Skip to content
All free classes

Free class · Boundaries + Reply Pack

How do I stop clients texting me late at night as a real estate agent?

Send clients one message with your reply windows, your after-hours rule and what counts as an emergency, then answer non-urgent late texts in the morning. AI can draft that message, ready replies in your voice and a Monday update, as long as you take out names, addresses and amounts first and send each one yourself.

In the window

The slides, in the shop window

The public slides from the class. Use the buttons or your arrow keys, and press F for full screen. Every slide is also written out below.

Boundaries + Reply Pack. End the 7-to-10 Text Trap

Slide 1 of 5: Boundaries + Reply Pack

The same slides, in words

  1. Slide 1 · Opening

    Boundaries + Reply Pack

    AI Acceleration

    Real Estate

    INNOVATION LAB · NORTH

    End the 7-to-10 Text Trap

    Presented by Build Things That Build Things

  2. Slide 2 · Three rules

    Ryan Wanner

    Build Things That Build Things

    I teach real estate professionals how to use AI without the hype. Utility, not slop. Systems, not one-offs. Three rules for today: (1) Tool Minimalism — ChatGPT or Claude plus copy-paste, nothing else. (2) Privacy first — redact names, addresses, amounts before pasting. (3) Draft does not equal send — a human reads and decides every time.

  3. Slide 3 · The idea

    Boundaries are not punishments.

    "Clear expectations prevent disappointment. Proactive communication prevents chasing. Reply systems prevent blank-page paralysis."

    The boundary insight

    A CRM does not set expectations. A human does. This class teaches the human conversation that makes everything else work.

  4. Slide 4 · The habit

    A draft is not a send.

    AI builds the scaffold. You own every send.

    Every boundary card, every reply, every batch update is a draft for your review. You read it, edit it to sound like you, fill in the [SLOTS], and decide whether to send it. Human approval gate on everything.

  5. Slide 5 · The build

    Communication Boundary Card (live)

    Open ChatGPT or Claude. Paste Prompt 1 from the Prompt Pack (scanned QR has it).

    Fill in the brackets. Current pattern: "7am-10pm, within 30 minutes." Desired pattern: "I batch replies at 10am and 3pm on business days."

    Define emergency. Wire issue, closing-day crisis, or deadline in next 4 hours. Everything else waits.

    Weekly update promise. "Every Monday 9am you get an update, even if nothing changed."

    Review the output. Does it sound like you? If not: "Make it more casual" or "Make it more formal." Pair-share with neighbor.

Door sign

Taught in person. Replay not yet found.

Class recording

Taught in person

No verified replay

Ryan Wanner taught this class in person at Innovation Lab North, Compass Lakeside, Hendersonville on August 14, 2026. We have not located a verified recording of it.

The slides above, the steps below, and the workbook with its prompts provide the current practice companion.

Each class was taught live by Ryan Wanner to working real estate agents.

Three signs

What goes on your door

Boundaries aren't punishments. They tell clients how your shop runs, so nobody has to guess. A CRM can't set those expectations. You can, with three signs that each do one job.

How I work

Replies
10am and 3pm, business days
After 8pm
Emergencies only
Emergency
A wire issue, a closing-day crisis, or a deadline in the next 4 hours
Your update
Every Monday at 9am, even if nothing changed
The class example, from slide 5. These are starting settings, not research. Post hours you can keep.
  1. The hours sign

    Communication Boundary Card

    One message you send each client once. It sets when you reply, what happens after hours, what counts as an emergency, and when they'll hear from you without asking.

  2. The flip signs

    8 ready replies in your voice

    Short templates with fill-in slots for the messages you write over and over. Built from 3 messages you actually sent, so they sound like you.

  3. The Monday notice

    Weekly proactive batch

    A short Monday routine. One or two sentences to every active client, even when nothing changed. You move first, so they don't have to chase you.

Opening routine

Put your hours up, step by step

The order the class runs it. AI writes every draft. You decide what gets sent.

  1. Choose reply windows under your actual policy

    Check brokerage requirements, active deadlines and agreed client expectations before choosing reply windows. The class times—10am, 3pm and an 8pm non-urgent cutoff—are examples, not universal response or emergency rules.

