Feature · Reimagine Hub

AI text triage: emergency, estimate, or support classified in 3 seconds

Deco AI reads the message before your dispatcher does. Emergency goes to the on-call queue. Active customer support routes with job history attached. Every message lands in the right place.

Book a live demo See how it works in 4 steps

Coming soon · 4 industries planned · Founded by PhDs (ex-Google, Meta)

✓ Response under 2 minutes✓ Included in Starter✓ No per-message fee✓ Onboarding in 1 week

The pain you recognize

Why an emergency buried in 40 unread texts costs you customers and revenue

Two hours of manual triage every morning, missed emergencies over the weekend, longtime customers waiting behind new leads. Sound familiar?

Emergency buried in 40 unread texts at 9 PM

A pipe burst Friday night. The customer texted 'URGENT water everywhere'. Your dispatcher was off. It sat in the general inbox until Monday morning. Between those two days the customer called three competitors and filed a claim. You lost the job and the relationship.

Dispatcher reads every message just to decide who handles it

Eighty inbound texts per day. Your office manager spends the first two hours reading and deciding: estimate queue, emergency queue, support queue, or spam. That is 10 hours of classification per week that produces no revenue.

New estimate request and a warranty callback land in the same pile

A homeowner asking for a new HVAC quote is not the same as a customer whose unit you installed six months ago and is now not cooling. One needs a sales follow-up, one needs a technician dispatched today. Without triage, both wait in the same queue.

Longtime customer waits in the new-lead line for hours

Sarah has been your customer for four years. She texted asking for a copy of her invoice from last month. Without triage she queued behind 12 new estimate requests. By the time your dispatcher replied, she had already called your office twice and was frustrated.

How it works

How AI text triage works in the Reimagine Hub

From an inbound message to the right queue in 4 steps. Classification in 3 seconds, around the clock:

Step 1: Message arrives, Deco AI reads it before the dispatcher

Every inbound text, photo, or voicemail transcript is intercepted by Deco AI before appearing in the dispatcher's queue. The AI reads the full content, checks if the sender is a known customer, and evaluates urgency signals.

Step 2: Classified into one of four categories in under 3 seconds

Emergency (pipe burst, no power, safety risk), field visit needed (on-site diagnosis required), standard estimate (pricing inquiry or scheduling), or active-customer support (question from an existing customer). Confidence above 85% routes automatically. Below 85% goes to the dispatcher with the suggested category highlighted.

Step 3: Each category follows its own workflow

Emergency: on-call tech is alerted immediately and an automated acknowledgment with ETA goes to the customer. Field visit: enters the diagnostic calendar with a pre-estimate text sent before arrival. Standard estimate: enters the normal lead pipeline. Active-customer support: opens a support ticket with the customer's full job history attached.

Step 4: Dispatcher opens an already-sorted queue

When your office manager opens Hub in the morning, the inbox is organized by priority: emergencies resolved or in progress at the top, visits scheduled, new leads in order, active-customer support separated. No reading through every message to decide.

Real example

Triage in action: two real scenarios

How Deco AI separated an emergency from a routine estimate and kept both customers happy.

Plumber receives an emergency text after hours and responds in under 4 minutes without a dispatcher on duty

Persona

Lisa is in Phoenix. Friday at 10:40 PM a supply line under her kitchen sink burst. She texted the plumbing company in a panic. The dispatcher had clocked out at 6.

10:40 PM
Lisa
Texts: 'SOS pipe under sink burst water everywhere help'.
Lisa
SOS pipe under sink burst water everywhere help I don't know what to do
10:40 PM
10:40 PM
Deco AI
Classifies as EMERGENCY. Confidence 97%. Fires on-call alert to the on-call technician's phone.
Hub
🚨 EMERGENCY classified. Lisa M. — 'pipe burst water everywhere'. Confidence 97%. Alert on-call now.
10:40 PM
10:41 PM
Hub
Automated acknowledgment sent to Lisa with ETA and first step to limit damage.
Lisa, our on-call tech has been alerted and will call you within 10 minutes. While you wait: shut off the water supply valve under the sink or at the main shutoff near the meter.
10:44 PM
On-call tech
Calls Lisa, confirms address, and heads out. Job opened automatically in Hub with Lisa's full customer history.
Lisa
Thank you! I turned it off. So relieved someone is coming.
10:48 PM

Outcome

Lisa stayed calm. Emergency handled in under 5 minutes with no dispatcher on duty. Job created with full history. No competitor had time to respond.

