This is a fictional composite. No real person completed this intake. The client, the employer, and every detail are invented to show what a real report looks like. Published so you can judge the work before you buy it.

AI CAREER RESILIENCE AUDIT

PREPARED FOR: Dana, sample composite

REPORT DATE: 24 July 2026

INTAKE RECEIVED: 20 July 2026

1. EXECUTIVE SUMMARY

WHAT YOU TOLD US

You are 53, a Director of Communications at a regional healthcare system, nineteen years in the field and six in this role. You run a team of five, own the executive communications calendar, and told us that in the last eighteen months your organization licensed an AI writing tool and your department headcount request was declined for the second year running.

WHAT THIS REPORT FOUND

FINDING 1. Roughly half of your reported week sits in tasks where current tools can already carry most of the production load. You reported that drafting, editing, and internal messaging together consume about 55 percent of your time, and those are the three task types where capability has moved fastest.

FINDING 2. Your most defensible work is invisible to the people who set headcount. You described talking a service line director out of a public statement that would have created a regulatory problem. That episode does not appear in any report your CFO reads.

FINDING 3. Your exposure is not really about writing. It is about concentration. Almost everything you do routes through one skill category, and you told us you have not led a project outside communications in six years.

FINDING 4. You are further ahead on AI than you think and getting no credit for it. You use two tools weekly and have built an approval workflow nobody asked you to build. You have never told anyone above you that you did.

THE THREE MOVES THAT MATTER MOST

Move 1. Make the crisis judgment work visible in writing, monthly, in a format your executives already read. Move 2. Take ownership of the AI writing tool your organization already bought, before someone in IT does. Move 3. Build one deliverable outside communications in the next 90 days.

WHAT THIS REPORT DOES NOT DO

It does not predict whether your job will exist. It does not forecast hiring, pay, or timing. It maps how the tasks in your role relate to current AI capability, names what is defensible, and gives you a sequence you can test in 90 days.

2. CURRENT ROLE SNAPSHOT

ROLE AS PRACTICED

Director of Communications. Regional healthcare system, roughly 4,000 employees, as reported. Nineteen years in field, six in role. Manages five. Owns a budget you described as small and shrinking.

WHERE YOUR TIME ACTUALLY GOES, AS YOU REPORTED IT

Drafting and editing executive communications, 25 percent. Reviewing and approving team output, 20 percent. Meetings, coordination, and alignment, 20 percent. Internal messaging and intranet content, 15 percent. Crisis and sensitive issue handling, 10 percent. Vendor and agency management, 10 percent.

WHAT YOU ARE MEASURED ON

Stated: campaign delivery, executive satisfaction, and absence of media problems. Implied, from your own answer to how you are really judged: whether the CEO's material sounds right, and whether anything embarrassing reaches the public. You wrote that nobody has ever asked you for a metric.

WHAT THE TITLE HIDES

You are the organization's institutional memory for what the system has promised the public. You told us you are the only person who remembers the 2019 commitment that resurfaced in a community meeting last spring. You also described yourself as the person who tells senior leaders no. Neither of those appears in your job description, and both are where your value actually sits.

3. TASK EXPOSURE MAP

HOW WE READ EXPOSURE

Jobs are bundles of tasks. Exposure is assessed per task, not per job, because one role routinely contains highly exposed and highly protected work at the same time. This is the standard task based approach in labor research, including the International Labour Organization's 2025 occupational exposure update, which found most exposed jobs are transformed rather than eliminated, because human input remains required. Source: ILO, Generative AI and Jobs, 2025 update, 20 May 2025.

THE THREE CATEGORIES

Automated candidate. Current widely available tools can complete the task end to end at acceptable quality with a competent person reviewing. Augmented candidate. Tools carry a large share of the effort, but a person must direct, judge, and be accountable. Protected for now. Tools add little because the binding constraint is not producing text or analysis.

Protected for now describes today, not tomorrow. We recheck the line, not the label.