  2. Use approved, minimal inputs

    Practice with fictional messages. For real work, confirm the account and task are approved, then supply only permitted excerpts needed for that decision. Remove unnecessary identifiers and identifying combinations; aliases alone do not make material anonymous. Do not paste wire instructions. Check training, retention and sharing controls separately.

  3. Propose urgency labels for a small practice set

    Use fictional messages, or a few permitted excerpts, plus the applicable deadlines and brokerage response policy. Ask AI to propose URGENT, TODAY, THIS WEEK or NO REPLY NEEDED with evidence. Review each label yourself; AI does not decide the response obligation or delete anything.

  4. Draft and review a boundary message

    Ask for a short note using your approved reply windows, after-hours policy, emergency escalation and update cadence. Check it against brokerage requirements and existing client commitments. Keep it as a draft until the exact content and recipient are reviewed for an authorized send.

  5. Apply the policy to an after-hours example

    Check the message against actual deadlines and your approved escalation policy before deciding whether it can wait. Draft the next-window response for a routine case. Route possible emergencies through the established human process; do not rely only on an AI urgency label.

  6. Build draft replies from permitted samples

    Use three fictional or permitted examples that show your tone without unnecessary client details. Ask for short drafts with blanks for recurring situations. Compare each draft with the source facts and voice samples before saving it for reuse.

  7. Prepare one accurate update

    Start with one fictional or permitted deal and its source-dated status. Ask for a brief draft that separates confirmed changes, unknowns and the next check. Removing names and addresses alone does not authorize uploading a client roster. Review facts, recipient and permission before sending.

  8. Read every draft before it goes out

    A draft is not a send. Fill the blanks, make it sound like you, and cut fake urgency, facts you haven't checked, promises you can't keep and anything about who a place suits.

Use the class workbook · Get the supporting files

Read and work through the lesson here. Downloadable files use your free student account.

Reply times

How soon each message gets an answer

Copy a day of client texts, take out the private details, and ask AI to sort each one into four levels with a one-line reason. Ask it to point out any this-week message you've been treating as urgent. That's your urgency bias. Run the sort every day for five days to reset it.

How fast I answer

  1. UrgentWithin 2 hours

    A wire issue, a closing-day problem, a contract deadline in the next few hours, or a client stuck at a property.

  2. TodayBy end of business day

    A lender question, a showing request for tomorrow, or offer feedback someone's waiting on.

  3. This weekWithin 2 to 3 business days

    A general market question, a property research request, or a check-in with no deadline.

  4. No reply neededNo reply

    Thanks, got it, see you then. Or anything already resolved.

These are the class's starting windows. Set your own and hold them. The sorting prompt is Prompt 2 in the pack.

Lettering

Words from the slides, quoted

With no verified replay available, these lines come straight from the class slides. Each slide label opens that slide in the window above.

Slide 3Boundaries are not punishments.

  • “Clear expectations prevent disappointment. Proactive communication prevents chasing. Reply systems prevent blank-page paralysis.”

    Each part of the system has one job: the boundary message sets expectations, the Monday update gets ahead of check-ins, and ready replies save you from a blank screen.

  • “A CRM does not set expectations. A human does.”

    No tool sets your reply hours for you, so the fix is one clear conversation with each client.

Slide 5Communication Boundary Card (live)

  • “I batch replies at 10am and 3pm on business days.”

    A reply window is a set time you answer messages, which replaces a promise to answer the moment they arrive.

  • “Wire issue, closing-day crisis, or deadline in next 4 hours. Everything else waits.”

    An emergency is a short list you write down ahead of time, and anything not on it waits for your next window.

  • “Every Monday 9am you get an update, even if nothing changed.”

    The boundary message also promises something back: a weekly update on a fixed day, so clients don't have to ask.

Slide 2Ryan Wanner

  • “redact names, addresses, amounts before pasting.”

    Client names, addresses and dollar amounts come out of anything you paste into an AI chat.

Slide 4A draft is not a send.

  • “You read it, edit it to sound like you, fill in the [SLOTS], and decide whether to send it.”

    AI writes the first draft of every message in this class, and you decide what actually gets sent.

Taken down

Signs to take down

Habits that undo a boundary, and what to do instead.

  • Answering non-urgent texts at night

    Every quick reply at 9pm teaches the client that 9pm works. If it isn't on your emergency list, it can wait for your morning window.

