Categorise SMS With AI

Automatically categorise and classify SMS messages using artificial intelligence. Perfect for Australian businesses needing smart message routing, support ticket categorisation, and intent detection.

Action•Pro & Enterprise Plans

What Does This Action Do?

The Categorise SMS With AI action uses advanced artificial intelligence to automatically classify incoming SMS messages into categories, topics, intents, and urgency levels. Perfect for Australian businesses that need to route customer messages to the right team, prioritise urgent requests, or automate support workflows based on message content.

Popular Australian Use Cases: Support ticket routing, product enquiry classification, appointment request categorisation, lead qualification, complaint detection, delivery issue sorting, and department-based message routing.

How It Works

1

SMS Received as Trigger

Use the "New SMS Received" trigger to capture incoming customer messages. Each SMS is passed to the AI categorisation action.

2

AI Analyses Message Content

DataFlows AI reads the SMS text and intelligently categorises it based on your custom category definitions or our pre-built templates.

3

Returns Category & Metadata

Action outputs: Primary Category, Subcategory, Intent, Urgency Level, Confidence Score, and Suggested Action. Use these in subsequent Zap steps.

4

Route & Automate Based on Category

Use Zapier Paths/Filters to route messages: High urgency → Call manager, Product enquiry → Sales CRM, Complaint → Support ticket, etc.

AI Categorisation Output Fields

Primary Category

The main classification of the message. Examples: "Support Request", "Product Enquiry", "Complaint", "Appointment Request", "Feedback", "Sales Lead".

Subcategory

More specific classification within primary category. Examples: "Technical Issue - Internet", "Billing Question", "Delivery Concern", "Urgent Medical".

Intent

Customer's underlying intention. Examples: "Purchase", "Get Information", "Report Problem", "Request Callback", "Cancel Service", "Provide Feedback".

Urgency Level

Assessed priority: "Critical", "High", "Medium", "Low". Helps prioritise response. Critical = immediate action needed, Low = can wait 24+ hours.

Confidence Score

AI's certainty in the categorisation (0.0 to 1.0). Higher = more confident. Use to filter: only auto-route if confidence > 0.8, otherwise escalate to human.

Suggested Action

AI's recommendation for next step. Examples: "Create support ticket", "Assign to sales team", "Schedule callback", "Offer refund", "Escalate to manager".

Detected Topics

Key topics/keywords identified. Examples: "Billing, Late Payment", "Product: Nike Air Max, Size 10", "Complaint, Rude Staff, Refund Request". Useful for tagging.

Department

Recommended department for routing. Examples: "Technical Support", "Sales", "Billing", "Customer Success", "Warehouse", "HR". Based on your organisation structure.

Action Configuration Options

Message Text Required

The SMS message content to categorise. Typically mapped from "New SMS Received" trigger: {{ Message Text }}.

Category Template Optional

Choose from pre-built templates: "Customer Support", "Retail Enquiries", "Healthcare", "Real Estate", "General Business". Or create custom categories.

Custom Categories Optional

Define your own category list (comma-separated): "Technical Support, Billing Question, Sales Enquiry, Complaint, Feedback". AI will classify into these.

Include Sentiment Optional

Also run sentiment analysis alongside categorisation. Returns: Positive/Negative/Neutral + Emotion. Useful for complaint detection.

Language Optional

Message language. Defaults to "English (Australian)". Also supports: English (US/UK), Mandarin, Vietnamese, Arabic, Spanish. AI adjusts for dialects.

Real-World Australian Use Cases

Here are 12 proven ways Australian businesses use AI categorisation to automate message routing and improve response times:

Automated Customer Support Ticket Routing

Customer Support

Automatically categorise incoming Australian customer SMS messages and route to the correct department or support team.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Customer sends SMS: 'My internet has been down for 3 hours!'
  2. AI categorises message → Category: Technical Support, Subcategory: Internet Outage, Urgency: High
  3. Create Zendesk ticket in 'Technical Support' queue with High priority
  4. Assign to on-call technical team member
  5. Send Slack alert to #tech-support channel
  6. Auto-reply SMS: 'Technical support notified. We'll call you within 15 minutes.'

