AI adoption among administrative professionals jumped from 26% to 53% in one year, according to ASAP's 2025 State of the Profession Report (3,700+ respondents)Source ↗
Executive assistants are the administrative professionals most likely to use AI (65%); among AI users, 73% use ChatGPT and Microsoft Copilot usage rose from 13% to 43%Source ↗
ASAP's 2026 report found 76.9% of administrative professionals now use AI tools daily — nearly a threefold increase over two yearsSource ↗
88% of managers say their teams already use AI, but only 35% of workers feel very confident using AI tools effectively, per Robert Half researchSource ↗
communicationCopilotChatGPTClaude

Triaging the inbox and drafting replies in your executive's voice

The executive's inbox is a daily firehose, and the real skill is spotting the handful of messages that need the principal now while clearing the routine dozens in a voice that sounds like them. AI can sort a batch of threads by urgency and produce first-draft replies that match the executive's tone. The judgment — what's urgent, what's sensitive, and what your executive would actually agree to — stays with you.

Prompt
You are my executive's email assistant. I am the executive assistant to {{executive_role}}. Help me triage a batch of incoming messages and draft routine replies in my executive's voice.

My executive's voice — match this greeting, tone, and sign-off: {{voice_sample}}
The messages, each with sender and full text: {{email_batch}}
This week's priorities and anything off-limits: {{priorities}}

Do two things:
1. Triage: return a table with columns Sender | One-line summary | Urgency (Now / This week / FYI) | Suggested action (Draft reply / Exec must decide / Delegate / Archive).
2. Draft: for every message marked "Draft reply," write a reply in my executive's voice, matching its warmth and brevity.

Rules:
- Do not commit my executive to any meeting, decision, price, or deadline that isn't in the priorities I gave you. Where a reply needs their call, insert [EXEC TO CONFIRM] and leave the specific blank.
- Never invent facts, names, numbers, or context you were not given.
- Flag anything that looks confidential, legal, financial, or personnel-related for me to handle directly — do not auto-draft a substantive reply to it.

Fill in your details and the prompt updates live — then copy.

planningClaudeChatGPTCopilot

Building tight meeting agendas from scattered inputs

Before a leadership or board meeting the EA collects agenda items from a dozen people across email, chat, and hallway asks, then has to shape them into a timed, prioritized agenda that actually fits the slot. AI turns that messy pile into a structured draft with time allocations, owners, and pre-reads flagged. You still confirm priorities and owners with the principal before it goes out.

Prompt
You are a meeting-planning assistant helping an executive assistant build an agenda. Turn the raw inputs below into a tight, timed agenda.

Meeting purpose, length, and attendees: {{meeting_context}}
Proposed items (each with who requested it and any time estimate): {{raw_items}}
Items that must be covered no matter what: {{must_cover}}

Produce:
1. A timed agenda: Topic | Owner | Minutes | Desired outcome (decision / discussion / FYI) | Pre-read needed?
2. A "parking lot" list of items that don't fit, ordered by priority.
3. A short "missing info to chase" list — owners not named, time estimates absent, or pre-reads not yet supplied.

Rules:
- Use only the items I provided. Do not invent agenda topics, owners, outcomes, or decisions.
- The total of the Minutes column must not exceed the meeting length; if it does, move the lowest-priority items to the parking lot and say so.
- If an item has no owner or unclear purpose, list it under "missing info to chase" rather than guessing.
- Keep every desired outcome concrete enough that we'll know whether we hit it.

Fill in your details and the prompt updates live — then copy.

automationCopilotClaudeChatGPT

Turning meeting notes into minutes and tracked action items

Every recurring staff or board meeting ends with the EA owing clean minutes and a follow-up list, usually the same day. Typing them from rough notes or a transcript is repetitive, deadline-driven work. AI converts your notes into structured minutes and pulls out every action item with owner and due date — the same format each time. The discipline is confirming that decisions, owners, and attributions are captured exactly right.

Prompt
You are a minutes assistant for an executive assistant. Convert my raw meeting notes into clean minutes and an action-item tracker. Use only what is in the notes.

Meeting title, date, and attendees: {{meeting_meta}}
My raw notes or transcript: {{raw_notes}}

Produce:
1. Minutes: attendees present, then, for each agenda item, a 2-3 sentence discussion summary and any decision reached (label it "Decision:").
2. An action-item table: Action | Owner | Due date | Status (default "Open").
3. A "To confirm" list for anything ambiguous.

Rules:
- Do not invent decisions, action items, attendees, numbers, or attributions. If who owns an action or what was decided is unclear, put it under "To confirm" instead of guessing.
- Keep names and who-said-what exactly as in my notes; do not reassign a statement or decision to a different person.
- Do not add recommendations or opinions of your own — summarize only what happened.
- Preserve any figures or dates verbatim; do not round or "correct" them.

Fill in your details and the prompt updates live — then copy.

analysisClaudeChatGPTCopilot

Condensing long documents into a one-page executive briefing

The executive is handed a 40-page vendor contract, a board deck, or an industry report and asks for "the short version" before a meeting they're about to walk into. Reading and distilling it is the EA's value-add. AI produces a fast first-pass summary with the key numbers, decisions needed, and risks. Every figure and claim still has to be checked against the source before it reaches the principal.

Prompt
You are a briefing assistant for an executive assistant. Summarize the document below into a one-page brief for a busy executive. Base everything strictly on the document.