  • Making the boundary sound like a scolding

    “Don't text me after 8” puts the problem on the client. “Here's how I work so I can serve you well” sets the same limit and keeps things warm.

  • Leaving “emergency” undefined

    If you never say what an emergency is, every message can feel like one. Write down the few things that can't wait, and check the list against your brokerage's policy.

  • Pasting client details into a chat

    Names, addresses, amounts and contact details don't belong in an AI chat. Wire instructions never do, even with details removed. Verify wires by phone, on a number you already know.

  • Sending the AI draft as written

    AI builds the first draft, not the last one. If it doesn't sound like you, tell it to make it more casual or more formal. Then read it once more before you send.

  • Buying a tool to fix expectations

    A CRM or automation tool can't set expectations with a client. One clear conversation does. The class uses the AI chat you already have, plus copy and paste.

Three checks before a reply goes out

  • No fake urgency

    Cut ASAP when it isn't urgent, made-up deadlines, guilt language, and routine things called critical.

  • No unchecked facts or promises

    No property details you haven't checked, no guarantees, and no legal advice you aren't licensed to give.

  • Fair Housing clean

    No comments on protected classes and no steering language about who a neighborhood suits.

Prompt 4 and Prompt 8 in the pack help you run these checks. You still read every draft yourself.

Next door

Other doors on this street

Next questions, and the pages that answer them.

All free classes

Behind the counter

The workbook, from behind the counter

What you build: Communication Boundary Card + 8 voice-matched reply templates + weekly proactive batch plan to stop the 7-to-10 text trap.

Pick one AI to work in

You only need one. Both work the same way here: open a chat, paste, read.

ChatGPT (free or paid tier)

One AI tool and copy and paste. Check the current plan, file support and usage limits for your account; practice can begin with manual review.

Best for All workflows in this class. Boundary Card design, message triage, reply pack generation, weekly batch planning.

Set it up

  1. Go to chat.openai.com and sign in or create a free accountCheck the current plan, file support and usage limits for your account; practice can begin with manual review.
  2. Check your data settings: Settings (bottom left) > Data controls > Improve the model for everyoneCheck the current official setup instructions and account limits. Do not assume a subscription, feature or authorization from this historical class.
  3. Start a new chat and paste Prompt 1 from the Prompt PackTest it with a fake scenario first: pretend client, pretend boundary. See if the output feels like something you would actually send.

Your first five minutes

  1. Check the data settings before pasting any client-adjacent content.
  2. Run the privacy and account-permission check on anything you paste: names? addresses? amounts? Redact to [CLIENT], [PROPERTY], [AMOUNT].
  3. Copy Prompt 1 (Communication Boundary Card Writer) and fill in the brackets with your actual desired boundaries.
  4. Read the output. Does it sound like you? If not, tell ChatGPT: 'Make it more casual' or 'Make it more formal' until it does.
  5. Keep every output in draft. A draft is not a send.

Claude (free or paid tier)

Alternative to ChatGPT. Same copy-paste workflow, slightly different voice.Check the current plan, file support and usage limits for your account; practice can begin with manual review.

Best for Same workflows as ChatGPT. Some agents prefer Claude's tone for client-facing drafts. Try both and pick one.

Set it up

  1. Go to claude.ai and sign in or create a free accountCheck the current plan, file support and usage limits for your account; practice can begin with manual review.
  2. Check your data settings: Click your profile (bottom left) > Settings > PrivacyCheck the current terms and settings for your exact Claude account. Consumer and commercial accounts have different data-use rules; feedback and safety cases also matter. Do not infer privacy from a paid plan.
  3. Start a new chat and paste Prompt 1 from the Prompt PackTest it with a fake scenario first, same as ChatGPT.

Your first five minutes

  1. Check the privacy policy in Settings before pasting any client-adjacent content.
  2. Run the privacy and account-permission check: names? addresses? amounts? Redact.
  3. Copy Prompt 1 (Communication Boundary Card Writer) and fill in the brackets.
  4. Compare the output to ChatGPT's version if you ran both. Pick whichever sounds more like you.
  5. Keep every output in draft.

The jobs, one prompt each

Job 1

Communication Boundary Card (expectation reset)

You'll need Prompt 1 from Prompt Pack + Your honest current + desired communication patterns

See Prompt 1 in the prompt pack below for the full prompt. Paste it into ChatGPT/Claude, fill in the brackets with your patterns, and generate your boundary card.

