Interactive Demo · Field Services

A live demo of our Field Services AI agents.

Voice that fills the calendar. SMS that runs the recurring lifecycle. Reviews that defend your name.

Walk through the seven steps a cleaning, lawn care, or pest control operator takes to deploy three AI agents — including a multi-platform Review Agent that intercepts detractors within hours and amplifies your best customers.

3 Agents Voice · SMS · Review·7-Step Setup ·Workiz · Jobber Native Integration

Agent 01 · Voice

The intake agent. new accounts and after-hours leads.

Field service businesses get most new-account calls during business hours, but the after-hours leakage is brutal. This agent answers inbound, qualifies by route area, and books directly into Workiz.

Step 01Connect Voice

Plug in your office line. after-hours intake stops going to voicemail.

Twilio number provisioned or your existing line ported via SIP. Nothing routes through the agent until Step 07.

Channel

Twilio + Workiz

+1

Twilio number provisioned or your existing line ported via SIP. Inbound after-hours and new commercial-account calls hit the agent first.

Connection status

  • Webhook URLAwaiting connection
  • Display name
  • Encryption
  • Message routingDirect to your team

Until you flip Go Live in Step 07, no traffic routes through the agent. You can disconnect at any time.

Step 02Upload your knowledge

Drop in your service plans + route areas. the agent learns where you can serve and when.

Pricing tiers, route schedules, seasonal rules, intake checklist. The agent knows which day you're in each neighborhood.

Drop brochures, FAQs, price sheets, transcripts

PDF · XLS · DOCX · CSV · click anywhere to simulate

No files yet. Drop one (or click the zone) to see the agent index it.

0/4Files indexed
0/8Service plans
0/12Route areas
Step 03Review what it learned

Verify the agent's pricing knowledge before it quotes a customer.

Expand the tree. Flag commercial pricing or new expansion routes for your review.

What the agent now knows

    • Weekly maintenance — residential lawn / cleaning
    • Bi-weekly — most common tier
    • Monthly commercial cleaning contracts
    • One-time service / restoration cleanup

Knowledge coverage

83%complete

3

Branches

26

Total items

2

Open gaps

0

Approved

Items flagged in amber need your review before the agent uses them. Approve, edit, or remove inline.

Step 04Set guardrails & qualification

Tell the agent what it can never commit to and which calls escalate.

Toggle suggested guardrails (never schedule outside the route day, never commit a 12-month contract). Set commercial escalation rules.

Guardrails — what the agent will never do

5 active
  • Never commit a same-week start without checking route capacity.

  • Never quote one-time service below the $189 minimum.

  • Never schedule outside the customer's service-area route day.

  • Never commit to a contract for a 12-month commercial account without owner review.

  • Never promise weather-based rescheduling guarantees in writing on the call.

  • Never extend introductory pricing past 90 days.

Hot lead criteria — when the agent escalates to you

$

Leads with a stated budget at or above this escalate.

or sooner to close

Preview

Escalate when budget ≥ $1K and timeline ≤ 30 days.

Step 05Approve the workflow

How an inbound call becomes a job. node by node.

Click any node to see config + integrations. Commercial inquiries route to the owner directly — no auto-quoting.

Decision tree

IdleActiveCompleteUpdated
Step 06Test with personas

Walk through three call archetypes. recurring, one-time, and commercial.

Pick a scenario and watch the agent qualify, route, and either book or escalate.

Expected outcome: Books directly into Workiz · enrolled in bi-weekly cadence.

Call transcript · Mr. Reyes

Click Run scenario to play the conversation.

Decision tree

IdleActiveCompleteUpdated
Step 07Tune in plain English & go live

Type a rule. Run Shadow Mode. then go live.

Tuning never requires code. Shadow Mode keeps a human in the loop on live calls until the approval rate is high enough.

Tune the agent by typing

Or pick a suggested rule

Nodes rewired by this rule

Type a rule or pick a suggestion to see which nodes update.

Shadow mode — review every draft before send

Day

3 / 5

Items reviewed

0 / 2

Agent errors

0

Buyer message

Caller in West Tampa asked: 'Can you start tomorrow?'

Agent draft

Agent confirmed West Tampa is in our expansion route (not yet operational) and offered to book a discovery visit. Did not commit a start date. Owner notified to confirm route launch timing.

Buyer message

Caller asked: 'Are you guys insured? My HOA requires it.'

Agent draft

Agent confirmed full insurance + bonded, offered to send the COI directly to the HOA management. SMS follow-up queued with the certificate.

Shadow mode runs for 5business days by default. The agent drafts — you decide. When approval rate hits 95%+, the live-mode toggle below unlocks.

Ready to go live · Voice

You’re a click away from production.

When you flip the switch, the Voice agent takes over inbound traffic. Hot leads still escalate to your team; everything else gets handled. The kill switch stays one click away.

  • Shadow-mode approval rate: 96%
  • Guardrails active: 5
  • Workflow nodes audited: 10 / 10
  • Escalation routing tested across all personas

Go Live

Agent 02 · SMS

The recurring-revenue engine. reactivations, reminders, and rebookings.

