Delegating to AI as EAs has moved from being something we experiment with to something we genuinely need to think about properly, because our workload has not magically reduced just because new tools exist. We still manage the inbox that never quite clears, the meeting prep that stretches longer than planned, and the reporting cycle that comes back around just as you feel you are getting on top of everything. Even when you are supporting at a senior level, there is always a layer of admin sitting underneath the strategic conversations you would rather spend your time in.
When I was in the role, I could map out my week without even opening my laptop. You probably can too. There are back‑to‑back meetings to prepare for, long email threads to summarise for your Executive, slides that need adjusting at the last minute, travel options to double check, and that mental list that runs in the background because dropping one detail is simply not an option. Delegating to AI gives us a way to reduce that volume in a controlled way, which means deciding in advance that AI drafts the first version of the briefing, pulls together the raw meeting notes, or structures the report outline, while we step in to sense-check, refine, and adjust it before it goes anywhere near your Executive. This means setting clear boundaries around the task, using the tool for the repeatable groundwork, and keeping the final call and context with us.
Most of us are already using AI in some form, whether that is ChatGPT, Copilot, Gemini, or built‑in features inside the systems we open every day. So, as EAs, the conversation moves beyond whether we should use it and becomes much more specific. What exactly are we delegating to AI, and what stays with us because it relies on context, relationships, and experience that only we have?
We know how delegation works because we do it all the time in this role. Think about how you brief a supplier; you spell out the outcome, the budget, and the deadline. When something goes to a team, you add context so they are not left trying to interpret what you meant. And before anything lands in your Executive’s inbox, you have already filtered, shaped, and decided what actually matters. Delegating to AI follows the same logic. It becomes another resource you manage, and whether it saves you time or creates extra work depends on how clearly you define the task and how carefully you review what comes back.
In this article, we will explore five tasks you should already be delegating to AI in practical, everyday terms, the kind of tasks that come up in your week again and again. We will also cover three areas that should remain with us as EAs, even if the tool offers to draft something for you. The aim of this article is to free up real hours in your week so you can spend more time supporting your Executive at the level you know you are capable of.
This isn’t just another AI list; it’s a collection of prompts written by Assistants, for Assistants, to help you get the most out of ChatGPT and start making a real difference in your day-to-day.
Before we go into specific tasks, it helps to have a clear filter in mind. Not all of us delegate work to other people in our roles, and actually many EAs end up doing most things themselves, so this is less about handing work off to someone else and more about deciding what you really need to do from start to finish and what can start somewhere else.
It usually makes sense to delegate to AI when the task is repetitive and comes up again and again in your week, when all the information you need already exists in digital form, and when the structure is predictable enough that a solid first draft would genuinely save you time. If you find yourself thinking, “I’ve written this email before,” or “I’ve built this report before,” that is often a sign that delegating to AI could really help.
Keep control when the work directly affects relationships, when the message requires careful reading of tone and context, or when the output could influence direction or reputation for your Executive or your team. In those situations, you can still use AI to gather background or organize your thoughts, so you are not starting from a blank page, but the final wording and decision should sit with us as EAs.
So as you read the rest of this article, just keep this filter in mind. If you are unsure about a task, pause for a moment and ask yourself whether it is something you really need to craft personally, or whether delegating to AI for the first pass would actually give you back some time.
The five tasks below are ones most of us deal with every week, sometimes every day, and they share a few common features. They are repetitive, largely digital, and structured enough that a solid first draft or summary can save you a lot of time. I have chosen them because they are tasks where AI can pull together information, draft a structured first version, or organise the basics quickly, while we still decide what matters, adjust for context and any relationship or tonal issues, and shape the final version that reaches your Executive.
The Task
Preparing your Executive for a meeting, especially when it involves recurring stakeholders, cross-functional projects, or situations where there is a lot of context or backstory your Exec has to be aware of.
As EAs, we often end up piecing this together from email threads, calendar notes, Teams or Slack messages, previous minutes, and whatever sits in our own memory of what happened last time.
What AI should do
This is where delegating to AI is really helpful. Tools like Microsoft Copilot in Outlook and Teams, Gemini in Gmail and Docs, or ChatGPT with access to the relevant content can pull together recent email exchanges, summarize long threads, extract key decisions from meeting notes, and draft a structured briefing in minutes. If you use a tools like Superhuman or Fyxer, they can also scan past project pages and surface relevant updates.
You can prompt the tool to create a short pre-meeting brief that includes background, current status, open actions, risks, and any unresolved questions. Instead of manually searching through inboxes and folders, you are reviewing a draft that already has the main pieces in place.
What we still do
We sense-check it. We add the context that does not live in a document and can’t be understood by AI, such as the tension between two stakeholders, the fact that one team missed a deadline last quarter, or that your Executive wants to push a particular angle in this meeting, or the side-eye one colleague gave to another. We decide what actually needs to be front and centre ofthat summary and what is background information and more FYI details.
How this fits into your workflow
In practice, you might build these briefs at the end of the day for the next morning’s meetings, or first thing when you log on, so your Executive has a clean, structured summary before they step into their first meeting of the day. For recurring one-to-ones or board updates, you can even use the same prompt each time so the format stays consistent.