YOUR TASK MAP

TASK. Drafting and editing executive communications. REPORTED TIME SHARE. 25 percent. CATEGORY. Augmented candidate. WHY. Production of a first draft is now cheap. What is not cheap is knowing that this CEO will not say the word leverage, and that the board chair reads the third paragraph first. That knowledge is not written down anywhere. WHAT WOULD CHANGE THIS. If your organization built a model trained on five years of your CEO's approved material and gave editing authority to a coordinator, this moves toward automated within the year.

TASK. Reviewing and approving team output. REPORTED TIME SHARE. 20 percent. CATEGORY. Augmented candidate, trending protected. WHY. Accountability attaches to a named person. You are the signature. A tool can flag inconsistency and tone drift, but it cannot absorb the consequence of a bad approval. WHAT WOULD CHANGE THIS. Nothing in current capability. It changes if your organization decides the approval is a formality, which is an organizational choice, not a technical one.

TASK. Meetings, coordination, and alignment. REPORTED TIME SHARE. 20 percent. CATEGORY. Split. Roughly half automated candidate, half protected. WHY. Status meetings, note taking, follow up capture, and scheduling are already automatable. The meetings where you read a room and adjust are not. You told us you can tell when a service line director is about to go around you. That is not a transcript problem. WHAT WOULD CHANGE THIS. Nothing near term for the second half. The first half will disappear whether or not you participate in the decision.

TASK. Internal messaging and intranet content. REPORTED TIME SHARE. 15 percent. CATEGORY. Automated candidate. WHY. Repeatable inputs, documented rules, text in and text out, low error cost, visible errors. This is the clearest automated candidate in your week. WHAT WOULD CHANGE THIS. Only if the content became legally sensitive or required verification a tool cannot perform.

TASK. Crisis and sensitive issue handling. REPORTED TIME SHARE. 10 percent. CATEGORY. Protected for now. WHY. Confidential inputs that cannot go into a tool. Judgment under time pressure. Accountability that must attach to a person. A counterparty who must believe a human is deciding. This task has all four protective signals at once. WHAT WOULD CHANGE THIS. Very little in current capability. This is your most defensible ten percent, and it is the smallest slice of your week.

TASK. Vendor and agency management. REPORTED TIME SHARE. 10 percent. CATEGORY. Augmented candidate. WHY. Scoping, brief writing, and deliverable review are partially automatable. Negotiation and the decision to fire an underperforming agency are not. WHAT WOULD CHANGE THIS. If agency spend shrinks because AI absorbs the work you outsource, this task shrinks with it rather than automating.

WHAT THE MAP ADDS UP TO

Roughly a quarter of your reported week sits in automated candidate work. Roughly half sits in augmented candidate work. Roughly a quarter is protected for now.

The pattern is not that your job disappears. It is that the protected quarter is the part nobody measures, and the automated quarter is the part that fills your calendar and justifies your headcount. If the automated quarter goes away and the protected quarter stays invisible, your role looks smaller on paper while being no less necessary in practice. That is the actual risk, and it is a visibility problem before it is a technology problem.

4. RESILIENCE PRIORITY SORT

This is a priority framework, not a test and not a score of you. It sorts six factors into three levels so you know what to work on first.

FACTOR 1. TASK CONCENTRATION. HIGH.

Roughly 40 percent of your reported week is writing or editing in some form, and that is the single fastest moving capability area.

FACTOR 2. OUTPUT VISIBILITY. LOW.

You wrote that nobody has ever asked you for a metric, and your best work this year was preventing something rather than producing something.

FACTOR 3. RELATIONSHIP CAPITAL. MODERATE.

Deep internally, thin externally. You named seven internal executives you can call directly and could not name three people outside the organization.

FACTOR 4. TOOL FLUENCY. MODERATE, TRENDING HIGH.

Two tools weekly, self built approval workflow, no formal training, no recognition.

FACTOR 5. OPTION BREADTH. LOW.

You reported no project outside communications in six years and named one credible alternative direction.