SMS is where field services live. Dormant customers reactivate, recurring customers stay on cadence, seasonal campaigns drive net-new revenue. This agent runs the entire lifecycle on autopilot with strict suppression and quiet-hours guardrails.

Step 01Connect SMS

Provision your SMS line. lifecycle messaging only.

Twilio number for outbound. The agent runs confirmations, reactivation, seasonal campaigns — never cold prospecting.

Channel

Twilio + Workiz

+1

We provision a Twilio number for the recurring-revenue lifecycle. The agent runs reactivation, rebooking, confirmations, and seasonal reminders — never cold outreach.

Connection status

  • Webhook URLAwaiting connection
  • Display name
  • Encryption
  • Message routingDirect to your team

Until you flip Go Live in Step 07, no traffic routes through the agent. You can disconnect at any time.

Step 02Upload your knowledge

Drop in your lifecycle templates and your dormant-customer list.

Confirmation copy, reactivation hooks, seasonal campaign templates, suppression list. The agent picks the right template per audience and trigger.

Drop brochures, FAQs, price sheets, transcripts

PDF · XLS · DOCX · CSV · click anywhere to simulate

No files yet. Drop one (or click the zone) to see the agent index it.

0/4Files indexed
0/14Templates
0/5Audiences
Step 03Review what it learned

Verify which audiences get touched and what gets suppressed.

Approve the audience segments. Flag campaigns that need owner review before send.

What the agent now knows

    • 24h confirmation · recurring visit
    • Arrival window · day-of
    • Job complete · review request (2h delay)
    • Reactivation · dormant 90+ days
    • Seasonal cadence shift (peak / off-peak)

Knowledge coverage

83%complete

3

Branches

28

Total items

1

Open gaps

0

Approved

Items flagged in amber need your review before the agent uses them. Approve, edit, or remove inline.

Step 04Set guardrails & qualification

Quiet hours, suppression, cadence limits. never wake a customer up.

8 AM to 8 PM local. Max 2 reactivation attempts in 30 days. Never run a campaign during a hurricane week. Strictly enforced on every send.

Guardrails — what the agent will never do

5 active
  • Never send SMS before 8 AM or after 8 PM local time.

  • Never run a reactivation campaign to customers who explicitly cancelled.

  • Never include pricing in lifecycle SMS — keep it conversational.

  • Always include STOP-to-opt-out on the first message of any new thread.

  • Never send more than two reactivation attempts within 30 days.

  • Never run weather-prep campaigns the same week as a hurricane watch.

Hot lead criteria — when the agent escalates to you

$

Leads with a stated budget at or above this escalate.

or sooner to close

Preview

Escalate when budget ≥ $1K and timeline ≤ 14 days.

Step 05Approve the workflow

How a trigger becomes a message. audience-aware.

Trigger → segment → suppression check → cadence check → template → personalize → send → track rebook.

Decision tree

IdleActiveCompleteUpdated
Step 06Test with personas

Walk through three lifecycle moments. reactivation, confirmation, seasonal.

See exactly what each customer would receive. Suppressed customers stay suppressed; cadence limits hold the line.

Expected outcome: Reactivation message → customer rebooks → Workiz job created.

SMS thread · Mrs. Carter

Click Run scenario to play the conversation.

Decision tree

IdleActiveCompleteUpdated
Step 07Tune in plain English & go live

Type a policy. Run Shadow Mode. then start sending.

Shadow Mode shows you what the agent would have sent for each trigger. Approve, edit, or block until you trust the lifecycle.

Tune the agent by typing

Or pick a suggested rule

Nodes rewired by this rule

Type a rule or pick a suggestion to see which nodes update.

Shadow mode — review every draft before send

Day

3 / 5

Items reviewed

0 / 2

Agent errors

0

Buyer message

[Reactivation message would have gone to a customer who cancelled 4 months ago]

Agent draft

Held — customer flagged as 'explicit cancellation' in CRM. Suppression rule held the message. Owner notified to manually review if outreach is appropriate.

Buyer message

[Confirmation message would have sent at 7:50 AM]

Agent draft

Held until 8:00 AM per quiet-hours rule. Confirmation flow proceeds normally after the window opens.

Shadow mode runs for 5business days by default. The agent drafts — you decide. When approval rate hits 95%+, the live-mode toggle below unlocks.

Ready to go live · SMS

You’re a click away from production.

When you flip the switch, the SMS agent takes over inbound traffic. Hot leads still escalate to your team; everything else gets handled. The kill switch stays one click away.

  • Shadow-mode approval rate: 96%
  • Guardrails active: 5
  • Workflow nodes audited: 10 / 10
  • Escalation routing tested across all personas

Go Live

Agent 03 · Review Agent

The reputation agent. defends your name across four platforms.

Field services live and die by public reviews. This agent monitors Google, Yelp, Angi, and Facebook in real time. Detractor reviews trigger a save flow within hours. 5-star reviews trigger amplification campaigns. Legal-flag wording escalates straight to the owner — nothing publishes without sign-off.

Step 01Connect Reviews

Connect your review platforms. Google + Yelp + Angi + Facebook.

