■ HIGH RISK ■ Administrative
The calendar-and-inbox layer of the EA job is being absorbed by AI scheduling and drafting tools right now. What survives is the trusted-operator layer — judgment, discretion, and managing the executive's world — but fewer executives will get a human for it.
“AI assistants schedule meetings and don't need birthday cake.”
Our AI replacement risk score — how we score jobs
A good executive assistant runs an executive's operating system: triaging a flooded inbox, playing calendar Tetris across time zones, booking travel with the boss's unstated preferences memorized, prepping briefing docs, taking board-meeting minutes, screening who gets access, and quietly fixing the things that would otherwise become the executive's problems. The best EAs function as chiefs of staff without the title — they know where the bodies are buried and who actually needs to be in the room.
AI is coming straight for the mechanical layer. Scheduling assistants negotiate meeting times without human email ping-pong; LLM tools draft correspondence in the boss's voice, summarize long threads, and prep meeting briefs from calendars and CRM data; travel platforms rebook a canceled flight before you've heard the announcement. Copilot-style assistants bundled into office suites mean every manager now has a baseline synthetic EA for free, which quietly erodes the argument for giving mid-level executives a shared human one. Administrative support has been shrinking for decades as software absorbed typing, filing, and phones; this is that curve steepening, and WEF's Future of Jobs surveys keep placing executive-secretary roles among the declining ones.
The resistant core is trust and judgment. An AI does not notice the CFO seemed rattled and quietly move the hard conversation; it doesn't know which board member requires white-glove handling or when 'no' from the calendar should be 'yes' because of politics invisible to any system. EAs handle confidential material — compensation, deals, firings — where executives want a person bound by loyalty, not a model bound by a terms-of-service. Expect the role to bifurcate: fewer, more senior EA/chief-of-staff hybrids supporting top executives, while the routine-scheduling tier below them dissolves into software.
Automatability: our editorial assessment of current and near-term AI capability
The pressure is immediate: AI scheduling and drafting tools are deployed in most large organizations already, and the ratio of executives to human assistants keeps stretching through the late 2020s. By ~2030 expect shared-EA and junior-EA roles to be substantially thinner, while senior EAs to C-suite executives persist as trust-based operator roles. The job's survival strategy has a deadline measured in years, not decades.
The bottom tier is. AI now handles scheduling, drafting, travel, and inbox triage well enough that mid-level executives increasingly get software instead of a shared human assistant. Senior EAs supporting C-level executives remain in demand because that job is really about trust, discretion, and judgment. The occupation shrinks and moves upmarket rather than vanishing outright.
Read the room. AI can't gauge which meeting request hides a political landmine, handle a confidential termination logistics chain, manage a difficult board member's ego, or notice the boss is overloaded and quietly restructure the week. EAs operate on context, loyalty, and institutional knowledge that no tool ingests — which is exactly the part of the job worth doubling down on.
The mechanical work — calendars, email drafts, travel — is being absorbed now; most organizations already run some of it through AI. The realistic near-term picture is not mass layoffs but stretched ratios: one EA supporting more executives, fewer junior openings, and 'assistant' expectations folded into everyone's software. Assume the routine layer of the job is gone within a few years.
Convert the role's hidden seniority into an explicit one. EAs already do project management, operations, and stakeholder handling — retitle and reskill toward chief of staff, business operations, or event/program management. Meanwhile, master the AI tools yourself; the person who orchestrates the automated assistant stack is far harder to cut than the person doing what it does.