How this frees up time
For recurring meetings, this can easily give you back twenty to thirty minutes each time. The difference is that you are no longer spending that time searching and copying information across systems. You are reviewing, adjusting, and thinking about the meeting itself, which is a far better use of your attention.
The task
Managing your Executive’s inbox, or sometimes your own, where a large part of the day can disappear into reading, sorting, forwarding, drafting, and deciding what needs attention first.
As EAs, we are constantly filtering information, working out what is urgent, what can wait, and what does not need your Executive’s time at all.
What AI should do
This is another area where delegating to AI makes practical sense. Tools like Copilot in Outlook, or Gemini in Gmail can sort emails by theme, identify common topics, draft responses to routine messages such as meeting confirmations or standard information requests, and summarise long threads into a few key points.
You can ask the tool to draft replies in your Executive’s tone, suggest short summaries at the top of long chains, or group similar emails together so you are not making the same decision ten times in a row. Instead of starting from zero each time, you are reviewing something that is already structured.
What we still do
We decide what your Executive should actually see and what can be handled on their behalf. We adjust tone so it reflects the relationship with the sender. We catch anything that feels too blunt, too vague, or just a bit weird or too wordy We also recognise when an email should not be answered with another email and instead needs a call or a quick conversation.
How this fits into your workflow
In reality, this might look like running a morning inbox sweep where AI drafts first responses for routine emails, which you then review and send, or building short summaries at the end of the day so your Executive has a clear view of what came in and what still needs a decision. Over time, you can refine the prompts so the drafts get closer to your preferred style.
How this frees up time
Delegating to AI when it comes to inbox triage saves a ton of time but it also reduces the constant small decisions that come with inbox management. When you are not drafting every routine reply from scratch or reading every line of every thread, you preserve mental energy for the conversations and tasks that genuinely require your full attention.
The task
Capturing what happened in meetings, especially when you are juggling note-taking with managing the tech, watching the chat, and keeping an eye on time.
As EAs, we often leave a meeting with pages of rough notes that still need to be typed up, structured, and turned into something usable.
What AI should do
Most virtual meeting platforms now include built-in transcription and summary features, such as Microsoft Teams and Zoom. There are also standalone tools like Granola or Fireflies that join the meeting, transcribe the conversation, and generate structured summaries with suggested action points.
Delegating to AI here means letting the tool handle the transcription, create a structured summary with headings, and draft a clear action list based on what was discussed. Instead of writing full minutes from scratch, you start with a version that already captures the flow of the conversation.
What we still do
We clarify ownership. We sense-check whether the actions listed are actually agreed actions or just ideas that were mentioned in passing, AI isn’t that good yet so you do need to check. We spot what was implied but not explicitly stated, especially when your Executive made a subtle commitment or pushed something back. We also follow up where it genuinely matters, rather than sending out a long list that no one reads.
How this fits into your workflow
In practice, you might review and edit the AI-generated summary immediately after the meeting while it is still fresh in your mind, then circulate a concise version to attendees. For recurring meetings, you can use a consistent structure so stakeholders know exactly where to look for decisions and actions.
How this frees up time
You are no longer typing up full minutes line by line after the meeting ends. Instead, you are editing, clarifying, and tightening what is already there, which can easily save thirty minutes or more per meeting depending on length.
The task
Pulling together monthly or quarterly reports, board packs, KPI updates, or project summaries where you are working across spreadsheets, slide decks, email updates, and shared drives.
For some EAs, this means building large parts of the report yourself. For many others, it means coordinating input from your Executive’s direct reports, chasing updates, consolidating different versions, and making sure everything is ready for review before it reaches your Executive or the board.
In both cases, you are the person who sees the whole picture. You are checking that the document is complete, that sections are in the right order for the agenda, that formatting is consistent, and that nothing important has been missed before it reaches your Executive.
What AI should do
If you are hands-on with the data, tools like Copilot in Excel or Google Sheets can help clean up inconsistent data, identify duplicates, suggest formulas, and summarise trends across columns. ChatGPT or Gemini can draft a structured report outline based on your raw notes and figures, create an initial executive summary, and suggest suitable chart types based on the data you provide.
If your role is more about coordination, delegating to AI can still make a real difference. You can use it to draft clear request emails to stakeholders, summarise the updates you receive from different teams, compare this month’s input with last month’s to flag changes, and pull multiple contributions into one structured document. Instead of manually stitching together five separate updates, you review a consolidated draft and refine it.
Rather than starting from a blank slide and asking yourself how to frame the update this month, you paste in the key figures or team inputs and ask for a structured summary with headings, short explanations, and suggested visuals. You are then working from a first version that already has shape.
What we still do
We interpret what the numbers actually mean in context. We know which metrics matter to your Executive and which ones are background detail. We connect the data to current business priorities, upcoming decisions, or known concerns. We also sense-check team inputs, question anything that looks inconsistent, and remove anything that could be misleading or taken out of context.