FACTOR 6. CONSTRAINT LOAD. HIGH.

You reported being the primary earner, geographically tied, and under two hours per week available.

ADDRESS FIRST

Output visibility, and option breadth. Visibility first, because it has the shortest feedback loop. You can change what your executives see about your work within one month. Option breadth takes a year. Start both, but expect the first to move first.

BUILD STEADILY

Tool fluency and relationship capital. Both compound. Neither pays inside 90 days.

MONITOR ONLY

Task concentration and constraint load. Concentration is high but it is a consequence of the other factors, not an independently fixable thing. Constraint load is a fact about your life, not a problem to solve. A plan that ignores it is a fantasy, so this report sizes everything to under two hours per week.

5. DEFENSIBLE STRENGTH INVENTORY

Defensible means hard to reproduce without your specific history, relationships, or context. It does not mean permanent.

STRENGTH 1. ORGANIZATIONAL MEMORY OF PUBLIC COMMITMENTS

EVIDENCE. You recalled a 2019 commitment when it resurfaced in a community meeting, and you were the only person in the room who did. WHY IT IS HARD TO REPRODUCE. It is not in a document. It exists because you were there and you were paying attention. HOW TO MAKE IT VISIBLE. Build a one page commitments register. Circulate it quarterly. This converts a personal memory into an organizational asset with your name on it.

STRENGTH 2. THE ABILITY TO TELL SENIOR PEOPLE NO

EVIDENCE. You talked a service line director out of a statement that would have created a regulatory problem. WHY IT IS HARD TO REPRODUCE. It requires standing, accumulated credibility, and a willingness to absorb a bad conversation. A new hire cannot do it in year one and a tool cannot do it at all. HOW TO MAKE IT VISIBLE. When you prevent something, write two sentences about it in your monthly update. Neutral language. No hero framing. The record is the point.

STRENGTH 3. EXECUTIVE VOICE CALIBRATION

EVIDENCE. You described knowing which words your CEO will not use and where the board chair starts reading. WHY IT IS HARD TO REPRODUCE. This is exactly the knowledge that would be required to train a tool to replace it, and you are the only source of it. HOW TO MAKE IT VISIBLE. This one is a genuine tradeoff and we address it in Section 9.

STRENGTH 4. CROSS FUNCTIONAL TRANSLATION

EVIDENCE. You reported spending 20 percent of your week in coordination, and described your role as making clinical people and finance people understand each other. WHY IT IS HARD TO REPRODUCE. Requires fluency in two vocabularies and trust in both directions. HOW TO MAKE IT VISIBLE. Volunteer for the next project that fails for communication reasons rather than technical ones. There is always one.

THE COMBINATION THAT IS RARE

Individually, each of these is common. What is uncommon is holding organizational memory, the standing to refuse an executive, and fluency in two functional vocabularies at once. That combination is why sensitive things route to you. It is also why nobody has written it down. Roles built on rare combinations are usually described by their most common component, which in your case is writing. That is the mislabeling this report exists to correct.

WHAT IS BEING TAKEN FOR GRANTED

Your organization currently prices your judgment at the market rate for writing. Environments that price judgment correctly include regulated industries under scrutiny, organizations in the middle of a merger, and any place that recently had a public failure. Those are where your combination is worth more than it is worth today.

6. FIVE PRIVACY SAFE AI WORKFLOWS

GROUND RULE FOR ALL FIVE

Do not put patient information, employee records, unreleased financials, legal matters, or anything covered by an agreement you signed into any AI tool. Every workflow below runs on your own reasoning, public information, or material you have anonymized yourself. You reported that approved tools exist but use is discouraged in practice. Treat that as a reason to confirm in writing before you scale any of these, not as permission.

WORKFLOW 1. THE COMMITMENTS REGISTER

WHAT IT REPLACES. Nothing. It creates something. WHY IT FITS YOU. It converts your invisible memory into a visible artifact.

HOW IT RUNS.

