time needed to carry it. If you remain answerable for the final result but cannot correct a faulty output, ask who holds that decision. If a new review step has been added, ask which task can be deprioritized or what capacity is available. If two deadlines compete, ask who sets the order. If an error has no clear route, ask who should receive it and what to do while awaiting a decision. These questions describe operational mismatches; they do not presume that management intended to transfer responsibility unfairly. Before the meeting, mark each point as known, unclear, or disputed. Bring the changed workflow and the gaps that affect delivery or accountability. You might say: “I understand the drafting step has changed. I want to confirm who owns the final check, what I can correct directly, and which priority should move if review adds work.” A specific question lets a manager confirm, correct, or assign an open decision. If the answer is not available yet, ask who will resolve it and when to revisit it. The aim is an understandable working arrangement: clear ownership, authority that fits responsibility, a workable priority, and a route for uncertainty.

Sources: Management Standards: Role

Why ask before uncertainty becomes the new normal?

Ask while the change is still specific enough to describe: which step stopped, what outcome still has to be delivered, and where the handoff now sits. The reason to speak early is narrower: when several changes overlap, it can become harder to tell which duties, objectives, and expectations still apply. A short clarification can keep an unresolved assumption from silently becoming the everyday arrangement. “Role clarity” means knowing what your work requires and what others expect of you. In “Blurred Lines: Exploring the Impact of Change Complexity on Role Clarity in the Public Sector,” researchers analyzed 2014 employee census data from Australia’s federal public service, covering 99,392 employees across 89 agencies. Their matched analysis compared people experiencing higher and lower complexity of workplace changes. Complexity meant experiencing more than three different types of change over a year, such as changed priorities, a new supervisor, structural changes, or staffing shifts. Clarity was measured by whether respondents understood their duties and responsibilities. Before matching, the share who said they always had role clarity was eight percentage points higher in the lower-change group than among employees facing more complex change. After matching, the gap narrowed to five points. That reduction matters: some difference may reflect people or organizational conditions associated with change, rather than change complexity alone. The authors still found an association between greater complexity and lower reported clarity, but matching cannot prove that change caused the difference. Nor was this a study of AI or private-sector workplaces. It offers a reason to notice overlapping changes, not a prediction about your job. There is an important boundary. A single automated step may leave responsibilities clear if the outcome, workload, authority, quality checks, and exception route remain understood. Then a brief operational update may be enough. The concern is more concrete if automation arrives alongside a new manager, revised priorities, redistributed duties, altered performance measures, or new review work. Each may be manageable alone; together, they can make it less obvious what you are expected to deliver and which expectation takes precedence. The UK Health and Safety Executive’s “Management Standards - Role” offers a practical benchmark: employees should understand their responsibilities, requirements should be clear and compatible where possible, and systems should let people raise uncertainty or conflict. This is UK workplace guidance, not a universal legal rule. Its useful point is that uncertainty is a legitimate work issue to raise, especially when requirements conflict. You need not frame the discussion as opposition to automation. Name the changes affecting your work and ask which expectations now govern. For example, you might say: “The drafting step has changed, and my priorities and review responsibilities have changed too. Can we confirm which deliverable takes precedence, who handles exceptions, and what I should do if review work exceeds the time available?” That is a request to define current work, not a claim that the technology has removed your role. If the task change is bounded and answers are clear, keep the conversation brief. If overlapping changes leave ownership or priorities uncertain, ask now and make those uncertainties explicit.

Sources: Blurred Lines: Exploring the Impact of Change Complexity on Role Clarity in the Public Sector; Management Standards: Role

How should you prepare and open the conversation?