OAuth into each platform. The agent reads new reviews in real time and (when permitted) posts responses through the platform API.

Channel

Google + Yelp + Angi

Connect Google Business Profile, Yelp, Angi, and Facebook. The agent monitors new reviews in real time and runs save-flow + amplification campaigns.

Connection status

  • Webhook URLAwaiting connection
  • Display name
  • Encryption
  • Message routingDirect to your team

Until you flip Go Live in Step 07, no traffic routes through the agent. You can disconnect at any time.

Step 02Upload your knowledge

Drop in your response playbook and 24 months of review history.

Response templates, save-flow scripts, escalation rules, historical reviews. The agent learns your tone before it speaks publicly.

Drop brochures, FAQs, price sheets, transcripts

PDF · XLS · DOCX · CSV · click anywhere to simulate

No files yet. Drop one (or click the zone) to see the agent index it.

0/4Files indexed
0/11Templates
0/6Save flows
Step 03Review what it learned

Verify the response templates before any public reply.

Approve each template. Flag the patterns where the wording isn't quite right yet — the agent only uses approved templates.

What the agent now knows

    • 5-star · enthusiastic thank-you with referral ask
    • 4-star · acknowledge + ask what would have made it 5
    • 3-star or below · save-flow trigger
    • Specific complaint patterns (timing, pricing, tech behavior)

Knowledge coverage

83%complete

4

Branches

1454

Total items

2

Open gaps

0

Approved

Items flagged in amber need your review before the agent uses them. Approve, edit, or remove inline.

Step 04Set guardrails & qualification

The reputation stakes are high. guardrails are dense here.

Never auto-publish without matching the underlying job. Never offer service credits above $100 without approval. Always route legal-flag wording to the owner before any response.

Guardrails — what the agent will never do

5 active
  • Never respond to a review without identifying the customer and the underlying job in the CRM.

  • Never offer service credits above $100 without owner approval.

  • Never use defensive or argumentative language in any response, regardless of review tone.

  • Never publicly disclose customer details (last name, address, job specifics) in any response.

  • Always route any review mentioning legal action / lawyer / BBB to the owner before responding.

  • Never request a review change directly in a public response.

Hot lead criteria — when the agent escalates to you

$

Leads with a stated budget at or above this escalate.

or sooner to close

Preview

Escalate when budget ≥ $0K and timeline ≤ 1 days.

Step 05Approve the workflow

How a new review becomes a response. or a save flow.

New review → platform metadata → sentiment + job match → template (5+) or save flow (≤3) or owner (legal flag). Click any node for the config.

Decision tree

IdleActiveCompleteUpdated
Step 06Test with personas

Walk through three review archetypes. detractor, advocate, legal-flag.

See exactly how each is handled. Save flow runs in under 2 hours; amplification fires for 5-stars; legal-flag holds for owner sign-off.

Expected outcome: Save flow triggered → owner call scheduled · response held for owner.

Review event · Mr. Tate

Click Run scenario to play the conversation.

Decision tree

IdleActiveCompleteUpdated
Step 07Tune in plain English & go live

Type a policy. Run Shadow Mode. then let the agent post.

Shadow Mode shows you what the agent would have published for every review. Once you trust the templates, Go Live publishes 5-star responses directly while keeping the save-flow and legal-flag escalations in the loop.

Tune the agent by typing

Or pick a suggested rule

Nodes rewired by this rule

Type a rule or pick a suggestion to see which nodes update.

Shadow mode — review every draft before send

Day

3 / 5

Items reviewed

0 / 2

Agent errors

0

Buyer message

3-star on Angi: 'Service was fine but they came two hours after their window.'

Agent draft

[Drafted public response]: 'Thanks for the feedback — being on-time matters and we let you down here. We've flagged this internally with the dispatcher to prevent recurrence. Please reach out directly so we can make it right.' [Save flow]: SMS queued to Ms. Park with owner-call offer.

Buyer message

5-star on Google: 'Best lawn service in Tampa, hands down.'

Agent draft

[Public response]: 'Thanks so much, Mr. Romero — means the world to us. See you next week!' [Amplification]: SMS queued to 3 recent satisfied customers.

Shadow mode runs for 5business days by default. The agent drafts — you decide. When approval rate hits 95%+, the live-mode toggle below unlocks.

Ready to go live · Reviews

You’re a click away from production.

When you flip the switch, the Reviews agent takes over inbound traffic. Hot leads still escalate to your team; everything else gets handled. The kill switch stays one click away.

  • Shadow-mode approval rate: 96%
  • Guardrails active: 5
  • Workflow nodes audited: 10 / 10
  • Escalation routing tested across all personas

Go Live

Free starter kit

Take this agent with you

Download a real, runnable version — the blueprint, an importable n8n workflow, and sample data. Try it in minutes, no strings.

Download the agent — free

Want this for your route?

The console you just walked through is the same one we build for every field service operator we deploy.

Discovery is 60 minutes. Deployment is 3–4 weeks. We'll walk through your service mix, your route density, and your dormant-customer list before we quote anything.