Occupations

Secretaries & Administrative Assistants

SOC 43-6014Knowledge / AIOctober 2026 assessment2025 employment 1.89M

Perform routine administrative support such as correspondence, scheduling, records, phone communications, and office coordination.

Occupation description · O*NET

Current assessmentOctober 2026Independent analysis for a changing world of work.

Current EOL outlook

Forecast and analytical state

See how this assessment evolved →

2030 working outlook−5% to −2%

−10%0%+10%

2035 working outlook−10% to −6%

−10%0%+10%

2035 reference points

BLS structural projection−6.0%

Metaculus probabilistic forecastNo current match

BLS remains a structural reference; no defensible Metaculus occupation match is currently used.

Bottom line

Administrative information handling is increasingly machine-scalable, and agentic systems can combine scheduling, correspondence, documents and workflow coordination. The occupation is already in structural decline and AI reinforces that trajectory.

Six-stage assessment

Where the evidence stands now

Same framework used across every EOL occupation

01CapabilityAdvanced; progressing rapidlyView evidence → 02Economic ViabilityStrongView evidence → 03AdoptionHigh use and broadening; agentic workflow automation remains uneven across organizationsView evidence → 04Task SubsumptionMaterial, heterogeneousView evidence → 05Human Labor DependencyModerateView evidence → 06Labor-Market EffectsEstablished historically; current AI increment emergingView evidence →

Transparent by design. Every stage judgment is traceable to the evidence supporting the current assessment.

View all Secretaries & Administrative Assistants evidence →

Evidence TiltStrong evidence tilt toward displacement

What this means →

DisplacementBalancedAugmentation

Historical ContinuityIncreasingly questionable

What this means →

ReasonableIncreasingly questionableUnlikely reliable

View all evidence in Signal Tracker →Read the methodology →Jump to labor supply →

Detailed occupation assessment

Forecast ContextView detail ›

Employment / inflection

Contraction underway

Forecast rationale

The 2030 range is retained. By 2035, the core downside still assumes additional agentic consolidation beyond BLS, but the more aggressive -12% case is treated as a stress scenario rather than a core bound.

Demand / countervailing pressure

Organizational growth creates administrative activity, but agents can absorb more of the incremental workload.

Key mechanism

Agentic administrative workflows; employees or executives supported per assistant.

How EOL translates evidence into a forecast range →Worked methodology example →

Six-Stage Transition Assessment

Expanded analysis

Capability

Scheduling, correspondence, information handling, drafting and routine office coordination are highly compatible with current AI systems. | INCREMENT X26: Salesforce Agentforce agent offerings; limitation: Capability/product claims not occupation labor effect | Qualification: Group average overstates SOC-specific adoption.

Advanced; progressing rapidly
Moderate-high confidence

Evidence Synthesis

Evidence tilt

Strong evidence tilt toward displacement

Directional summary of whether current evidence pushes EOL toward greater or lesser human labor demand relative to the occupation’s existing structural trajectory. It is categorical, not a probability, percentile, automation share, or mathematical adjustment.

Historical continuity

Increasingly questionable

Administrative information handling can increasingly scale through software and agents, while managers and professionals become more self-sufficient, weakening historical support-staffing relationships.

Labor Supply

Pipeline: Open-entry / employer-trained
Supply trend: No discrete pipeline; entry remains broad and flexible
Replenishment burden: 8.7% of employment annually
Demand-supply alignment: Well positioned to accommodate contraction through attrition and reduced replacement hiring

Supply implication for the forecast

Secretarial and administrative support has few formal entry barriers and a relatively old incumbent workforce. As AI reduces routine administrative workload, retirements, occupational exits, and transfers create a strong natural-attrition channel. The occupation can therefore contract materially through nonreplacement without requiring layoffs on the same scale as the decline in employment.

How EOL analyzes labor supply →

Forecast Bridge

The analysis above is the evidence base. The bridge below summarizes how those findings translate into the 2030 and 2035 ranges without repeating the full six-stage assessment.

What BLS already captures

BLS explicitly identifies AI-enabled routine back-office productivity as a factor reducing O10 utilization. EOL does not count document preparation, scheduling or notetaking a second time.

Why EOL still differs

The residual departure is broader agentic, cross-system administrative workflow execution that raises support ratios beyond the routine AI and office automation already recognized by BLS.

2030−5% to −2%

Neutral structural bridge: -3%. This is an EOL benchmark derived from the BLS ten-year path, not a BLS 2030 forecast.

Lower-employment edge

At the lower-employment edge, EOL assumes agentic administrative workflows let managers and professionals handle more coordination themselves and allow each remaining assistant to support more people. That puts employment about 2.0% below the neutral structural reference. The 2030 bound remains constrained by the limited time available for diffusion and organizational redesign.

Upper-employment edge

At the upper-employment edge, EOL assumes BLS already captures much of the routine AI effect and organizational growth plus persistent coordination needs keep the additional EOL departure small. That puts employment about 1.1% above the neutral structural reference.

2035−10% to −6%

BLS structural projection: −6.0%.

Lower-employment edge

Agentic workflows and higher support ratios compound through attrition and reduced replacement hiring, producing a decline materially below BLS but short of the prior stress case.

Upper-employment edge

BLS-like decline prevails: AI raises productivity but exception handling, coordination and organizational growth prevent additional staffing compression.

Why not materially higher or lower?

The prior -12% 2035 lower edge is retained as a stress scenario pending stronger O10-specific support-ratio and nonreplacement evidence. A materially higher outcome would require the employment-supportive conditions to dominate more strongly than current evidence supports: Executive coordination, exception handling and relationship management remain labor-intensive; organizations preserve dedicated support despite automation; growth in organizational complexity offsets productivity.

September revalidationv1.8 independent validation adjudicated

What Would Change Our View?

Toward greater displacement

Agentic scheduling, correspondence and document workflows become default.

⌄

Agentic scheduling, correspondence and document workflows become default; employees or executives supported per assistant rises materially; dedicated support roles are consolidated.

Toward greater employment

Executive coordination, exception handling and relationship management remain labor-intensive.

⌄

Executive coordination, exception handling and relationship management remain labor-intensive; organizations preserve dedicated support despite automation; growth in organizational complexity offsets productivity.

Signal Tracker

See how this assessment evolves

EOL preserves meaningful assessment updates, forecast holds, forecast changes, and BLS structural rebases over time.

Open Signal Tracker →

Sources and Assessment History

Current assessment

October 2026EOL occupation assessment

Structural reference

BLSEmployment Projections · SOC 43-6014

External calibration

No current matchNo defensible Metaculus occupation match

Supply evidence

BLS / CPSAge profile, separations and replacement demand