Before the meeting, make a one-page map of the workflow before automation and now. Name the changed step, its input, what the system produces, what someone still checks or approves, and what happens when the result is wrong. Separate confirmed changes from questions. For example, write “drafting is now generated in the shared tool” only if you observed that; put “who approves exceptions?” under open questions. This keeps a demonstration from being mistaken for a settled team process. Add work that is easy to miss when a visible step disappears: review, correction, handoffs, exception handling, and questions from people who rely on the output. Use recent examples to describe how often these tasks arose, if records allow. Do not turn a short sample into a universal percentage or forecast. If you lack time records, say so. Then identify the outcome you remain responsible for, any quality or service standard, decisions you can make without approval, and decisions needing someone else’s authorization. The Health and Safety Executive’s “Management Standards - Role” says requirements should be clear and compatible, with a way to raise role uncertainties or conflicts. This is a practical benchmark, not a rule that determines rights in every country. Bring capacity into the same conversation. List current priorities, new duties, and needed access, training, or time. If two priorities cannot both fit, ask which takes precedence. If health, care, location, salary, or training constraints affect what is feasible, state the work constraint and support you need; you do not need a speculative career plan to clarify today’s assignment. You could open with: “The workflow changed when [specific step] moved to [new process]. I still need to deliver [outcome], and I’m seeing [specific review or exception work]. Could we agree who owns that work, what I can decide, what should be escalated, and what takes priority if the time does not fit?” Ask: “What standard should I check against?”, “Who owns an exception?”, “Can I approve this, or should it go elsewhere?”, and “When should we review how this is working?” Request a short meeting with people who can clarify priorities. If your manager does not yet know, avoid treating uncertainty as refusal or filling the gap with assumptions. Ask who will decide, what process and priorities apply meanwhile, and when to return to the question. Acas’s “What to consult on” guidance for Great Britain says employers should consult when an issue or change is identified and before a final decision; it describes smaller discussions about new equipment or ways of working as good practice, while noting that some situations have legal consultation requirements. This is jurisdiction-specific guidance, not a promise that every worker elsewhere has the same process. A specific request gives both sides something observable to resolve, rather than turning the meeting into “Am I still needed?”

Sources: When to hold a consultation; Management Standards: Role

What should a temporary agreement and follow-up contain?

If the workflow is still unsettled, ask for an interim agreement rather than relying on a general assurance that “we’ll see how it goes.” The agreement should make clear who owns each changed responsibility, which decisions you may make, what must be checked, where exceptions go, what takes priority when time is limited, what support or access is available, and when the arrangement will be reviewed. It describes how to work during uncertainty; it does not settle whether employment terms have changed. Keep the record short enough to use. Separate what has been decided from what remains open. For example, the manager may confirm that you still own delivery of a report, while the team has not yet decided who approves unusual results. Mark the approval question as open, name who will resolve it, and agree how to handle an unusual result in the meantime. This keeps an unresolved exception from silently becoming your responsibility without decision authority. A useful follow-up tracks what actually happens: recurring exceptions, corrections or rework, decisions delayed while waiting for approval, and whether the checking work fits within available time. These observations show whether the workflow is workable, not whether a job is at risk. Agree on a review date or a practical trigger, such as repeated errors, workload regularly exceeding the time allocated, or a proposal to change formal terms. Choose a review point that allows real examples to emerge, while raising urgent quality or workload problems sooner. The Health and Safety Executive’s “Management Standards - Role” says employees should understand their responsibilities, requirements should be clear and compatible where possible, and systems should let workers raise uncertainty or conflict. This is a role-clarity benchmark, not a universal legal rule. Acas’s “What to consult on” guidance says employers in Great Britain can consult on smaller issues such as new equipment or a new performance-management approach as good practice, and recommends consultation when an issue is identified and before a final decision. This does not establish a universal duty for every workflow change. After the meeting, send a neutral recap through an appropriate work channel: “My understanding is that I own X, can decide Y, and should send Z exceptions to [role or team]. A and B are still open; [owner] will confirm by [date]. We will review the workload and error examples on [date]. Please correct anything I have misunderstood.” Follow workplace documentation and confidentiality policies. The recap records shared understanding; it does not change contractual rights. At review, close the loop if ownership, authority, capacity, and actual demand for the work fit. If one does not, raise that specific mismatch and ask for a redesign discussion or the appropriate formal process.