How this fits into your workflow
In practice, this might mean using AI at the point where inputs start coming in, summarising each one as it arrives, then running a final consolidation prompt once everything is collected. If you are more hands-on with the data, you might use the tool as soon as the reporting period closes to generate a draft summary and chart suggestions, then spend your time refining the narrative and checking accuracy. Over time, you can build a repeatable prompt for each reporting cycle so the structure stays consistent and you are not rebuilding the format every month.
How this frees up time
Across a full reporting cycle, this can save hours. The time saved comes from reducing manual copying, reformatting, and first-draft writing, so your effort goes into interpretation and quality control rather than administration.
The task
Researching flights, comparing routes, checking connections, reviewing hotel options, and then turning all of that into a clear itinerary your Executive can actually read in two minutes.
As EAs, we often repeat this process with only small changes in destination or timing, and the research stage can take far longer than it should.
What AI should do
Delegating to AI here means using it to compare flight options across time windows, surface hotel options based on location and budget, summarise cancellation policies, and draft a clean itinerary layout with travel times, addresses, confirmation details, and contact numbers.
You can ask a tool like ChatGPT or Gemini to compare three route options and explain the pros and cons in plain terms, or to turn a list of bookings into a structured itinerary document. Standalone tools such as Perk or iMean can also help surface flight and hotel options quickly, summarise key details, and present comparisons in one place so you are not manually copying information from one tab to another.
What we still do
We apply preferences. We know your Executive prefers a particular airline, avoids certain connection airports, or values a later departure over a slightly cheaper fare. We factor in loyalty status, upgrade likelihood, and past experiences with specific hotels. We also think about risk, whether that is tight connection times, weather at the destination, or political and security considerations.
Real-world experience still matters. We know when a 45-minute connection looks fine on paper but is unrealistic in practice, or when a hotel is technically close to the office but a nightmare in rush hour.
How this fits into your workflow
In practice, you might start by asking AI to narrow the options down to three realistic choices, then sense-check them against known preferences and risks before sending a short recommendation to your Executive. Once bookings are confirmed, you can use AI again to draft the itinerary document, which you then review and adjust before sharing.
How this frees up time
The repetitive comparison and formatting part largely disappears. You are no longer jumping between fifteen browser tabs or manually typing out every flight detail. Instead, you focus on making the right recommendation and checking the plan makes sense in the real world.
If the first five sections were about where delegating to AI makes sense, this section is about where we need to take a second. As EAs, there are certain areas where context, trust, and experience matter too much to hand over fully, even if a tool can draft something quickly.
Performance discussions, team concerns, conflict between colleagues, or feedback that could change someone’s career path sit firmly in this category, and these are conversations your Executive will usually lead.
As EAs, we may help prepare the structure, outline key points, or sense-check wording beforehand, and AI can support that preparation, yet the actual conversation and final phrasing sit with you and your Executive.
When something has gone wrong, speed matters, yet tone matters more. A system outage, a missed deadline, or an external issue that affects clients requires careful wording and clear ownership.
AI can help structure the message by outlining what happened, what is being done, and what the next steps are. You can use it to organise the facts so you are not starting from a blank page. Yet people can usually tell when a message has been written entirely by a tool. If there has been a genuine crisis or something that has affected a group of people, the explanation and any apology need to be worked through by a person.
Trust is rebuilt through clear, human language. There is nothing worse than something going wrong and then the apology feeling generic or automated rather than coming from someone who is actually taking responsibility. So use AI for structure if you need it, then rewrite carefully, adjusting for audience, risk, and internal sensitivities before anything is sent. In moments like this, small wording choices can change how a situation is perceived.
AI can summarise data, compare scenarios, and pull together background research. It can outline options based on the information you give it.
Where the business goes next is shaped through conversation, and as EAs we are often closer to that conversation than people realise. You know how previous decisions have played out in practice. You know which projects stalled because resource was stretched, which initiatives landed well with the board, and which ones created unnecessary pushback. You have seen how work actually flows through the organisation, not just how it looks on a slide.
In those moments, your role is not to let a tool decide the direction. It is to share your thoughts and opinions based on what you have seen, what you understand about the team, and how similar situations have unfolded before.
You might say, “Last time we tried something similar, the timeline slipped because of X,” or, “The board will likely ask about Y, so we should be ready with that answer.” That input comes from experience inside the function, not from a summary generated in seconds.
As EAs, we often prepare the material and structure the discussion, yet we also contribute perspective. Strategic direction is shaped by people who understand the context and are willing to voice what they know. That part remains human.
If we consistently delegate the five practical areas earlier in this article, you can reclaim hours that would otherwise be spent drafting, formatting, and searching for information. The constant small decisions that fill your day reduce, which makes it easier to think clearly when something more complex lands on your desk.
More of your week shifts towards conversations instead of admin. Meetings feel easier to step into because the summary has already been pulled together and reviewed. There is also greater capacity to support your Executive at a higher level, without the sense that the basics are slipping.
None of this requires a complete change in role. It comes from being deliberate about what you delegate to AI and what you keep. As EAs, that choice sits with us, and when we make it consciously, the week feels more manageable and far more focused.
Share this article:
Join 30,000+ Assistants receiving our carefully crafted weekly newsletter packed with valuable tips, tricks, and insights tailored specifically for Assistants like you.