Product Enquiry Classification for Retail

Retail & E-commerce

Classify Australian customer product enquiries by category, intent, and priority for better sales response.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Customer sends SMS: 'Do you have the Nike Air Max in size 10?'
  2. AI categorises → Category: Product Enquiry, Product Type: Footwear, Brand: Nike, Intent: Purchase
  3. Check inventory system for Nike Air Max size 10
  4. If in stock: SMS with availability and store location
  5. If out of stock: SMS with alternatives and restock date
  6. Log enquiry in Salesforce with category tags

Medical Appointment Request Categorisation

Healthcare

Automatically categorise patient SMS requests by urgency, department, and appointment type for Australian healthcare practices.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Patient sends SMS: 'I need to see Dr. Smith about my test results'
  2. AI categorises → Category: Appointment Request, Doctor: Dr. Smith, Type: Follow-up, Urgency: Medium
  3. Check Dr. Smith's calendar for next available follow-up slot
  4. Send SMS: 'Dr. Smith available Thursday 2:30 PM or Friday 10 AM. Reply 1 for Thu or 2 for Fri.'
  5. Patient reply triggers booking confirmation
  6. Add to practice management system with category tags

Real Estate Enquiry Segmentation

Real Estate

Categorise Australian property enquiries by type, budget range, and urgency to prioritise agent follow-up.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Prospect sends SMS: 'Interested in 123 Smith St. Looking for 3BR under $1M in Sydney'
  2. AI categorises → Category: Property Enquiry, Property: 123 Smith St, Bedrooms: 3, Budget: Under $1M, Location: Sydney, Intent: Serious Buyer
  3. Assign lead score based on category (3BR + $1M budget = High value)
  4. Alert agent via SMS with prospect details and category tags
  5. Add to CRM with tags: Budget: $1M, Location: Sydney, Urgency: High
  6. Agent calls prospect within 5 minutes

Event Registration Category Routing

Events & Conferences

Categorise event-related SMS enquiries to route to tickets, venue, catering, or logistics teams.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Attendee sends SMS: 'I have a dietary requirement - gluten free. How do I update my registration?'
  2. AI categorises → Category: Event Enquiry, Topic: Dietary Requirements, Department: Catering, Urgency: Medium
  3. Create task for catering team in Asana
  4. Send SMS: 'Thanks! Our catering team will email you a dietary form within the hour.'
  5. Notify [email protected] with attendee details
  6. Log in event management system under 'Catering Requests'

Restaurant Reservation & Enquiry Categorisation

Hospitality

Automatically categorise Australian restaurant SMS messages by type: bookings, menu enquiries, feedback, or complaints.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Customer sends SMS: 'Can I book a table for 6 people Saturday night?'
  2. AI categorises → Category: Reservation Request, Party Size: 6, Date: Saturday, Intent: Book
  3. Check OpenTable for Saturday availability
  4. Send SMS: 'Available at 6:30 PM or 8:00 PM Saturday. Reply 1 for 6:30 or 2 for 8:00.'
  5. Customer reply triggers automatic booking creation
  6. Confirmation SMS sent with booking details

Feedback & Complaint Categorisation System

Customer Experience

Categorise Australian customer feedback by type, sentiment, and department for better response management.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Customer sends SMS: 'Your delivery driver was rude and my order was cold!'
  2. AI categorises → Category: Complaint, Topic: Delivery Service + Food Quality, Sentiment: Negative, Urgency: High
  3. Create priority support ticket in Zendesk
  4. Alert customer success manager via Slack
  5. Auto-reply SMS: 'We're very sorry! Manager will call you within 10 minutes.'
  6. Log in complaints dashboard for trend analysis