Sources: Management Standards: Role; When to hold a consultation

When is this more than a workflow clarification?

A removed task calls for a formal-process question when the change reaches your employment terms or structure, rather than only changing how an agreed outcome is produced. If a manager says your pay, hours, employment status, location, reporting arrangement, or another agreed term may change, ask: “Is this a proposed change to my employment terms, or a change to how my current responsibilities are carried out? What process applies, and when will I receive the details?” An announcement of restructuring is also a reason to ask whether a formal process is underway. Task removal alone does not establish that a contract has changed or that a particular legal right has been triggered. An operational conversation and a terms-change process answer different questions. In the first, you and your manager can clarify who owns an outcome, how exceptions are handled, or which work takes priority. In the second, you may need information about a proposal, consultation or representation, and time to understand its effect. Ask what the employer is actually proposing instead of inferring legal significance from automation itself. For Great Britain, Acas’s “Consulting - Changes to your contract” guidance says that when an employer proposes contract changes affecting employees, it should fully consult affected employees and relevant representatives. It describes consultation as two-way discussion: explaining reasons, inviting feedback, hearing concerns, answering questions, considering alternatives, and considering revisions before a decision. Acas’s “What to consult on” guidance says employers should consult when a change or issue is identified and before a final decision as good practice; it separately notes legally required consultation in certain circumstances, including some contract changes. They do not mean every changed task requires formal consultation, or establish what law applies outside Great Britain. If unsure, ask for the proposal and practical effects in writing, including any timeline, then check your contract and applicable workplace policies or collective arrangements. Ask whether a union, employee representative, or staff forum is involved. Outside Great Britain, consult the relevant official labor authority or a qualified local adviser because procedures depend on jurisdiction and contract. This is not a conclusion about whether a proposed change is lawful, agreed, or enforceable. Once formal terms are in play, the next step is process and support, not only a better task checklist. Keep relevant communications in line with workplace policy and seek local advice before a consequential decision. If management confirms there is no proposed terms change, return to the operational discussion and clarify duties, authority, workload, and review. If it confirms a formal proposal, find out what process applies and use the appropriate representation or advice route.

Sources: Consulting - Changes to your contract; When to hold a consultation

What if clarity reveals a workload or career decision?

Once the work is clear, separate a task-mix change from a workload mismatch and from a decision about whether the role still fits. Each calls for a different response. If the role still serves a goal you care about and the revised work is feasible, staying and redesigning it may be the smallest useful move: ask which workflow skills matter, what support is available, and whether you can test a change on a bounded piece of work. If drafting is faster but you remain responsible for source checks and exceptions, a focused learning step might improve verification in that workflow. Start from the capability gap the conversation exposed; do not assume you need a new credential, occupation, or vendor-specific tool.

If review, coordination, or exception work has been added without time, authority, or a priority trade-off, treat that first as a workload and role-design issue. Ask what should move down the list, which decisions you may make, and who resolves cases outside the agreed rules. Do not silently convert saved minutes into extra output while checks remain your responsibility. Describe the mismatch and seek a revised allocation or support. This alone does not show that your job is disappearing; it shows that the current arrangement needs a decision.

A formal proposal about pay, hours, status, location, or other agreed terms changes the question. Ask what process applies where you work and get appropriate local support before making a consequential decision. If the process confirms the role or its availability may change, compare three paths: negotiate a workable redesign; seek an adjacent role using more of your experience; or consider a larger change. Weigh them against local openings and requirements, training time and cost, income needs, location, health, and family responsibilities. Postings show current requests, and official labor-market information adds context; neither establishes what your employer will do. Occupational exposure is not demand or a personal displacement forecast. The ILO brief “Workers’ exposure to AI: what indicators tell us—and what they don’t” explains that susceptibility indicators do not capture adoption constraints or labor-market outcomes. Use them to inspect tasks, not to abandon experience or rush a move.