Before vs. after

Manual dispatcher vs. menu chatbot vs. CRM tags vs. Reimagine Hub: four approaches to message triage

Without Reimagine Hub

Dispatcher reads 80 messages per day just to classify them. Two hours of triage every morning with no revenue output.
Emergency lands in the general queue over a weekend. Customer calls three competitors before Monday.
New estimate and active-customer warranty callback treated the same. Wrong response sent.
Longtime customer waits in the new-lead line for hours. Leaves a one-star review about response time.
Holiday weekend: inbox fills up unchecked. Tuesday morning starts with 120 messages in arrival order.
No priority logic: dispatcher replies to whoever texted most recently, not most urgently.

With Reimagine Hub

Deco AI classifies every message in 3 seconds. Dispatcher opens an already-sorted priority queue.
Emergency identified in seconds. On-call tech alerted. Automated acknowledgment sent. Customer stays.
Field visit follows a diagnostic workflow. Estimate follows the lead pipeline. Each in the right place.
Active-customer support opens with full job history. One-tap reply. Customer resolved in minutes.
Weekend: emergencies handled, visits scheduled. Monday starts clean with a sorted inbox.
Queue sorted by urgency, not arrival time. Dispatcher focuses on whoever needs help most right now.

By the numbers

What businesses report after adopting AI text triage

AI triage does not replace your dispatcher. It gives them a pre-sorted queue so they focus on responses that need a human. Metrics observed in similar operations:

2h → 0h

Daily manual triage time per dispatcher

Deco AI classifies before the dispatcher opens the inbox

3 sec

Average classification time per message

vs 2-5 minutes of reading and deciding per message

-85%

Emergencies without a same-day response

on-call alert fires automatically on any risk-signal message

+40%

Active-customer satisfaction with support speed

support question resolved without waiting behind new leads

Results vary by volume, service catalog, and configuration. Figures observed across Reimagine operations running this feature for 3+ months.

By industry

AI text triage across every industry

Emergencies, field visits, estimates, and active-customer support happen in every service business. See how Deco AI adapts triage for each industry:

Live

AI Inbound Text Triage for Home Services

Works for any trade: HVAC, plumbing, roofing, electrical, painting, landscaping

Home Services contractors using the Reimagine Hub

Live

AI Inbound Text Triage for Auto Shop

Repair orders, diagnostics, warranty, customer communication

Auto shops using the Reimagine Hub

Live

AI Inbound Text Triage for Real Estate

Lead pipeline, listing management, buyer/seller communication

Real estate teams using the Reimagine Hub

Live

AI Inbound Text Triage for Property Management

Tenant portal, rent collection, maintenance routing

Property managers using the Reimagine Hub

What sets us apart

Reimagine Hub vs. manual dispatcher vs. chatbot vs. CRM: what makes triage different

There are several ways to manage inbound messages today. Here is why AI triage is not the same as a menu bot or manual tagging:

Manual dispatcher, no AI

What it is: The standard model: a human reads and classifies every inbound message before routing.
Limitation: Two hours of triage per day, emergencies reach the wrong queue after hours, human error during high-volume periods, fixed labor cost for a function that can be automated.
Hub: Deco AI classifies in 3 seconds, 24 hours a day, 7 days a week. Dispatcher opens a priority-sorted queue. Emergency on-call alert fires automatically without a human reading the inbox first.

Menu-based chatbot (press 1, 2, or 3)

What it is: A bot that asks customers to pick a category before the conversation continues.
Limitation: A customer with a pipe burst is not going to navigate a menu. Active customers hate being forced through a phone tree. High abandonment rate under stress. Cannot read natural language, photos, or voicemail transcripts.
Hub: Deco AI understands natural language, photos, and audio. The customer texts 'water everywhere' and AI classifies it as emergency without the customer doing anything. No menu, no friction.

Generic CRM with manual tagging

What it is: A platform that lets the dispatcher manually tag each conversation (urgent, visit, support) after reading it.
Limitation: Manual tagging requires a dispatcher to be available. Inconsistent tagging standards across team members. No on-call alert for after-hours emergencies.
Hub: Hub tags automatically via Deco AI. Consistent classification 24/7. Sunday night emergency alerts on-call before any human sees the message.

