Automate, or be automated. This is something I've been saying for a while.
It isn't a threat. It's an observation that the parts of this job a machine can do are going to be done by a machine sooner or later, and the assistant who has already handed those parts over is in a very different position from the one who is still defending them.
What has changed recently is that AI is finally good enough to be worth learning and applying. So the "sooner" is here – whether you like it or not.
What hasn't changed is which parts of the job matter, and almost none of them are the parts AI is good at.
Start with what doesn't transfer
Before taking a look at what can be automated with AI, let's talk about the part that gets skipped.
An assistant's real value is judgment, relationships and context. Knowing that this request from this person on this day means something different than it did last week. Knowing who to call. Knowing that your executive says yes to things on Friday afternoons that they regret on Monday.
None of that's in a large language model. It's in you, and it's built by being present for years. Talking with Fiona Young on AI and the future of the EA role was a great conversation about where the line falls, and Dawn Stallwood on automation and leadership gets at the same thing from thirty years of experience.
The goal isn't to do your entire job with AI. It's to hand over the painful, mundane, repetitive parts of your job so you can spend your time on the more interesting – and more influential parts.
"So how do I communicate my influence, Jeremy?"
I'm glad you asked. As assistants, our job responsibilities are all over the place, so it can be hard to communicate the effect we have on our executives and teams. Thankfully, over the years I've learned how to do this, and in part due to my friend Al-Husein Madhany's simple, yet powerful framework to use as I think (and talk) about my role as an EA.
This framework works well for performance reviews and job descriptions, but it's also a game-changing framework for everyday EA work.
In short, there are 3 primary areas of influence you have as an assistant:
- Your executive(s).
- Your team (your executive's direct reports, your peers, etc).
- Your company or organization.
Anything and everything you do – either directly or indirectly – causes ripples (good or bad) in one of the 3 areas above. So the next time someone asks you what you do, break it down in those buckets.
For example…
- I manage my executive's most valuable asset – their energy.
- I keep our team on the same page so we don't lose sight of the mission.
- I helped our company raise $15 million.
Now that's powerful.
Where AI genuinely helps
From multiple conversations on the podcast, and from my own work in the AI industry over the years, here are a few examples of where automation wins.
Drafting anything that has a shape.
Meeting agendas, slide decks, briefing docs, a first pass at an email you'll rewrite anyway, notes turned into a summary. The value isn't that the output is good. It's that a first draft is far easier to morph into a finished product than a blank page.
Summarizing volume.
Long email threads, recorded calls, a fifty-page board pack you need the three relevant pages from. This is one of the highest ROI use cases for assistants who want help with cognitive tasks, and it's low-risk because you can check AI's answers against the source.
Research you'd rather not do.
Background on a person before a meeting, an unfamiliar industry term, three options for a venue. Verify anything you'll act on, and treat the output as a starting point rather than a final answer.
Removing the repetitive middle.
Not the thinking, the plumbing. Moving information between systems, filing, formatting, chasing. In episode 339 of The Leader Assistant Podcast, Cortney Hickey, executive business partner to the CEO of Zapier, talks about scaling a role this way, and in episode 384, Beth Portesi talks about how she builds her own tools after twelve years as an EA.
Where AI doesn't work, and where it's dangerous
Of course, there are areas where AI doesn't work so well, or to put it bluntly – it's dangerous.
Anything where being wrong is expensive.
Numbers going to a board. Names and titles in an external email. Dates. A model that's right ninety percent of the time is useless for a task where the tenth time costs your reputation.
Anything that's really a relationship.
The message that needs to sound like your executive, to someone who knows them well. Sure, you can draft it with AI, but you butter make sure it's not full of AI slop. (Yes, I said butter. It's my quirky human thing.) Close relationships will notice, and the cost of being caught sending generated warmth can be much higher than the time saved.
Confidential information, until you get the green light from IT/security.
This is the one that gets assistants in real trouble. Confidential board material, personnel matters, pre-announcement financials, anything covered by an NDA. Find out what your company's policy actually is before you paste anything into any AI tool, and if there's no policy, assume the answer is no. You have access to more sensitive material than almost anyone, which is exactly why this matters more for you than for a colleague experimenting with the same tool.
How to actually start
The assistants who've made AI work well didn't begin with a tool. They began with a task, a problem, a specific use case. Here's a quick guide to get started with AI.
- Pick one recurring task you do at least weekly, that takes more than fifteen minutes, and where being wrong is cheap. Meeting notes into a summary, or calendar automation are good first steps.
- Do it both ways for two weeks. Your way and the tool's way. You're finding out whether it's actually faster once you include the information verification (by the way, that would make a good band name), which is the part many assistants leave out of the estimate.
- Keep what survives. Most experiments don't. That's a successful experiment, not a failure. Don't be afraid to nix a workflow that simply isn't working.
- Then add a second task. Not a second tool. If an AI system works, give it more work to do!
A couple good episodes to check out on this topic are Mike Todasco on leveraging AI as an assistant (episode 329) where he covers the landscape from someone who ran innovation at PayPal, and Naomi White (episode 218), a senior EA at a company that builds AI models, who talks through what she actually uses them for.
If you want to be walked through some specific examples, Beth Portesi and I ran a Claude workshop for assistants on putting this into a real workflow.
Getting better output
One big difference between assistants who find these tools useful and those who don't is how much context they give.
A one-line request gets a generic answer. What works is telling it who you are, who the output is for, what the constraints are, and what good looks like – roughly what you'd tell a capable new team member on their first day.
Two habits worth building:
Keep your good prompts.
When something produces a genuinely useful result, save the wording. Most of the value of a year with these tools is the dozen prompts you end up reusing weekly. Or better yet, create a reusable skill.
Give real world examples.
Paste in three agendas you wrote, then ask for a fourth in the same style. Output that sounds like you is almost always a context problem, not a tool problem. Much like you would teach a new hire, you have to teach the AI.
The elephant in the room
Some version of "will AI replace assistants?" comes up every time I train corporate EA teams, post a podcast about AI, or speak at a conference for assistants. So to put my position plainly: I don't think so, and I haven't for years – I wrote my argument down in my five ways to future-proof your career article back in 2019. This was long before Claude, ChatGPT, and other LLMs.
What I do think is that the role requires more flexibility and curiosity than it ever has. The scheduling and the inbox-sorting will get more and more automated, but the judgment and the partnership won't, and assistants who lean into the latter will continue to be more valuable to executives and their organizations.
I've written an updated version of my thoughts on how to future-proof as an executive assistant in the age of AI here, and episode 343 with Leah Warwick and Maggie Olson on the state of the profession is a good resource on where the whole field is going.
The assistants who should be concerned about being replaced by AI aren't the ones whose tasks get automated. They're the ones whose entire job is to complete tasks.
Tactics > Tools
I've deliberately not spent this guide talking a lot about specific tools because the tools are evolving and innovating so quickly. For example, right now I love Claude Code and Claude Design, but Google's Gemini AI could release a new model next week and change things significantly.
What I will say is that the tactics you employ matter much more than the specific AI you choose. Assistants getting real value out of AI aren't the ones who picked correctly. They're the ones who picked something, used it every day for a month, and found the three things it was genuinely good at.