Lead Source & Intent Classification

Sales & Marketing

Categorise inbound Australian sales enquiries by product interest, budget, and purchase intent for better lead qualification.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Lead sends SMS: 'I saw your ad for enterprise software. Need solution for 50+ users. Budget approved.'
  2. AI categorises → Category: Sales Enquiry, Product: Enterprise Software, Company Size: 50+ users, Budget Status: Approved, Intent: High
  3. Calculate lead score: Enterprise (5 pts) + 50 users (4 pts) + Budget approved (5 pts) = 14/15 (Hot Lead)
  4. Immediately assign to senior sales rep
  5. Alert via phone call + SMS to sales rep
  6. Create Salesforce opportunity with category: Hot Lead - Enterprise

Delivery Issue Categorisation for Logistics

Logistics & Delivery

Automatically categorise delivery-related SMS messages by issue type, urgency, and required action.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Customer sends SMS: 'Package not delivered today. Tracking says delivered but I never received it.'
  2. AI categorises → Category: Delivery Issue, Type: Missing Package, Urgency: High, Action Required: Investigation
  3. Create urgent case in delivery management system
  4. Alert local depot manager and driver
  5. Send SMS: 'Investigating immediately. Depot manager will call within 30 minutes.'
  6. Escalate to senior ops if not resolved in 2 hours

Financial Enquiry Categorisation for Banking

Banking & Finance

Categorise Australian customer banking SMS enquiries by type, account, and urgency for compliance and routing.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Customer sends SMS: 'I see a $500 charge I don't recognise on my credit card'
  2. AI categorises → Category: Fraud Concern, Account Type: Credit Card, Amount: $500, Urgency: Critical
  3. Immediately flag account for fraud review
  4. Create high-priority case in fraud prevention system
  5. Send SMS: 'Security team notified. Freezing card now. We'll call you in 5 minutes.'
  6. Alert fraud team via phone + email with category: Potential Fraud - Credit Card

IT Help Desk Ticket Auto-Categorisation

IT & Managed Services

Automatically categorise IT support requests from Australian clients by issue type, severity, and SLA priority.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Client sends SMS: 'Email server down for entire Sydney office - 40 staff affected'
  2. AI categorises → Category: IT Issue, Type: Email Server, Severity: Critical, Affected Users: 40, Location: Sydney Office
  3. Create P1 (Priority 1) ticket in ServiceNow
  4. Alert on-call engineer via phone + SMS
  5. Notify client: 'P1 ticket created. Engineer John responding now. ETA: 10 minutes.'
  6. Escalate to senior engineer if not resolved in 30 minutes per SLA

Educational Enquiry Categorisation for Schools

Education

Categorise parent/student enquiries for Australian schools by topic, grade level, and department.

Trigger: DataFlows SMS - New SMS Received

Workflow Breakdown:

  1. Parent sends SMS: 'My daughter is sick today. Year 7. Won't be at school.'
  2. AI categorises → Category: Absence Notification, Student: Year 7, Reason: Illness, Department: Attendance Office
  3. Log absence in school management system
  4. Notify Year 7 homeroom teacher
  5. Send SMS confirmation: 'Absence recorded. Hope she feels better soon!'
  6. Add to attendance report for the day

Quick Setup Guide

Step 1: Set Up New SMS Received Trigger

  1. Create new Zap in Zapier
  2. Search for "DataFlows SMS" and select it
  3. Choose trigger: "New SMS Received"
  4. Connect your DataFlows account
  5. Test trigger to get sample SMS data

Step 2: Add Categorise SMS With AI Action

  1. Click "+" to add action step
  2. Search for "DataFlows SMS" and select it
  3. Choose action: "Categorise SMS With AI"
  4. Your DataFlows account should already be connected
  5. Click "Continue"