Direct comparison

AI text triage: Reimagine Hub vs. the alternatives

How each approach handles inbound message classification.

Feature	Reimagine Hub	Manual dispatcher	Menu chatbot	CRM manual tags
Classifies natural language messages	✓	✓	—	—
Classifies photos and voice transcripts	✓	✓	—	—
Response in seconds, 24/7	✓	—	✓	—
Automatic on-call alert for emergencies	✓	—	—	—
Links active customer to full job history	✓	~	—	~
Included in base plan at no extra charge	✓	—	—	—

✓ Has feature~ Partial / limited— Not available

Technical honesty

When AI triage needs a human review

Deco AI classifies with high confidence in most cases. A few situations require a dispatcher review.

Ambiguous messages with no context (e.g., 'hey are you guys available?') go to the dispatcher queue with the suggested category highlighted

A new number that uses trade-specific jargon consistent with a vendor or subcontractor may be misclassified as a lead on first contact

Voice transcripts from messages with heavy background noise or strong accents may score below 85% confidence and prompt a manual confirmation

What Hub does instead

Hub places the message in a 'needs review' queue with the AI's suggested category and confidence score visible. Dispatcher confirms or corrects with one tap. Each correction feeds back into Deco to improve future classifications.

Comes with the Hub

Owner, manager, front desk, team member, client. Each role sees only what it needs. Audit trail on every access, two-factor confirmation for sensitive actions.

View feature

Every feature above comes inside the Reimagine Hub on the same plan. See all Reimagine Hub features or get in touch if you are looking for something specific.

Pricing

Included in every Reimagine Hub plan

Automatic classification, on-call emergency alerts, 24/7 triage, and active-customer job history routing. Included in all plans starting at $397/month.

Book a demo Explore all Hub features

Glossary

Triage terms explained

Plain-English definitions for how AI triage works in practice.

Triage: The process of classifying inbound messages by type and urgency before routing them. In Hub, Deco AI does this automatically before the message appears in the dispatcher queue.
AI confidence score: The percentage certainty Deco AI assigns to its classification. Above 85%, the message routes automatically. Below 85%, it goes to the dispatcher with the suggested category highlighted for a one-tap confirmation.
On-call alert: An automatic notification sent to the designated on-call technician or manager when Deco AI classifies a message as an emergency. The alert fires immediately, regardless of business hours.
Support ticket: A record of an active-customer question or request. Different from a job: a ticket does not generate an invoice. It tracks the question, the response, and the outcome with full audit trail.

FAQ

Frequently asked questions about AI text triage

How does the AI know an emergency is real and not just an impatient customer?

Deco AI was trained on patterns from real emergency messages: 'burst', 'flooding', 'no power', 'smoke', 'gas smell', 'car stalled on highway', 'ceiling collapsed'. Subjective urgency (e.g., 'I need this ASAP') is classified differently from a genuine safety or damage risk. On-call alerts only fire above 85% confidence.

What if the AI misclassifies an active customer as a new lead?

Hub shows the confidence score to the dispatcher. If confidence is below 85%, the message goes to the review queue with the suggested category visible. The dispatcher corrects it with one tap and the correction improves Deco's accuracy going forward. After 30 days of live operation, error rate typically falls below 5%.

Does it work with photos and voice messages?

Yes. Photos are analyzed visually (e.g., a photo of a burst pipe or flooded floor is recognized as an emergency signal). Voice messages are transcribed and analyzed for urgency and context. Mixed messages (text plus photo) use both signals for higher accuracy.

Is there a monthly message limit?

No. Triage processes every inbound message with no volume cap. High-traffic periods do not increase cost.

Can I customize the four categories for my business?

Yes. The four default categories cover most service businesses. Companies with specific needs (e.g., separating residential from commercial estimates as distinct categories) can request customization during the onboarding week. Reimagine configures it at no additional charge on Pro and Elite plans.

What about vendor texts, spam, or wrong numbers?

Deco AI identifies vendor, spam, and wrong-number patterns and routes them to an 'other' queue automatically. Dispatcher reviews at their own pace without those messages polluting the customer queue. Numbers flagged as spam three times are blocked automatically.