Level 4 taught you to govern AI across a brokerage. Level 5 is about building the part that acts. A drafting assistant suggests; an automation does. It answers an enquiry, schedules a disclosure deadline, proposes a listing change or writes to the CRM, and when it is wrong the error has already reached a consumer, the MLS or a transaction file by the time anyone looks. This level teaches brokerage automation whose authority is bounded in writing, whose every action is recorded, whose fair housing and consent controls are built in rather than remembered, and whose failures are caught by design. It is written for brokerage technologists, operations and transaction management leaders, CRM, MLS-feed and website owners, marketing automation owners and licensees who own automation projects. No coding is assumed or required.
Modules one to three set authority and build the first orchestrations. Module one places every automation on a ladder from suggesting to acting inside the brokerage, names the three root causes of excessive agency, writes the agent authority register before anything touches a live system, and lists the acts no automation performs, from signing and negotiating to AI-voice calls and payment instructions. Module two builds enquiry intake that records calling and texting consent as data, honours every revocation and do-not-call request, answers availability questions from the record, and prepares the written buyer agreement for a licensee to explain and conclude. Module three automates the dated work of a listing and a transaction: the one-business-day MLS filing clock and exempt listings, altered images labelled before syndication, lead-based paint and California transfer disclosure timing, agreement expiry and a closeout that keeps records as long as the rules require.
Modules four to six are the connections. Module four maps every system an automation touches, explains RESO Data Dictionary resources, fields, lookups and local fields, and writes a data contract at each boundary that keeps altered-image indicators intact. Module five starts MLS, IDX and CRM connections read-only and earns each write, separates proposed from submitted listing changes, leaves other participants' listings untouched, honours seller controls, keeps searches compensation-neutral and records every write. Module six keeps automation near earnest money and closing funds without moving them: no automation both receives and relays payment instructions, every change request goes to a person on a channel held in advance, and a misdirected wire halts the automations while the recall and IC3 report proceed.
Modules seven to nine are human control. Module seven designs approval gates a licensee really decides at, explains how automation bias and workload turn review into rubber-stamping, and measures whether gates catch anything. Module eight builds fair housing controls into automated copy and images, audience selection under 24 CFR 100.75(c)(3) with the Meta settlement as a platform example and HUD's withdrawn 2024 guidance as history only, steering tests for routing and recommendation logic, and chatbot hand-offs for neighbourhood, availability and accommodation questions. Module nine gives every automation its own least-privilege identity, keeps licensed activity with licensees using California's guidance and Texas's written-scope rule, and reviews access.
Modules ten to twelve are defence and observation. Module ten threat-models enquiries, listing remarks, transaction email and uploads from the defender's side only, and teaches prompt injection as a residual risk that controls reduce but no product removes. Module eleven specifies the action log a supervising broker, commission investigator or insurer can follow, what it must never hold, and retention mapped to record rules a brokerage already follows. Module twelve monitors for silent failure, drift after vendor updates, runaway cost and outcomes that differ across neighbourhoods, with thresholds the brokerage chooses while HUD's proposal on its discriminatory effects rule is pending.
Module thirteen responds when an automation has already acted: detect, respond and recover, a clean stop, work in flight, the manual route and a staged return, funds recall and breach notice with California as one example. Module fourteen measures time, quality and cost against the brokerage's own baseline with review time and rework counted, refuses savings claimed twice, writes stop criteria before the pilot, and prepares the capstone: the Automation Implementation Project for a fictional brokerage. The level ships with a printable workbook and ten templates, and the examination draws forty scenario questions from a reviewed bank.
Everything here is professional education. It is not legal advice, and it does not replace a real estate licence, brokerage supervision, MLS rules, counsel or the fair housing, licensing, consumer protection, telemarketing, settlement services and privacy laws of any jurisdiction. Completing the level earns an independent educational certificate issued by AI Coalition Network with a public verification page. It is not a real estate licence or a professional designation, carries no jurisdictional education hours, and satisfies no pre-licensing, post-licensing, renewal, fair housing training or association requirement.