What you get

  • Subject line + 3-4 short paragraphs + bulleted list of 4 boundaries.
  • Tone: warm, professional, matter-of-fact. Not defensive, not apologetic.
  • Frame: 'This is how I work so I can serve you well,' not 'You are a problem.'

Test it on one client first. The first boundary conversation is the hardest.

Go to Prompt 1

Job 2

24-hour message triage (learn your urgency bias)

You'll need Prompt 2 from Prompt Pack + 24 hours of redacted client messages

See Prompt 2 in the prompt pack below. Paste your 24 hours of messages (redacted) and let AI categorize: URGENT / TODAY / THIS WEEK / NO REPLY NEEDED.

What you get

  • Each message labeled with category + one-sentence explanation.
  • Compare the AI classification with source deadlines and policy; a person decides whether escalation is needed.
  • Run this daily for 5 days to recalibrate your urgency meter.

Count the messages in each category using the actual deadline and your agreed policy. Other messages may be TODAY, THIS WEEK or NO REPLY NEEDED. Use the actual deadline and agreed response policy.

Go to Prompt 2

Job 3

8 ready replies in your voice (Style Match)

You'll need Prompt 3 from Prompt Pack + 3 sent messages where you sounded like yourself

See Prompt 3 in the prompt pack below. Paste 3 sent messages, get 8 templates: acknowledgment, delay, update, no, next-step, holding-pattern, checking-in, emergency-only.

What you get

  • Each template 1-3 sentences with [BRACKETED] fillable slots.
  • Compare the draft voice with your supplied writing samples: casual if you're casual, formal if you're formal.
  • Keep them in a note. Fill the relevant slots, review facts and permission, and send only when authorized.

Update quarterly. Your voice evolves. Re-run Prompt 3 every 3 months.

Go to Prompt 3

Job 4

Monday proactive batch (one update per active client)

You'll need Prompt 6 from Prompt Pack + List of active deals with current status

See Prompt 6 in the prompt pack below. Paste your deal list, get one short update per client formatted as a standalone text.

What you get

  • 1-2 sentences per client.
  • If nothing changed, say so honestly: 'No update this week — lender still reviewing.'
  • Review the updates at your chosen time and deliver only through authorized channels.

Send an authorized update when it helps. Measure replies and unresolved questions instead of assuming a reduction.

Go to Prompt 6

Rules before anything goes out

  • I run the privacy and account-permission check before every paste: names? addresses? amounts?

    Use an approved account and minimum permitted context. Check identifying combinations and hidden data; aliases alone do not anonymize a person.

  • I never paste wire instructions into AI, even redacted

    Wire verification happens by voice on a known phone number. That's it.

  • I apply the four categories using the agreed brokerage/client policy and actual deadlines.

    Agree on the policy with the client, apply it consistently and review observed misunderstandings.

  • I never send a reply with fake urgency (ASAP, artificial deadlines, guilt language)

    Prompt 4 checks for this. Use it.

  • Every output I generate is a draft for my review. A draft is not a send.

    I read it, edit it to sound like me, fill in the [SLOTS], and decide whether to send it.

  • I sent the Communication Boundary Card to at least one client this week OR ran triage daily for 5 days

    Pick one and do it. The system only works if you implement it.

Opening week

  1. TODAY

    Write your Communication Boundary Card and send it to one client

    Pick the easiest client first — someone who will understand. Test the boundary conversation with them.

  2. TOMORROW

    Run 24-hour triage on your messages

    Copy-paste 24 hours (redacted) into Prompt 2. Learn what's actually URGENT vs what trained you to reply instantly.

  3. DAY 3

    Generate your 8 ready replies (Style Match)

    Paste 3 sent messages into Prompt 3, get your reply pack, save it in a note on your phone.

  4. DAY 5

    Run your first Monday proactive batch

    List your active deals, paste into Prompt 6, send the updates at 9am Monday. Record whether the update answered the client’s actual question.

  5. ONGOING

    Hold the boundary: reply windows, not instant replies

    URGENT follows the agreed escalation procedure. TODAY and THIS WEEK use the documented deadline and response window. NO REPLY NEEDED requires human confirmation. Review the actual result and adjust the agreed policy when needed.