Step 3: Configure AI Categorisation

  1. Message Text: Map the message from trigger: {{ Message Text }}
  2. Category Template: Choose pre-built template or select "Custom"
  3. If using custom categories, enter your list:
    Technical Support, Billing Question, Sales Enquiry, Product Question, Complaint, General Feedback
  4. Include Sentiment: Toggle ON if you want emotion detection too
  5. Language: Select "English (Australian)" or leave default
  6. Click "Continue"

Step 4: Test AI Categorisation

  1. Click "Test action" to run AI on your sample SMS
  2. Review the output fields:
    • Primary Category (e.g., "Support Request")
    • Subcategory (e.g., "Technical Issue - Internet")
    • Intent (e.g., "Report Problem")
    • Urgency Level (e.g., "High")
    • Confidence Score (e.g., 0.91)
    • Suggested Action (e.g., "Create urgent support ticket")
  3. Verify the AI categorised correctly. If not, refine your custom categories.

Step 5: Add Routing Logic Based on Category

  1. Click "+" to add next step
  2. Choose "Paths by Zapier" for conditional routing based on category
  3. Path A: High Urgency Messages
    • Condition: Urgency Level = "Critical" OR "High"
    • Actions: Create urgent ticket + Alert manager via phone call + SMS customer
  4. Path B: Sales Enquiries
    • Condition: Primary Category = "Sales Lead" OR "Product Enquiry"
    • Actions: Create Salesforce lead + Assign to sales rep + SMS confirmation
  5. Path C: Complaints
    • Condition: Primary Category = "Complaint" OR Sentiment = "Negative"
    • Actions: Escalate to customer success + Offer apology + Priority follow-up
  6. Path D: General Enquiries
    • Condition: Everything else (Urgency = "Low" or "Medium")
    • Actions: Create standard ticket + SMS with expected response time

Step 6: Activate & Monitor

  1. Turn on your Zap
  2. Monitor Zap history to see how messages are being categorised
  3. Check categorisation accuracy after first 50-100 messages
  4. Refine custom categories if needed based on real data
  5. Track metrics: Average confidence score, category distribution, routing effectiveness

Best Practices for Australian Businesses

✅ Start With Pre-Built Templates

Use DataFlows' pre-built category templates for your industry first. They're trained on thousands of real messages. Customise later if needed.

✅ Keep Custom Categories Clear & Distinct

Define 5-10 categories maximum with clear differences. Avoid overlap: "Technical Support" and "IT Issues" are too similar. AI works best with distinct categories.

✅ Use Confidence Score for Human Escalation

Only auto-route messages with confidence >= 0.75. Lower confidence? Send to human for manual categorisation. Prevents mis-routing critical messages.

✅ Combine with Sentiment Analysis

Enable "Include Sentiment" to catch angry customers even if category is neutral. Example: "Product Enquiry" + "Negative Sentiment" = Unhappy customer, prioritise!

✅ Route Critical Urgency to Humans Immediately

When Urgency = "Critical", bypass automation and alert humans via phone/SMS. Examples: "Server down", "Medical emergency", "Security breach".

✅ Log All Categorisations for Analysis

Send category data to Google Sheets or Airtable: Message, Category, Confidence, Timestamp. Helps identify trends: "50% of messages are billing questions - need FAQ".

✅ Test with Real Australian Language

AI is trained on Australian English ("mobile" not "cell", "organisation" not "organization"). Test with real Aussie customer messages for accuracy.

✅ Use Suggested Action Field

Don't ignore this! AI recommends next steps based on message content. Use in Paths: If Suggested Action = "Escalate to manager", trigger manager alert.

✅ Review & Refine Categories Monthly

After 30 days, analyse category distribution. If 80% fall into "Other" or "General", your categories are too specific. Adjust based on real data patterns.

✅ Handle Multi-Topic Messages

Some messages cover multiple topics: "I want to buy your product but I'm also unhappy with previous purchase". Check Detected Topics field for all themes.

