Will AI replace executive assistants? A balanced view.
By The Min · · 6 min read
It's the question on every founder's mind: if AI can draft emails, summarise documents and book meetings, why pay for a human at all? It's a fair question, and the honest answer is more interesting than either side of the hype. AI is genuinely transformative for some of the work. It's also nowhere near ready to own the rest. The point that gets lost in the noise is simple.
Where AI genuinely wins
Let's be fair to the technology, because it deserves it. Used well, AI is brilliant at the speed work:
- First drafts. Emails, briefs, replies and documents, in seconds, ready to be shaped.
- Summarising. A 40-message thread or a 60-page board pack, distilled to what matters.
- Research and options. Suppliers, comparisons, background, gathered in minutes instead of hours.
- Logic and admin. Time-zone maths, scheduling permutations, tidying notes, formatting.
This is real and it's not going away. Ignoring it isn't "responsible", it's just slow. A good assistant who uses AI well gives you more, faster, than one who refuses to.
Where AI quietly falls short
But speed isn't the hard part of executive support. The hard part is everything AI can't be accountable for:
- Judgement. Knowing which of today's twelve "urgent" things actually matters, and which can wait.
- Relationships. The tone with a key investor, the read on a difficult client, the favour quietly remembered.
- Discretion. Knowing what should never be written down, let alone typed into a chatbot.
- Accountability. When something goes wrong, "the AI got it wrong" is not an answer a founder can give a board.
AI also makes confident mistakes. It will invent a detail, miss the context only you know, or cheerfully send the wrong thing to the wrong person. Without a human checking the output, speed becomes a liability.
The responsible-use line
This is where most of the real risk lives, and it has nothing to do with how clever the AI is. It's about the data and the oversight:
- Keep confidential, personal or regulated information out of public tools. Convenience is never worth a leak.
- Keep a human in the loop. AI drafts; a person decides, checks and owns the result.
- Be transparent. The people you work with should be able to trust how their information is handled.
Done this way, AI is a safe accelerator. Done carelessly, it's a confidentiality incident waiting to happen. Same tool, different operator. (It's why discretion, and an NDA as standard, matter more in an AI world, not less.)
So, hand it to AI or keep a human?
The honest answer is "both, deliberately." Here's the split that works:
| Great for AI (speed) | Keep a human accountable (judgement) |
|---|---|
| First drafts of emails, briefs and docs | Final tone on sensitive or high-stakes messages |
| Summarising long threads and documents | Decisions involving money, people or risk |
| Research, options and background | Relationships and stakeholder nuance |
| Scheduling logic and time-zone maths | Knowing what actually matters today |
| Transcribing and tidying notes | Anything confidential, personal or regulated |
The real future isn't AI or a human
It's a capable operator using AI. The assistant who pairs human judgement with AI speed will out-perform both a human working alone and AI working alone, comfortably. That's the model we believe in: AI-powered where it helps, human where it matters. You can see it in practice in our CEO Morning Brief, the kind of thing AI helps assemble but a person curates so it's actually right for you.
The technology will keep getting better. The question that matters won't change: who is holding the leverage, and can you trust how they use it? Pick the operator, not the tool.
Frequently asked questions
Will AI replace executive assistants?
Not for the work that matters most. AI is excellent at speed: drafting, summarising, research and scheduling logic. But executive support is mostly judgement, relationships, discretion and knowing what actually matters today, which AI cannot own or be accountable for. The realistic future is a skilled assistant using AI as a tool, not AI replacing the assistant.
Is it safe to let AI handle my inbox, calendar or documents?
It depends entirely on how it's used. Used responsibly, with a human in the loop and confidential or regulated information kept out of public tools, AI can speed up drafting and triage safely. The risk is never the AI itself, it's putting sensitive data into the wrong place or trusting an output without a person checking it.
What should I automate with AI and what shouldn't I?
Hand AI the speed work: first drafts, summarising long threads and documents, research, transcribing notes and scheduling logic. Keep a human accountable for the judgement work: final tone on sensitive messages, decisions involving money, people or risk, relationships, and anything confidential.
Does The Min use AI?
Yes, where it genuinely helps, and never where it shouldn't. We use AI to move faster on drafting, research and summaries, with a human reviewing the output and confidential information handled discreetly. AI-powered where it helps, human where it matters.