Take one

Take one: the prompt pack

Inbox Without Burning Out prompt pack, 8 prompts. Copy one, paste it into ChatGPT or Claude, and fill in the brackets with your own details. Swap names, addresses, and amounts for [CLIENT], [PROPERTY], and [AMOUNT] first. Read every draft before you send it.

  1. Prompt 1

    Communication Boundary Card Writer

    What it's for Generate your one-page expectation script that resets client communication.

    When to use it Before you onboard a new client, or when you need to reset expectations with an existing client who texts at 9pm.

    I am a licensed real estate agent. I need to write a Communication Boundary Card — a one-page expectation script I send to clients that sets response time, after-hours policy, emergency definition, and weekly update cadence.
    
    Here is my current communication pattern (be honest):
    - I currently respond to texts: [describe your current pattern, e.g., "7am to 10pm daily, usually within 30 minutes"]
    - My after-hours boundary is: [current state, e.g., "nonexistent" or "8pm cutoff for non-emergencies"]
    - I define emergency as: [your definition or "I haven't defined it"]
    - I currently send proactive updates: [frequency, e.g., "never" or "when I remember"]
    
    Here is my DESIRED communication pattern (the boundary I want to set):
    - Response time: [e.g., "I batch replies at 10am and 3pm on business days"]
    - After-hours: [e.g., "Emergency texts only after 8pm. Emergency = wire issue or closing-day crisis."]
    - Emergency definition: [e.g., "Wire instructions, closing-day issue, or contract deadline in next 4 hours"]
    - Weekly update: [e.g., "Every Monday 9am you get a proactive update on your deal, even if nothing changed"]
    
    Write a Communication Boundary Card I can send to clients. Keep it warm, professional, and matter-of-fact. The tone should be "this is how I work so I can serve you well," not "I'm setting rules because you're a problem."
    
    Format:
    - Subject line
    - 3–4 short paragraphs
    - Bulleted list of the 4 boundaries
    - Closing that reassures them this is FOR them, not AGAINST them
    
    Do not invent my voice. Write it neutrally and I will adjust the tone myself.
  2. Prompt 2

    24-Hour Message Triage

    What it's for Learn what's actually urgent vs what trained clients to expect instant replies.

    When to use it Daily for 5 days, or when you need to reset your own sense of urgency.

    I am a licensed real estate agent. I have copied 24 hours of client messages below (redacted for privacy). Triage every message into exactly one category:
    
    URGENT — Follow the supplied brokerage/client escalation policy and actual deadline. Do not impose a universal two-hour response window or delay an immediate emergency.
    
    TODAY — Needs action during the current business day under the supplied policy and source deadline. Examples require human confirmation.
    
    THIS WEEK — The known deadline and agreed policy allow later handling. Return the actual next review/action date; ask if it is missing.
    
    NO REPLY NEEDED — Informational, or already resolved. Examples: "Thanks!", "Got it", "See you then".
    
    For each message, label it with the category and explain in one sentence why you assigned that category. If I have been treating THIS WEEK messages as URGENT, point that out.
    
    Here are my 24 hours of messages (all names/addresses replaced with [CLIENT] and [PROPERTY]):
    
    [PASTE YOUR REDACTED MESSAGES HERE]
  3. Prompt 3

    Batch Reply Pack in Your Voice (Style Match)

    What it's for Generate 8 ready replies in your voice that you can review and send.

    When to use it Once during class, then update quarterly as your voice evolves.

    First swap names, addresses, and amounts in your three messages for [CLIENT], [PROPERTY], and [AMOUNT].

    I am a licensed real estate agent. I need 8 ready-reply templates in MY voice for common client communication situations. To match my voice, here are 3 real messages I sent where I sounded like myself:
    
    [PASTE 3 SENT MESSAGES HERE — texts, emails, or DMs where you liked how you sounded]
    
    Now generate 8 reply templates in that same voice for these 8 situations:
    
    1. ACKNOWLEDGMENT — "I got your message, working on it now"
    2. DELAY — "I'm researching this, I'll have an answer by [TIME]"
    3. UPDATE — "Here's where we are: [STATUS]. Next step: [ACTION]."
    4. NO — "That's outside my expertise, here's who to call: [REFERRAL]"
    5. NEXT-STEP — "Got it. Next step for you: [ACTION]. Next step for me: [ACTION]."
    6. HOLDING-PATTERN — "We're waiting on [PARTY]. I'll follow up [DATE] and update you."
    7. CHECKING-IN — "No update yet, but I'm still on it. I'll know more by [DATE]."
    8. EMERGENCY-ONLY — "I'm off-hours now. If this is an emergency (wire, closing issue, deadline in next 4 hours), text me. Otherwise I'll reply at [TIME] tomorrow."
    