Use a task-level tool only if the meeting leaves a broader career choice. AI Proof Work’s free task checker can organize change-pressure signals around your actual work; its paid career roadmap can compare scenarios against your preferences and constraints. Neither predicts a layoff or guarantees employment or income. If the assignment is workable, agree on a small learning or workflow experiment and review point. If capacity is the problem, return to priorities and authority. If terms or availability are changing, follow the local process and compare verifiable options. Let the decision follow what changed and what you can sustain, not an exposure label alone.

Sources: Workers’ exposure to AI: What indicators tell us—and what they don’t

What is the next step after the meeting?

After the meeting, send a neutral note stating the outcome you own, decisions or checks assigned to you, priorities, and the route for exceptions. Mark unresolved points with an owner and a date to revisit them. Invite corrections so the note reflects shared understanding. The UK Health and Safety Executive’s Management Standards – Role says organizations should clarify responsibilities and provide a way to raise uncertainties or conflicts; use it as a benchmark, not a rule for every workplace. At review, check actual work: are responsibilities clear, do you have authority for assigned decisions, and is there capacity for checks and workload? If something remains open, give a specific example and ask who will decide it and by when. If the discussion raised a proposed change to pay, hours, status, location, or another agreed term, ask what formal process applies where you work and seek local advice or representation. Acas’s Consulting on changes to your employment contract describes consultation in Great Britain; it does not establish the process elsewhere. To map task-level change pressure, the free checker offers signals, while the paid roadmap compares work scenarios against your constraints. Neither provides a validated displacement probability or guarantees employment or income. The useful result is a clear working arrangement or a named formal next step, not certainty about your job’s future.

Sources: Management Standards: Role; Consulting - Changes to your contract

Questions readers ask

Does automation removing one task mean my job is at risk?

No. A changed task shows that one step in the observed workflow changed; it does not establish employer-wide adoption, changed demand, or whether your position will be removed. Ask what outcome remains, who owns checks and exceptions, and whether management has proposed a formal change.

Sources and notes

  1. Workers’ exposure to AI: What indicators tell us—and what they don’t

    Explains that exposure indicators describe task susceptibility and do not establish adoption, labor-market outcomes, or individual displacement.

  2. The impact of GenAI on jobs, productivity and work organization: a review of the empirical evidence

    The accessible record confirms a review of experiments, firm data, platform studies, and worker surveys; it does not expose detailed findings.

  3. Generative AI and jobs: A 2025 update

    Summarizes a global occupational exposure assessment based on nearly 30,000 tasks and four exposure gradients.

  4. From Exposure to Adoption: Generative AI in European Workplaces

    Reports analysis of 2024 worker data across 35 European countries and no detectable average effect of early adoption on reported restructuring.

  5. Generative AI & Changing Work: Systematic Review of Practitioner-led Work Transformations through the Lens of Job Crafting

    Its abstract reports a review of 23 studies describing delegated routine tasks alongside monitoring, refinement, and shifted task boundaries.

  6. Management Standards: Role

    Provides UK workplace guidance on understood responsibilities, compatible requirements, and routes for raising uncertainty or conflict.

  7. Blurred Lines: Exploring the Impact of Change Complexity on Role Clarity in the Public Sector

    Reports an association between complex concurrent changes and lower role clarity in Australian public-sector employee data.

  8. When to hold a consultation

    Great Britain employer guidance recommends discussing changes before a final decision and distinguishes good practice from legal requirements.

  9. Consulting - Changes to your contract

    Describes two-way consultation for proposed employment-contract changes affecting employees in Great Britain.

Apply this to your own work

See the whole job market at once.

Explore which occupations AI may reshape, then turn the signal into a practical response.

Explore the job map