Signal Scanner · WORKFORCE, SKILLS & ORGANISATIONAL CHANGE

The Second Workforce: AI Agents Enter the Employment System of Record

Enterprises began rebuilding the machinery of employment in 2026, HRMS worker types, agent-manager roles, C-suite mandates, to govern AI agents as a workforce ahead of displacement evidence, with a 2026-2028 inflection for HR, IT and boards.

The consensus AI-and-jobs debate argues over counts: how many roles displaced, how many augmented, when. Beneath it, a quieter development surfaced in early 2026. Enterprises started rewriting the apparatus of employment to admit a second, non-human workforce: worker-type fields in HR systems, agent-manager roles in job architecture, Chief AI Officer mandates contested between HR and IT, pay bands for people who run agents. The apparatus is arriving ahead of the evidence, which shows limited displacement. The question boards should hold through 2028 is not how many humans AI replaces, but who governs digital labour and on what terms.

Signal Identification

A shift in organisational design and governance rather than a capability story: the categories through which firms administer work, worker types, org charts, manager roles, mandates, pay philosophy, now extend to AI agents. The weak signal: employment machinery built for humans is being rebuilt for a hybrid workforce before the productivity record justifies it.

Time horizon: 1–3 years (worker-type and mandate decisions land 2026-2027; EU AI Act high-risk obligations on HR uses of AI and rebuilt job architecture bind 2027-2028) Plausibility band: Medium–High Geographic / Jurisdictional Scope: Global; US-led enterprise adoption, EU regulatory overlay via the AI Act, India moving early on Chief AI Officer appointments Sectors exposed: HR and people functions, enterprise software (HCM and ITSM vendors), professional and financial services, IT operations, workforce planning, compensation and benefits

What's Changing

KPMG reports Chief AI Officer adoption rose from 11% of large enterprises in 2023 to 26% in 2025, projected above 40% of the Fortune 500 by end-2026, and that the HRMS employment-type pick list, unchanged for two decades, is about to gain an “AI Agent” entry; one US enterprise software company runs roughly 3,000 internal agents at a 3:1 agent-to-employee ratio, its workforce redrawn into builders, agent managers, system managers and front-liners, with million-dollar bands for those who build or manage AI systems. KPMG's formulation: "The CIO deploys the agent. HR integrates it as a worker." (KPMG in India, 30/06/2026).

Microsoft's 2026 Work Trend Index, from trillions of Microsoft 365 signals and 20,000 surveyed AI users in 10 countries, names four human-agent collaboration patterns, author, editor, director, orchestrator; organisational factors carry more than twice the AI impact of individual ones (67% vs 32%), and only 13% of workers say they are rewarded for reinventing work with AI (Microsoft, 05/05/2026). Independent analysis adds that 66% of AI users report more time on high-value work, yet only about one in five sit where skill and organisational readiness reinforce each other (Forbes, Moor Insights & Strategy, 19/05/2026).

Governance lags. Across 9,000-plus leaders in 89 countries, 65% of organisations say their culture must change significantly because of AI; 60% of executives use AI in decisions but only 5% manage it well; 66% say functions such as HR, IT and legal must change while 7% report progress, a gap Deloitte calls culture debt (Deloitte, 04/03/2026).

The categories run ahead of the governance, 2026

Large enterprises with a CAIO, 2023 vs 2025 (KPMG/IBM IBV) 11% 26% Organisations saying culture must change significantly because of AI (Deloitte) 65% Executives using AI in decisions vs saying they manage it well (Deloitte) 60% 5% Workers rewarded for reinventing work with AI: 13% (Microsoft, 2026 Work Trend Index)

Sources: KPMG in India (30/06/2026); Deloitte (04/03/2026); Microsoft (05/05/2026).

