The Macquarie Bank 2026 Benchmarking Report confirmed what the conference circuit has been buzzing about: 80% of Australian strata businesses are now using AI tools. But buried in that headline is a more telling number — 24% are deploying AI agents. Not chatbots. Not copilots. Autonomous systems that execute multi-step workflows with minimal human oversight.

That 24% figure marks the beginning of a structural shift in how strata operations work. This article separates the signal from the noise on agentic AI in property management, maps where it creates genuine value in strata, and draws the line where it must stop.

The Three Waves of AI in Strata

To understand where the industry is heading, it helps to see where it has been.

AI Adoption in Strata — Three Waves

WAVE 1
2023–2024
Assisted: Generative AI tools used for communication drafting, template generation, meeting minutes summarisation. The human initiates every action. AI suggests; the manager decides.
Productivity gain: incremental · Risk: low · Adoption: broad but shallow
WAVE 2
2024–2025
Copilot: AI embedded into platform workflows — auto-classifying maintenance requests, drafting owner correspondence, flagging compliance deadlines. The human remains in the loop but the AI handles more of the preparation.
Productivity gain: moderate · Risk: moderate · Adoption: growing among enterprise firms
WAVE 3
2025–2026
Agentic: Autonomous AI agents that execute complete workflows end-to-end. An agent receives a maintenance request, diagnoses priority from historical data, checks budget availability, generates a work order, contacts the preferred supplier, and schedules the job — without a human touching the workflow until the exception threshold is triggered.
Productivity gain: transformational · Risk: requires architectural controls · Adoption: 24% and accelerating

The jump from Wave 2 to Wave 3 is not incremental. It is architectural. A copilot helps you do your job. An agent does part of your job. That distinction changes everything about how platforms are designed, how compliance is maintained, and how firms scale.

Where Agentic AI Creates Value in Strata

Not every strata workflow is a candidate for autonomous execution. The value map has clear boundaries.

Agentic AI Value Map — High Value: Automate Aggressively

AUTOMATE
Owner inquiry triage and response — High volume, pattern-based, low statutory risk
AUTOMATE
Maintenance request classification and routing — Historical data enables accurate priority scoring
AUTOMATE
Supplier invoice matching and AP processing — Rules-based matching against POs and contracts
AUTOMATE
Meeting notice generation and distribution — Template-driven with jurisdiction-specific rules
AUTOMATE
Insurance renewal tracking and compliance certificate monitoring — Date-driven, penalty-linked, audit-critical

Medium Value: Agent Prepares, Human Decides

PREPARE
Budget preparation and levy setting — Financial authority requires committee approval
PREPARE
Work order approval above threshold — Expenditure authority limits apply
PREPARE
By-law breach escalation — Legal judgment and proportionality required
PREPARE
Committee report drafting — Strategic framing and relationship context

No-Go Zone: Deterministic Only

NO-GO
Trust account reconciliation — Statutory liability, zero-tolerance for error
NO-GO
Levy calculation and apportionment — Mathematical precision required by legislation
NO-GO
Sinking fund / capital works fund transfers — Regulated fund movements with audit obligations
NO-GO
AGM voting tabulation — Governance integrity, legal standing of resolutions

This last category is the critical boundary. As we wrote in February's Deterministic vs Probabilistic, trust accounting and financial compliance require hard-coded rules engines — not probabilistic models. Agentic AI amplifies that principle. An autonomous agent that makes a trust accounting decision based on pattern matching rather than statutory rules is not just a risk — it is a potential breach of fiduciary duty.

The Economics of Agentic Operations

The Macquarie benchmarking data tells the economic story. The average strata business manages 415 lots per FTE, up 19% from 2022. Firms at the frontier are pushing well beyond that. But the 415 figure is an average — and averages mask the distribution.

Productivity Distribution — Lots per FTE

280–350
Manual-heavy firms — Headcount-dependent, low automation
415
Industry average — Partial automation, some AI adoption
500–650
Automation-led firms — Workflow automation, Wave 2 AI
650+
Agentic-ready firms — End-to-end autonomous workflows for Tier 1 tasks
The frontier: $180,000–$250,000 annual salary savings per 10,000-lot portfolio vs industry average

The gap between 350 and 650+ lots per FTE is not explained by harder-working staff. It is explained by how much of the operational baseline runs without human intervention. Agentic AI is the technology that pushes the frontier from 500 to 650+ — by eliminating the human touchpoints in high-volume, pattern-based workflows.

What the Platform Must Do Differently

An agentic architecture is not a chatbot with more permissions. It requires five structural capabilities that most strata platforms do not have today.

Five Requirements for Agentic Platform Architecture

1
Deterministic Guardrails: Every agent must operate within explicit authority boundaries. Financial thresholds, compliance rules, and governance constraints are hard stops. The agent handles the workflow up to the boundary. At the boundary, it hands off to a human. No exceptions, no overrides, no learned behaviours.
2
Audit Trail by Default: Every action an agent takes must be logged with the same rigour as a human action. Who initiated the workflow? What data did the agent use? What decision did it make? What was the outcome? If you cannot audit an agent's decision chain, you cannot deploy it in a regulated environment.
3
Exception Escalation: Agents must know what they do not know. When a maintenance request doesn't match any historical pattern, the agent escalates — it does not guess. The quality of an agentic system is measured by the quality of its escalation logic, not by the volume of its autonomous decisions.
4
Jurisdiction Parameterisation: Strata regulation varies by state. An agent that sends a meeting notice must know whether it is operating under NSW, Victorian, or Queensland rules — and apply the correct notice periods, quorum requirements, and voting procedures. This is not a configuration setting. It is a runtime parameter.
5
Human Authority Primacy: No governance action without authorised human initiation. An agent can prepare an AGM pack, draft every notice, compile every financial report, and pre-populate every resolution. But the committee chair presses the button. This is not a limitation of the technology. It is a design principle that preserves the legal standing of every decision the scheme makes.

The Turnover Connection

The Macquarie data shows strata manager turnover at 24% — down from 33% in 2022, but still 60% above the national workforce average. The correlation between automation maturity and retention is not coincidental.

Firms where strata managers spend their days on owner communications, relationship management, and strategic advisory retain staff at 83%. Firms where managers spend their days chasing invoices, re-keying data, and assembling AGM packs retain staff at 76%.

Agentic AI does not replace strata managers. It replaces the parts of the job that drive people out of the profession. The 24% that are already deploying agents are not reducing headcount. They are retaining the headcount they have — and deploying it on work that justifies premium fees.

The Bottom Line

Agentic AI is not a feature. It is an operating model. The firms that deploy it successfully will manage more lots per head, retain more staff, and capture more margin — not because the AI is smarter, but because it handles the operational baseline while humans handle the exceptions, the relationships, and the decisions that carry legal weight.

The 24% adoption figure in the Macquarie report is a leading indicator. Within two years, agentic workflows will be table stakes for enterprise strata firms. The platforms that enable it — with deterministic guardrails, audit trails, and jurisdiction-aware agent logic — will define the next productivity frontier.

Agentic AI handles the baseline. Humans handle the boundaries. The platform enforces the difference.

The current release of StrataPort v3.0 scaffolds all concepts outlined in this article as we work to deeply integrate agentic AI into the platform.