    Keep each template short (1–3 sentences). Make the [BRACKETED] parts fillable slots I can customize per message. Use my writing samples as evidence and flag uncertain style choices — if I'm casual, be casual. If I'm formal, be formal. If I use contractions, use contractions.
  4. Prompt 4

    "Do Not Invent Urgency" Validator

    What it's for Check a draft reply for fake urgency or pressure tactics before you send it.

    When to use it Before sending any message where you're tempted to add urgency to get a faster reply.

    I am a licensed real estate agent. I wrote the draft message below to a client. Before I send it, check it for fake urgency, pressure tactics, or anything that trains them to expect instant replies when none is needed.
    
    Flag these red flags:
    - "ASAP" or "urgent" when it's not actually urgent
    - Artificial deadlines ("need to hear back by end of day" when you don't)
    - Guilt language ("I'm waiting on you to move forward")
    - Overstatement ("this is critical" when it's routine)
    
    If you find any, rewrite the sentence WITHOUT the fake urgency. If the draft is clean, say "No fake urgency detected."
    
    Here's my draft:
    
    [PASTE YOUR DRAFT MESSAGE HERE]
  5. Prompt 5

    Boundary Reminder Draft (After-Hours)

    What it's for Politely remind a client who texted you at 9:30pm that you have boundaries.

    When to use it The first time a client texts you after hours, or when you need to re-set a boundary.

    I am a licensed real estate agent. A client texted me at [TIME, e.g., 9:30pm] about [SUBJECT, e.g., "a showing request"]. This is not an emergency (emergency = wire issue, closing-day crisis, deadline in next 4 hours).
    
    I need to reply tomorrow morning with a warm, professional boundary reminder that:
    1. Acknowledges their message
    2. Answers their question
    3. Reminds them of my after-hours policy without sounding annoyed
    4. Reassures them I'm on top of their deal
    
    My after-hours policy is: [e.g., "Emergency texts only after 8pm. I batch replies at 10am and 3pm on business days."]
    
    Write a 2–3 sentence reply I can send tomorrow at 10am that hits all 4 points. Keep it friendly and matter-of-fact.
  6. Prompt 6

    Weekly Proactive Batch Planner

    What it's for Write Monday's proactive batch, one short update per active client.

    When to use it Every Monday at 9am (or your chosen batch time).

    I am a licensed real estate agent. It's Monday morning and I'm sending my weekly proactive batch — one short update per active client, even if nothing changed this week.
    
    Here's my current deal status (all names/addresses redacted):
    
    [CLIENT A] — Listing at [PROPERTY]. Status: [e.g., "Under contract, appraisal ordered Friday, expect report Wednesday"]
    [CLIENT B] — Buyer search. Status: [e.g., "Sent 3 new listings Friday, scheduling showings Tuesday"]
    [CLIENT C] — Pre-approval stage. Status: [e.g., "Lender confirmed pre-approval Friday, good to show homes"]
    
    For each client, write a 1–2 sentence proactive update in a warm, professional tone. If nothing changed, say so honestly (e.g., "No update this week — lender is still reviewing docs. I'll follow up Wednesday and let you know.").
    
    Format each update as a standalone text message I can copy-paste and send.
  7. Prompt 7

    After-Hours Auto-Reply Text

    What it's for Set a text auto-reply that enforces your boundary without sounding robotic.

    When to use it Draft the wording first. Check current official documentation for your exact phone, messaging app and account. Do not assume a scheduled general auto-reply exists. A manually saved reply is a valid fallback; test any supported automation with permission before relying on it.

    Put your own number where it says [PHONE], for real emergencies only.

    I am a licensed real estate agent. I need an after-hours auto-reply for texts I receive after [TIME, e.g., 8pm] and before [TIME, e.g., 7am].
    