1. Collect your organization's public statements from the last three years. Press releases, community meeting minutes, published reports. Public sources only. 2. Paste them into an AI tool and ask it to extract every forward looking commitment, with date and source. 3. Review every extraction yourself. It will miss implied commitments, which is where you add value. 4. Format as one page. Date it. Circulate quarterly. WHAT NEVER GOES IN. Anything not already public. TIME SAVED, ESTIMATE. Three to four hours the first time, one hour per quarter after. Assumes public documents are already collected. HOW YOU KNOW IT WORKED. Someone cites it in a meeting you are not in.

WORKFLOW 2. THE PREVENTED PROBLEM LOG

WHAT IT REPLACES. The monthly update you currently write from scratch. WHY IT FITS YOU. Your highest value work is prevention, which leaves no evidence.

HOW IT RUNS.

1. Keep a plain text file. When you prevent something, write one line. No detail, no names. 2. Monthly, paste the lines into an AI tool and ask it to turn them into three neutral sentences suitable for an executive update. 3. Edit for tone. Never let the tool's phrasing go out unchanged. 4. Append to the update you already send. WHAT NEVER GOES IN. Names, patient details, personnel matters, legal matters. Describe the category of risk, never the incident. TIME SAVED, ESTIMATE. Roughly neutral on time. This is a visibility workflow, not an efficiency one. HOW YOU KNOW IT WORKED. An executive references a prevented problem back to you within one quarter.

WORKFLOW 3. FIRST DRAFT DISPLACEMENT

WHAT IT REPLACES. The 15 percent of your week in internal messaging and intranet content. WHY IT FITS YOU. It is your clearest automated candidate and you are already doing part of this informally.

HOW IT RUNS.

1. Write a style guide of your own. Ten rules, one page. What your organization's internal voice does and does not do. 2. For each recurring internal content type, write one reusable instruction that includes the style rules and the structure. 3. Generate the first draft, then edit. Never publish unedited. 4. Track your minutes for two weeks so you have a real number. WHAT NEVER GOES IN. Anything about named employees, anything unannounced. TIME SAVED, ESTIMATE. Four to six hours per month. Assumes roughly six internal pieces monthly at your reported volume. HOW YOU KNOW IT WORKED. You reclaim the time and can state the number.

WORKFLOW 4. MEETING LOAD REDUCTION

WHAT IT REPLACES. The automatable half of your 20 percent coordination time. WHY IT FITS YOU. You are in meetings you attend only to capture what happened.

HOW IT RUNS.

1. List every recurring meeting. Mark which ones you attend to decide and which to capture. 2. For capture only meetings, propose that a written summary replace your attendance. Ask for one, not all, as a trial. 3. Use approved transcription only, and only where your organization permits it. If it does not, ask the owner for written notes instead. This step is as much a negotiation as a tool question. 4. Reinvest the time in Workflow 1 or 5, not in more meetings. WHAT NEVER GOES IN. Do not record any meeting without explicit permission. In a healthcare setting, treat this as a hard rule. TIME SAVED, ESTIMATE. Two to four hours per month if one recurring meeting is converted. Named as an estimate because it depends on a person saying yes. HOW YOU KNOW IT WORKED. One recurring meeting comes off your calendar and nothing breaks.

WORKFLOW 5. THE OWNERSHIP MOVE

WHAT IT REPLACES. Nothing. This is a position, not a task. WHY IT FITS YOU. Your organization has already bought an AI writing tool. In most organizations, ownership of a tool like that lands with whoever asks first. Nobody has asked.

HOW IT RUNS.

1. Write a two page proposal. Usage guidelines for the tool, a short list of approved and prohibited uses, and a review process. 2. Base the prohibited list on your organization's existing policies, not on opinions. 3. Propose yourself as the reviewer for communications output specifically. Narrow scope. Do not ask to own the whole enterprise. 4. Send it to the person who signed the contract. WHAT NEVER GOES IN. This is a document you write yourself. Use a tool for structure, not substance. TIME SAVED, ESTIMATE. None. This costs you four to six hours and buys position instead of time. HOW YOU KNOW IT WORKED. Somebody forwards your document to somebody else.