Advanced Automation Tips

Multi-Tier Routing Based on Category + Sentiment

Create sophisticated routing logic combining multiple AI outputs:

  • Tier 1 (Immediate): Urgency = Critical OR (Category = Complaint AND Sentiment = Negative)
  • Tier 2 (2 hour SLA): Urgency = High OR Category = Sales Lead
  • Tier 3 (24 hour SLA): Urgency = Medium AND Sentiment = Neutral
  • Tier 4 (48 hour SLA): Urgency = Low AND Category = General Enquiry

Dynamic CRM Field Population

Use category data to auto-populate CRM fields:

  • Primary Category → Salesforce Lead Source
  • Detected Topics → HubSpot Contact Tags
  • Urgency Level → Support Ticket Priority
  • Intent → Pipeline Stage (Purchase Intent = "Hot Lead" stage)

Category-Based Auto-Responses

Send different SMS replies based on category:

  • Support Request → "Ticket #1234 created. Response in 2 hours."
  • Sales Enquiry → "Thanks! Sales rep will call within 30 mins."
  • Billing Question → "Billing team notified. Reply in 24 hours."
  • Complaint → "We're very sorry! Manager will call you within 15 minutes."

AI Training Feedback Loop

Improve AI accuracy over time:

  • Log all categorisations to Google Sheets with columns: Message, AI Category, Human Review, Correct?
  • Staff review and mark incorrect categorisations
  • Weekly: Identify patterns in mis-categorisations
  • Adjust custom category definitions based on errors
  • Re-test with historical messages to verify improvement

Department-Specific Zaps

Create separate Zaps for each department based on category routing:

  • Sales Zap: Filter for Category = Sales Lead → Salesforce → Assign rep → Call within 5 mins
  • Support Zap: Filter for Category = Support → Zendesk → Route by subcategory → SLA timer
  • Billing Zap: Filter for Category = Billing → Xero lookup → Send account statement SMS

Compliance & Audit Trail

For regulated industries (healthcare, finance):

  • Log every message + category + confidence to immutable database
  • Include timestamp, staff assigned, response time
  • Use Category = "Compliance Risk" to trigger legal review
  • Monthly reports: Category trends, response times by urgency, escalation rates

Common Issues & Solutions

❌ Low Confidence Scores (< 0.5)

Solutions:

  • Your custom categories may be too similar or ambiguous. Make them more distinct.
  • Messages are too short or vague ("Hi", "Thanks") - add filter to require minimum 10 characters
  • Switch to pre-built template instead of custom categories
  • Route low-confidence messages to human review instead of auto-processing

❌ AI Choosing Wrong Category Consistently

Solutions:

  • Review your category definitions - are they clear and distinct?
  • Reduce number of categories (aim for 5-8 max)
  • Provide more context: Include previous message history if available
  • Use pre-built industry template which has better training data

❌ Most Messages Categorised as "General" or "Other"

Solution: Your categories are too specific. Broaden them. Instead of "Technical Issue - Internet - Router Not Working", use "Technical Support". Let subcategory handle specifics.

❌ Non-English Messages Not Categorising Well

Solution: Set the Language field to match your messages (Mandarin, Vietnamese, Arabic, etc.). AI accuracy drops significantly if language mismatch.

❌ Urgency Always Returns "Medium"

Note: AI assesses urgency based on message content keywords ("urgent", "asap", "emergency", "broken", "not working"). If customers don't use urgency language, most will be "Medium". Consider using sentiment as additional urgency signal.

Related DataFlows Features

Analyse SMS Sentiment With AI

Detect customer emotions (happy, angry, frustrated) in SMS messages. Perfect for complaint detection and customer experience monitoring.

Action

New SMS Received

Trigger your Zaps when customers send SMS messages. The starting point for AI categorisation workflows.

Trigger

Send SMS

Send automated SMS responses based on message category. Perfect for category-specific auto-replies.

Action

Ready to Automate Message Categorisation With AI?

Join Australian businesses using DataFlows AI to automatically route, prioritise, and categorise thousands of SMS messages every day.

Get Started with ZapierView All Features