Disruption Pathway

Stage one, 2026-2027: administrative admission. Agents acquire worker types in the HRMS, registries, lifecycle rules and named human managers; CAIO mandates split into a platform role under the CIO and a strategy role contested by the CHRO. Stage two, 2027-2028: consequences attach. Pay philosophy tilts toward value created, workforce plans state human and agent counts side by side, and the EU AI Act's high-risk classification of most HR uses of AI gives the apparatus teeth. Governments build their own registers: California's May executive order set a 180-day clock on AI-specific layoff-notice rules and a public dashboard of AI's employment effects, after Connecticut and New York amended their WARN laws for AI-driven layoffs (HR Executive, 22/06/2026).

Stresses concentrate at the HR-versus-IT boundary, where ownership of the agent record decides budgets and standards; at the middle manager, re-tasked as an agent manager without a settled playbook; and at the junior pipeline, as agents absorb the formative tasks apprenticeship was built on. Two adaptations follow: joint human-agent capacity plans and shared dashboards operationally; works councils and regulators extending employment-style oversight to agent decisions about people.

Why This Matters

For CHROs and CIOs, the mandate question is live now: whichever function holds the agent record will set job architecture, pay philosophy and governance defaults for a decade. For boards, the apparatus installed in 2026 embeds decision rights that outlast any single AI investment case, and Deloitte's numbers say most firms build it without matching governance. For HCM and ITSM vendors, agent-as-worker features are a contest for the system of record. For investors, the gap between category adoption and the ILO's evidence baseline shows which transformation stories are administrative rather than economic.

Decision-action posture for this signal: Prepare — the worker-type and mandate decisions are being made now but their consequences bind from August 2026 (EU high-risk HR obligations) through 2028; settle the agent-governance mandate and job-architecture approach before vendors and defaults settle it for you.

Counter-Argument

The strongest objection: the second workforce barely works yet. The ILO's review of the empirical record finds "productivity gains are real albeit often unverified and uneven", large-scale displacement limited, and worker time savings of a few per cent of hours not yet translated into measured output, earnings or employment (International Labour Organization, 01/06/2026). Its companion brief warns that exposure indicators "cannot be interpreted as predictions of job displacement, productivity gains or reskilling needs" (International Labour Organization, 17/04/2026). On this reading, agent worker types and CAIO titles are governance theatre, and Deloitte's 5% figure shows the practice hollow.

The counter-counter: the signal claims institutionalisation, not proven productivity. Categories are sticky; the mandates, registries and pay structures installed now will govern digital labour whether or not the output arrives, and the EU's high-risk rules attach regardless.

Implications

This reads as durable change in how employment is administered: categories, mandates and statutes move slower than sentiment, and all three shifted the same direction inside six months. The window runs to 2028, when the CAIO trajectory, EU enforcement and the first rebuilt job architectures will show whether HR or IT won the mandate. Deloitte's tipping-point framing is the canonical statement of the reset (Deloitte, 04/03/2026). Firms that settle governance early, with honest transition plans, gain; firms that let the mandate default to IT tooling inherit the culture debt.

Early Indicators to Monitor

Disconfirming Signals

Strategic Questions

Keywords

AI agents; digital labour; agentic workforce; HRMS; system of record; agent manager; Chief AI Officer; job architecture; human-agent teams; workforce planning; EU AI Act; organisational design

Bibliography

Source tiers: Tier 1, governments, regulators and intergovernmental bodies. Tier 2, think-tanks, academic institutes, major consultancies and quality data providers. Tier 3, quality journalism and specialist trade press. Tier 4, vendor, company and practitioner sources, used only as directional corroboration.

This Signal Scan is one component of a Shaping Tomorrow intelligence cycle on Workforce, Skills & Organisational Change. Every monitored topic also includes monthly Change Trackers that assess what is materially shifting in momentum, and a quarterly Decision Intelligence briefing that converts the cycle into board-level strategic implications. Cycles may also close with Action Triggers, an end-of-cycle synthesis that reads across the three formats and surfaces patterns no single deliverable carries on its own. Continuous intelligence on this topic is available as a Single-Topic or Multi-Topic programme.

Discuss a Tailored Programme

Prepared by Shaping Tomorrow: 4 July 2026