    The auto-reply should:
    1. Acknowledge receipt
    2. State my response time (e.g., "I batch replies at 10am and 3pm on business days")
    3. Use my supplied emergency definition and independently verified escalation contact; do not invent coverage or response times.
    4. Sound warm and professional, not robotic
    
    Write a 2–3 sentence draft I can use manually or with a separately verified, authorized auto-reply feature. Keep it under 160 characters if possible so it works as SMS.
  8. Prompt 8

    Quality Gate / Fair Housing Check

    What it's for Check a batch of replies for compliance issues before you send them.

    When to use it Before sending any batch of client messages, especially if you generated multiple replies at once.

    I am a licensed real estate agent. I wrote the draft messages below using AI. Before I send them, run a compliance check for Fair Housing issues, fabricated facts, or anything that could create legal risk.
    
    Flag these red flags:
    - Any comment on protected classes (race, color, religion, sex, national origin, familial status, disability, plus classes your state adds)
    - Any steering language ("this neighborhood is better for families like yours")
    - Any invented property facts I didn't verify (square footage, year built, school ratings stated as fact)
    - Any promise I can't keep ("I guarantee this will appraise")
    - Any legal advice I'm not licensed to give ("you don't need an inspection" or "this contract clause means X")
    
    For each message, identify issues and missing evidence. If none are found, say "No issue detected in this limited review," not legally cleared. A responsible person must review facts and applicable requirements.
    
    Here are my drafts:
    
    [PASTE YOUR DRAFT MESSAGES HERE, NUMBERED 1–8]

Pick-up

Pick up the class files

This week, pick one

Send your boundary message to one client, starting with the one most likely to understand. Or run the urgency sort every day for five days. Then name one boundary you'll actually keep. The first conversation is the hardest one.

Create a free student portal account to download this class's files.

  1. Inbox Without Burning Out prompt pack

    Eight paste-ready prompts: Communication Boundary Card Writer, 24-hour triage, 8 reply templates (Style Match), fake-urgency validator, boundary reminder draft, weekly proactive batch planner, after-hours auto-reply, and Fair Housing compliance check.

    Markdown · 12 KB · Free account

    Get the file: Inbox Without Burning Out prompt pack
  2. Communication Boundary Card (fillable template)

    The one-page expectation script: response time, after-hours policy, emergency definition, weekly update cadence. Fill in the brackets, adjust the tone, send to one client this week.

    Markdown · 5 KB · Free account

    Get the file: Communication Boundary Card (fillable template)
  3. Message Triage Rubric

    URGENT / TODAY / THIS WEEK / NO REPLY NEEDED. Learn what's actually urgent vs what trained clients to expect instant replies. Run this daily for 5 days to recalibrate your urgency meter.

    Markdown · 2 KB · Free account

    Get the file: Message Triage Rubric
  4. Privacy Redaction Checklist

    The check before pasting into AI: names? addresses? amounts? Redact to [CLIENT], [PROPERTY], [AMOUNT]. Print this and tape it next to your monitor.

    Markdown · 2 KB · Free account

    Get the file: Privacy Redaction Checklist
  5. Class skill card

    The class method on one page, written so you can hand it to your AI as instructions.

    Markdown · 15 KB · Free account

    Get the file: Class skill card

Questions about using this class

Practical answers from the lesson and workbook.

How do I choose reply windows without ignoring urgent issues?

Start with windows you can keep. Define which issues need immediate attention and who owns them, then tell the client how those exceptions work. Use the urgency exercise to check your rule against realistic messages.

Can I practice without pasting client messages?

Yes. Use a fictional message or remove identifying details from an example you are authorized to use. Keep account details, sensitive transaction information and unnecessary personal facts out of the input.

What should a boundary message actually say?

State when the client can expect a response, how to reach the appropriate person for a genuinely urgent issue, and what update you will provide next. Do not promise availability or progress you cannot deliver.

Should I let AI reply while I am unavailable?

This class's drafts still need your review. Check urgency, current facts, tone and recipient before sending. A tool connection or an existing draft is not permission to respond on your behalf.

Boundaries that hold

Boundaries are the start. A plan for your week is the rest.

One boundary message and a Monday update get you ahead of the check-in texts. The Agent Constraint Accelerator looks at the rest of your week, names the one constraint slowing your business down, and builds the first working system against it with you.

Fine print

The fine print, with sources

Check your AI account's data settings on the vendor's own page before you paste anything about a client. The Fair Housing check on every reply comes from the law itself.