What the executive cares about and their role: {{exec_focus}}
The decision or meeting this brief is for: {{purpose}}
Document text: {{document_text}}

Produce:
1. TL;DR — 3 bullets.
2. Key facts and numbers, each with the page or section it came from.
3. Decisions or actions required of the executive.
4. Risks and open questions.
5. Three sharp questions the executive should ask.

Rules:
- Use only what is in the document. Do not add outside facts, context, or benchmarks.
- For every figure or claim, cite the page or section. If something the executive would expect isn't in the document, write "not stated in document" rather than filling it in.
- Mark any number that drives a decision with [VERIFY] so I check it against the source.
- Be neutral: report what the document says, flag what's missing, and don't give me your own recommendation as if it were fact.

Fill in your details and the prompt updates live — then copy.

planningChatGPTCopilotClaude

Assembling a travel itinerary and pre-trip briefing

Booking and briefing executive travel means stitching flights, hotels, ground transport, meetings, and time zones into one clean itinerary the principal can follow without thinking. AI is fast at formatting a day-by-day itinerary and drafting a pre-trip brief from details you supply. But it must never be the source of a flight time, confirmation number, address, or visa rule — those come only from the actual bookings, because AI confidently invents them.

Prompt
You are a travel assistant for an executive assistant. Format a clean itinerary and pre-trip brief from the confirmed details I paste. Do not add any travel fact I did not give you.

Confirmed bookings (flights, hotel, cars, confirmation numbers, times): {{booking_details}}
Meetings and agenda during the trip: {{meetings}}
Executive's preferences and constraints: {{preferences}}

Produce:
1. A day-by-day itinerary with each item's local time, the time zone, and travel/buffer time between items.
2. A one-page pre-trip brief: who they're meeting and why, dress code, key addresses, and logistics notes.
3. A "Verify before travel" checklist and a "Missing — get from booking" list.

Rules (critical):
- Use ONLY the booking details I paste. Never generate, guess, or "helpfully fill in" a flight number, departure/arrival time, gate, terminal, address, confirmation code, price, or entry/visa/passport requirement.
- If any of those is missing, put it under "Missing — get from booking" — do not invent it.
- Convert times carefully and label every time with its zone; if you're unsure of a conversion, flag it rather than stating it.
- Do not recommend hotels, flights, or restaurants I didn't provide.

Fill in your details and the prompt updates live — then copy.

writingChatGPTClaudeCopilot

Drafting a run-of-show and coordination comms for events

Offsites, board dinners, and all-hands meetings mean the EA juggles vendors, headcounts, timing, and a stack of confirmation emails — and one dropped detail is visible to everyone in the room. AI drafts the run-of-show, the vendor and attendee messages, and the master prep checklist from your plan. You keep ownership of the real bookings, contracts, and headcounts.

Prompt
You are an events coordination assistant for an executive assistant. Turn my plan into a run-of-show, a prep checklist, and draft coordination emails. Use only the details I provide.

Event details (type, date, headcount, venue, budget notes): {{event_details}}
Segments and timing: {{segments}}
Vendors and what each is providing: {{vendors}}

Produce:
1. A minute-by-minute run-of-show: Time | Segment | Owner | Notes.
2. A master prep checklist grouped by timeline (T-2 weeks, T-1 week, day-of).
3. A short draft email to each vendor confirming their piece and asking for anything still open.

Rules:
- Use only the details I gave you. Do not invent vendor names, prices, contract terms, headcounts, or venue rules.
- Mark anything unconfirmed as [CONFIRM] and do not promise a vendor anything that isn't in my notes.
- Build realistic buffers between segments; if timing is too tight, say so rather than silently overlapping items.
- Keep vendor emails professional and specific, and flag any detail I still need to supply.

Fill in your details and the prompt updates live — then copy.

Common questions from executive assistants

Is it okay for an executive assistant to use ChatGPT or Copilot at work?

Yes, with guardrails — it's already mainstream, and ASAP found executive assistants are the administrative professionals most likely to use AI (65%). The conditions are consistent: check your organization's AI policy first, use approved tools rather than personal accounts, keep confidential information out, and treat AI output as a draft you verify. Start with low-risk tasks like formatting, summarizing your own notes, and first-draft replies.

What executive or company information should I never put into an AI tool?

Board materials, unreleased financials or earnings, M&A and deal details, litigation, personnel records, and your executive's personal data (passport and Social Security numbers, home address, health, family). Unreleased financial information can create insider-trading and disclosure exposure, and much of this is covered by NDAs. Use an approved enterprise account with no-training and retention controls, and redact or describe things generically. When in doubt, keep it out.

Can I trust AI to book or confirm travel details?

No. AI hallucinates flight times, operating hours, prices, and visa or passport rules, and presents wrong answers as confidently as right ones. Let AI format an itinerary from bookings you already confirmed, but verify every time, address, confirmation number, and entry requirement against the airline, hotel, or government site before it reaches your executive. AI organizes travel information; it does not source it.

Will AI replace executive assistants?

The parts AI does well — drafting, summarizing, formatting — are the commoditized parts of the job, not the whole job. Judgment about what's urgent, discretion with confidential information, relationships, and anticipating what the principal needs can't be outsourced to a model. As Fortune reported, the role is shifting toward strategic support; assistants who use AI to clear the routine work and move up the value chain are the ones getting ahead.

Related professions