SEQUENCING

Start with Workflow 2. It costs the least, carries no privacy risk, and addresses your Address First factor directly. Workflow 5 is the highest value and the highest effort. Do it in week five, after the log has given you evidence to reference.

7. STAY, CHANGE, EXIT INDICATORS

Observable signals, not instructions. You decide.

SIGNALS THAT STAYING IS WORKING

You are in the first conversation about a new initiative, not the third. Someone above you references your prevented problem log unprompted. Your AI tool proposal generates a meeting rather than a thank you. A budget conversation includes your function before the numbers are set. Someone outside communications asks you to lead something.

SIGNALS THAT THE ROLE NEEDS TO CHANGE FROM WITHIN

Your calendar fills with production work while decisions happen elsewhere. Your team's output is being reviewed by someone other than you. A new tool is adopted in your area and you learn about it after the fact. You are asked to document your process in unusual detail. Your headcount request is declined a third time with no discussion.

SIGNALS THAT BUILDING OPTIONS IN PARALLEL DESERVES SERIOUS ATTENTION

These are reasons to build options, not reasons to resign. Two or more of your five person team are not replaced when they leave. Your function is moved under a different executive without consultation. The word efficiency enters your organization's public language. An outside firm is engaged to assess your area. You are excluded from a conversation you would have led two years ago.

HOW TO USE THESE

Watch for three or more signals in one category over 60 days. One signal is noise. A cluster is information. Write the date next to each one you observe. Pattern beats memory.

8. YOUR 90 DAY PLAN

Sized to the under two hours per week you reported. Nothing here requires telling anyone you are doing it.

TIER 1. HIGH IMPACT, LOW EFFORT. START WEEK 1.

ACTION. Start the prevented problem log. WHY. Directly addresses output visibility, your top priority factor. TIME. Ten minutes to create, two minutes per entry. DONE LOOKS LIKE. Four entries by day 30.

ACTION. Write your one page internal style guide. WHY. Prerequisite for Workflow 3 and useful on its own. TIME. Ninety minutes. DONE LOOKS LIKE. Ten rules on one page, saved somewhere you can find it.

ACTION. Add three sentences of prevented work to your next monthly update. WHY. First test of whether visibility changes anything. TIME. Twenty minutes. DONE LOOKS LIKE. Sent.

TIER 2. HIGH IMPACT, HIGHER EFFORT. START WEEKS 3 TO 6.

ACTION. Build the commitments register. WHY. Converts your rarest strength into a durable artifact with your name. TIME. Three to four hours, split across three sessions. DONE LOOKS LIKE. One page, circulated once.

ACTION. Write and send the AI tool ownership proposal. WHY. Highest ceiling action in this plan. Position, not efficiency. TIME. Four to six hours. DONE LOOKS LIKE. Sent to the contract owner with a specific ask.

TIER 3. STEADY BACKGROUND. RUNS THROUGHOUT.

ACTION. Convert one recurring meeting to a written summary. TIME. One conversation, then zero. DONE LOOKS LIKE. The meeting is off your calendar for three cycles.

ACTION. Have one conversation per month with someone outside your organization in a similar role. Professional association, industry group, anywhere that is not your employer. WHY. Relationship capital is your thinnest compounding asset. TIME. One hour per month. DONE LOOKS LIKE. Three conversations by day 90.

THE 30, 60, 90 SEQUENCE

BY DAY 30. Log running with four entries. Style guide written. One update sent containing prevention language. BY DAY 60. Commitments register circulated. Two outside conversations held. One meeting converted or formally declined. BY DAY 90. Ownership proposal sent. Three outside conversations. You can state in one sentence what changed in how your work is seen.

IF YOU ONLY DO ONE THING

Send the AI tool ownership proposal. It is the only action in this plan that changes your position rather than your workload.

