The Supervised Automation Stack

// Proposed for the SRAR 2027 education calendar. Pending Business Technology Committee approval. 45 minutes on Zoom, recorded, free to members.

What This Class Does.

Rungs 2 and 3 taught agents to build AI systems they run by hand. This class puts those systems on a trigger. A new inquiry arrives, the AI drafts the reply, the CRM note, and the next task, and the agent approves from a phone. Attendees learn the three automations worth building first, the disclosure and consent rules that must be written before any bot talks to a consumer, and the 10-question red-team script that finds a bad answer before a client does. Money, contracts, and legal questions stay human. Everything else gets faster.

Come If This Is You.

• You completed rung 2. Rung 3 helps but is not required.
• You have a CRM, or you are choosing one this year.
• You have lost a lead to a slow reply or an after-hours voicemail and you know it.

You Leave Able To.

1. Explain the supervised automation loop: Trigger, Draft, Queue, Approve, Send, Log, and where the human gate sits in each step.
2. Decide which of the three first automations fits your business: speed-to-lead reply, after-hours voice and text answering, or database reactivation.
3. Write the approval rules and the disclosure language before the automation goes live.
4. Red-team a bot with a 10-question script and recognize what a failing answer looks like.
5. Measure the result: response time, contact rate, appointments set, and errors caught at the gate.

The Eight Points.

• The supervised loop: Trigger, Draft, Queue, Approve, Send, Log
• The three automations worth building first: speed-to-lead, after-hours answering, database reactivation
• Live demo: an inbound lead arrives, the AI drafts the reply, the CRM note, and the next task, and the agent approves from a phone in one tap
• Voice and text AI that answers when you cannot: what it should say, what it must never say, and where it hands off to you
• Disclosure and consent: California bot disclosure, texting consent, and DRE guidance on supervision
• The 10-question red-team script you run before any bot talks to a consumer
• Where automation belongs and where it does not: money, contracts, and legal questions stay human
• Measuring it: response time, contact rate, appointments, and the error log

Forty-Five Minutes, Timed.

Built On Screen, Not Described.

What You Walk Out With.

The Rules That Do Not Move.

• Bot disclosure: California requires a clear, conspicuous disclosure when a bot communicates with a person online for commercial purposes. The templates say it plainly and early.
• Texting consent: marketing texts need prior express written consent. Keep the opt-in record. Honor STOP immediately. Your CRM should enforce both.
• Broker supervision: DRE guidance is clear that the broker supervises licensed activity. An automated message is still your message and your broker's responsibility.
• Fair housing: the bot answers neighborhood, school, and demographic questions with sourced facts or a hand-off, never an opinion. The red-team script tests this directly.
• Money, contracts, legal: the bot never discusses price flexibility, commission, contract terms, or legal questions. It hands off. That rule is written before go-live, not after.
• Privacy: consumer data captured by the bot lives in the CRM under your brokerage's policy, not in a public consumer AI.

Every rung of the series runs the S.A.F.E. gate before anything leaves the office. Sensitive information removed. Accuracy confirmed through reliable sources. Fair housing and fiduciary responsibilities protected. Everything properly disclosed and approved.

For The Committee.

Instructor: Connor MacIvor, REALTOR®, CalDRE #01238257, SYNC Brokerage. 27+ years in Santa Clarita Valley real estate. 43+ years writing code. 1,000+ brokers and agents trained in AI. Former SRAR board member.

Questions Agents Ask.