Request → Offer Safe Steps → Capture Result → Resolved? → Dispatch With Context

Self-Help Troubleshooting Agent

Guides residents through safe, approved troubleshooting before a tech is dispatched and records what was tried, so the simple fixes get handled without a truck roll and the real ones arrive with context.

The Problem

A tech drives out to flip a breaker the resident could have flipped.

Unnecessary dispatches for simple issues. "Is the breaker tripped?" never gets asked before sending a tech.

What Changes

Safe steps first. Dispatch only when it's really needed.

Walks the resident through safe troubleshooting steps and records what was tried. If the issue resolves, no dispatch needed. If it doesn't, the tech arrives knowing what's already been ruled out.

Humans stay in control: Your team handles every dispatch and sets the approved steps. The agent runs safe self-help first. It removes the needless truck rolls, not the service.

HOW THE AGENT WORKS

TRIGGERLLMAPIDECISIONACTIONWAIT
TRIGGERRequest ArrivesA maintenance request comes in
LLMMatch TroubleshootingFind approved steps for this issue
DECISIONSafe To Self-Help?
ACTIONGuide ResidentOffer the steps one at a time
LLMCapture ResultRecord what was tried and the outcome
DECISIONResolved?
APIDispatch With ContextOpen a ticket noting steps tried
The agent doesn't withhold help. It offers safe, approved steps and captures what was tried, then dispatches the moment those don't resolve it. Anything with a safety angle goes straight to a person.

HOW TO BUILD IT

Three levels of depth. Pick the one that matches where you are.

For the ops person who needs to understand what's involved and brief their team.

Your RequestsResident issues
The AgentWorkflow engine + LLM
Your TechsFewer, better-briefed visits

WHAT YOUR TEAM PROVIDES ONCE

  • Approved troubleshooting - the safe steps per issue type
  • Safety exclusions - issues that must skip self-help and dispatch
  • Capture format - how to record what the resident tried
  • Dispatch handoff - how a ticket inherits the self-help notes
  • Tone and channel - how to guide residents step by step
What happens day-to-day: A resident reports something a safe step might fix, a tripped breaker, a reset. The agent walks them through it and records the result. If it works, no truck roll. If it doesn’t, the tech arrives already knowing what was tried.

For the IT person or technical lead who needs to scope the integration.

INQUIRY SOURCES

Portal / SMS
Resident concierge
Intake agent

WORKFLOW ENGINE (N8N)

Trigger
LLM - Match steps
Branch - safe?
Guide resident
LLM - Capture result
Branch - resolved?
PMS API - Dispatch w/ context

CONNECTS TO

PMS / CMMS (Yardi / Property Meld)Write: ticket with self-help notes
Knowledge baseRead: approved troubleshooting steps
SMS / portalGuide the resident two-way
Intake / triage agentsHand off unresolved with context
WITHOUT PMS API

Agent offers steps from a static guide and emails the team with what was tried. The deflection logic still runs.

WITH PMS API

Agent guides safe steps and, if unresolved, opens a ticket carrying the full self-help record. Best experience.

For the person who wants to build this agent step by step. Each node below maps to an n8n workflow node.

TRIGGERNode 1
Trigger - Request Arrives
From portal/SMS or chained from the concierge/intake agent.
Output: { unit_id, resident_contact, issue_text }
LLMNode 2
LLM - Match & Safety-Check Steps
Find approved steps for the issue and confirm it is safe for self-help.
Output: { safe_to_self_help, steps[], exclusion_reason }
You decide whether a maintenance issue is safe for resident self-help.

Using APPROVED STEPS and SAFETY EXCLUSIONS, return:
- safe_to_self_help: true/false
- steps: ordered safe steps if applicable
- exclusion_reason: why not, if false (e.g. gas, electrical hazard, water damage)

Never offer steps for excluded categories.
Respond ONLY with valid JSON.

APPROVED STEPS: {approved_steps}
SAFETY EXCLUSIONS: {safety_exclusions}
ISSUE: {issue_text}
INPUT
"Microwave and kitchen outlets all dead suddenly."
OUTPUT
{
  "safe_to_self_help": true,
  "steps": ["Check if the GFCI outlet has tripped and press reset", "Check the breaker panel for a tripped kitchen breaker and flip it fully off then on"],
  "exclusion_reason": null
}
DECISIONNode 3
Condition - Safe To Self-Help?
IF safe → guide the resident. ELSE → skip self-help and dispatch immediately.
Output: Branch: SAFE (guide) or EXCLUDED (dispatch now)
ACTIONNode 4
Guide Resident
Send the steps one at a time, waiting for the result of each.
Output: Resident walked through the safe steps.
LLMNode 5
LLM - Capture Result
Record what was tried and whether it resolved the issue.
Output: { resolved, steps_tried[], notes }
You record the outcome of resident self-help.

Given the steps offered and the resident's responses, return:
- resolved: true/false
- steps_tried: which steps they completed
- notes: anything useful for a tech if dispatch is needed

Respond ONLY with valid JSON.

STEPS: {steps_json}
RESPONSES: {resident_responses}
OUTPUT
{
  "resolved": false,
  "steps_tried": ["Pressed GFCI reset - no change", "Checked breaker - none tripped"],
  "notes": "GFCI reset and breaker both ruled out; likely wiring or dead GFCI."
}
DECISIONNode 6
Condition - Resolved?
IF resolved → close with a thank-you. ELSE → open a ticket with the self-help record attached.
Output: Branch: RESOLVED (close) or NOT (dispatch)
APINode 7
HTTP Request - Dispatch With Context
Open a ticket that inherits the steps tried so the tech does not repeat them.
Output: Ticket created with full self-help context.
MVP vs Advanced — Where To Start
MVP — START HERE
  • Single channel
  • Steps from a static guide
  • Safety exclusion list
  • Manual result capture
  • Email dispatch
  • Notes pasted manually
ADVANCED — ADD LATER

    Full Video Walkthrough Coming Soon

    From a request to a resolved fix or a smarter dispatch - step by step.

    Register to be notified

    Everything on this page is free to learn.

    When you're ready to build — do it yourself or bring us in.

    OTHER AGENTS IN LEASING & MARKETING

    Fills the gaps in a request before a tech is dispatched
    Catches duplicate tickets and flags recurring failures
    Converts QA failures into structured follow-up tasks
    Classifies trade, sets priority from history, triggers correct dispatch

    Frequently asked questions

    Self-Help Troubleshooting Agent is an AI agent that guides residents through safe checks before sending a tech.
    A tech drives out to flip a breaker the resident could have flipped. The Self-Help Troubleshooting Agent addresses this — it guides residents through safe checks before sending a tech.
    No — your team stays in control. Your team handles every dispatch and sets the approved steps. The agent runs safe self-help first. It removes the needless truck rolls, not the service.