9. DECISION CHECKPOINTS AND RISKS

CHECKPOINT AT DAY 30

QUESTION. Did anyone respond to the prevention language in your update? IF YES. Visibility is a solvable problem here. Increase the frequency and keep going. IF NO. The channel is wrong, not the content. Try a verbal mention to one executive instead of a written line to all of them.

CHECKPOINT AT DAY 60

QUESTION. Did the commitments register get used by anyone other than you? IF YES. You have created an organizational dependency on your judgment. That is exactly the goal. IF NO. Ask one specific executive whether it was useful. If the honest answer is no, stop producing it and put the hours into the outside conversations instead.

CHECKPOINT AT DAY 90

QUESTION. Did the ownership proposal generate a meeting? IF YES. Your standing is intact and the stay path is live. Rerun this audit in six months against a changed task map. IF NO, AND THERE WAS NO ACKNOWLEDGMENT AT ALL. That is one of the strongest signals in Section 7. Not a reason to resign. A reason to move the outside conversations from monthly to biweekly.

RISKS TO WATCH

RISK. Training your replacement. EARLY SIGNAL. You are asked to document your voice calibration knowledge in detail, or to help configure the AI writing tool with examples. LOW COST RESPONSE. Own the tool rather than feeding it. That is the entire logic of Workflow 5. Refusing to participate gets you excluded. Leading it keeps your name on it. This is the central tradeoff in your situation and there is no clean answer.

RISK. Visibility work reads as self promotion. EARLY SIGNAL. A peer comments on your new monthly format. LOW COST RESPONSE. Neutral language, no adjectives, no hero framing. Report categories of risk, never incidents. If it reads as a log, it survives. If it reads as a pitch, it does not.

RISK. Over correcting. EARLY SIGNAL. You find yourself reading job postings at eleven at night after a bad meeting. LOW COST RESPONSE. This plan takes 90 days on purpose. You reported being the primary earner and geographically tied. Under those constraints, an abrupt move is the most expensive available option. Build the options first. Nothing in this report suggests leaving.

RISK. Doing nothing because the plan is too small. EARLY SIGNAL. Day 20 with zero log entries. LOW COST RESPONSE. The plan is small because you have two hours a week. If even that fails, do only the log. One line, when it happens. That is the minimum viable version of this entire report.

WHAT WOULD MAKE THIS REPORT WRONG

This rests on four assumptions. That your reported time shares are roughly accurate. That your organization's AI tool adoption continues rather than stalling. That your standing with executives is as durable as your examples suggest. And that current tool capability in writing tasks is stable enough for a 90 day plan. If your organization abandons the tool, most of this report's urgency evaporates and the visibility work still holds. If your standing is weaker than your examples imply, the ownership proposal fails and tells you something important. The report is designed so that the failure of each assumption produces information rather than waste.

10. SCOPE AND BOUNDARIES

WHAT THIS REPORT IS

A structured analysis of the information provided in the intake form, reviewed by a human analyst, describing how the tasks in the role relate to current publicly documented AI capability, with a prioritized set of actions to test.

WHAT IT IS NOT

Not financial, investment, tax, insurance, legal, employment law, immigration, or mental health advice. It does not predict whether any job will exist, what anyone will earn, whether anyone will be hired, or when any change will occur. It is not a personality assessment and it is not a substitute for a qualified professional in your jurisdiction.

WHAT IT IS BASED ON

Only the intake information, plus public research and public tool capability as of the report date. No employer was contacted. No personnel file was reviewed. No system was accessed. An error in the intake is inherited by the analysis.

ON AI TOOLS AND CONFIDENTIALITY

Nothing here should be read as advice to enter confidential employer information, client information, or personal data about others into any AI tool. Follow your employer's policies and any agreements you have signed.

DECISIONS ARE YOURS

Any action taken on this report is your own decision. We provide analysis and options, not instructions.

END OF SAMPLE REPORT