Skip to main content

​
Why Detailed Descriptions Matter

When a customer calls with a service request, the AI analyzes the conversation and matches it to the most appropriate Job Type and Business Unit in your system. The quality of this matching depends entirely on how well you’ve described each option.[1] Without detailed descriptions, the AI must guess based solely on the name of the job type or business unit. This often leads to:
  • Jobs booked under the wrong job type
  • Appointments assigned to incorrect business units
  • Customer frustration when the AI asks unnecessary clarifying questions
  • Reduced booking accuracy and increased manual corrections
With detailed descriptions, the AI can:
  • Accurately distinguish between similar job types
  • Select the correct business unit on the first try
  • Provide a smoother, more confident customer experience
  • Book jobs correctly into your CRM the first time[2]

​
How to Add Descriptions in Your Avoca Dashboard

​
Accessing Business Units & Job Types

For Business Units:
1
Navigate to https://app.avoca.ai/team/TEAM_ID/st-integrations?tab=business-units
2
Click “Get unsynced Business Units” to see all available units
3
Click “Add” for any you want Avoca to consider for booking
4
Click “Configure” to open the settings dropdown[1]
For Job Types:
1
Navigate to https://app.avoca.ai/team/TEAM_ID/st-integrations?tab=job-types
2
Click “Get unsynced Job Types” to see all available types
3
Click “Add” for any you want Avoca to book
4
Click “Configure” to open the settings dropdown[3]

​
Writing Effective Descriptions

Focus on what distinguishes each option from others, not on writing a complete definition.[4] Good examples:
  • ✅ “Jobs that require a repair for a plumbing issue on systems that are 10 or more years old”
  • ✅ “Book this job type only if the call is related to well systems and current time is between 5 PM to 7 AM”
  • ✅ “Responsible specifically for moving water out of the building - drain pipes, wastewater removal, etc”
Poor examples:
  • ❌ “Plumbing jobs” (too vague)
  • ❌ “HVAC Service” (doesn’t distinguish from HVAC Install or HVAC Maintenance)

​
Using AI to Generate Descriptions

You can use ChatGPT to help create descriptions:
1
Go to the Retool Admin dashboard: https://avoca.retool.com/app/full-ai-admin > Onboarding > ST IDs
2
Select your team and download the CSV
3
Paste the CSV into ChatGPT along with a standard list (example: Turnpoint Master Job List)
4
Ask ChatGPT to create new definitions that focus on distinguishing features[3]

​
Understanding the Toggles

Each Business Unit and Job Type has two critical toggles that control AI behavior:

​
Booking Toggle

Controls: Whether the AI is allowed to book appointments into this business unit or job type
  • Enabled (ON): AI can book jobs under this option
  • Disabled (OFF): AI will take a message or escalate to a human instead of booking
When to disable: If you want to prevent the AI from booking certain types of work (e.g., installs, commercial jobs, or deprecated job types)[1]

​
Rescheduling Toggle

Controls: Whether the AI is allowed to reschedule existing appointments for this business unit or job type
  • Enabled (ON): AI can reschedule appointments
  • Disabled (OFF): AI will take a message or escalate to a human for rescheduling
When to disable: If certain job types require human approval before rescheduling (e.g., install appointments, warranty work)[1]
Important:After making any changes to toggles or descriptions, always click the blue “Save” button at the bottom of the page. Changes will not take effect until saved.

​
Frequently Asked Questions

Recommendation: Keep it to 10-30 active job types that Avoca can book.The more job types the AI has to choose from, the higher the likelihood of occasional misclassifications. We’ve seen clients with 44+ job types experience more booking errors than those with streamlined lists.
Review and remove any job types you don’t actively want Avoca booking into. The fewer options, the more accurate the AI will be.
You can include time conditions directly in your descriptions!As of recent updates, the AI now receives current time information when determining job types. This means you can write descriptions like:
“Book this job type only if the call is related to HVAC on-call emergencies and current time is between 9 PM to 7 AM”
This is particularly useful for after-hours or emergency job types that should only be selected during specific windows.
The AI uses descriptions to match the customer’s stated problem to the right business unit.Scenario: Customer calls about a well emergency after hours, and you want it booked under “Plumbing On-Call” business unit.Solution:
1
Add well-related keywords in the description of the “Plumbing On-Call” business unit.
2
Add the same keywords in the description of relevant job types.
3
In Booking Windows, add a service type for “Well Repairs/Services” associated with the Plumbing business unit.
4
Add “Well repair/service” under services provided in the human-onboarding dashboard.
The AI selects job types and business units based on your descriptions and booking window mappings. If a job type is only valid for certain business units, make sure:
1
Your business unit mappings in Booking Windows are configured correctly.
2
The descriptions clearly indicate which business unit(s) each job type belongs to.
3
The job type’s business unit associations in ServiceTitan are up to date.
Common issue: When business unit mappings in Booking Windows don’t match the job type’s allowed business units, capacity queries may fail or return incorrect availability.
No particular format is required, but descriptions should be clear and focused. Avoid:
  • Writing essays or overly long descriptions
  • Including irrelevant information
  • Duplicating information that’s already obvious from the name
Keep it simple: One to two sentences that highlight the distinguishing features is usually sufficient.
Not necessarily. If your job type or business unit name is very clear (e.g., “HVAC-Maintenance-Residential-Charlotte”), a description may not be needed. Focus your effort on items that could be confusing or overlapping.

​
Important Tips & Best Practices

​
Tip #1: Focus on Differences, Not Definitions

Don’t describe what plumbing or HVAC is in general. Instead, explain what makes this specific job type or business unit different from your other options.[4] Example: Instead of “Plumbing services for residential customers,” write “Plumbing work for systems less than 10 years old (use Plumbing-Demand-10+ for older systems).”

​
Tip #2: Use Keywords from Customer Language

Think about how customers describe their problems on the phone. Include those exact words in your descriptions:[7]
  • “No heat” or “No cool” for emergency HVAC
  • “Clogged drain” or “backed up” for drain services
  • “Won’t open” or “off track” for garage doors

​
Tip #3: Include Equipment Age Distinctions

If you have different job types based on equipment age (e.g., 10+ years vs. newer), explicitly state the age threshold in the description. The AI can ask customers about equipment age and use that to select the right job type.[3]

​
Tip #4: Reduce Overlap Between Job Types

If you’re experiencing frequent misclassifications, audit your job types for overlap:[5]
  • Do you have multiple job types that serve the same purpose?
  • Can any be merged or consolidated?
  • Are there deprecated job types that should have booking disabled?

​
Tip #5: Test After Making Changes

After updating descriptions:
1
Make test calls with realistic customer scenarios.
2
Check that jobs are booking into the correct job types and business units.
3
Review the AI’s confidence scores in the call debugger.
4
Iterate on descriptions that still cause confusion.

​
Tip #6: Coordinate with Booking Windows

Your descriptions work together with your Booking Windows configuration. Make sure:[7]
  • Service types in Booking Windows align with your job type descriptions
  • Business unit mappings are correct for each service/appointment type combination
  • Emergency qualifications match the language in your emergency job type descriptions

​
Tools for Managing Job Types & Business Units

​
Avoca Dashboard

​
Retool Admin Dashboard

https://avoca.retool.com/app/full-ai-admin
  • Export job type lists as CSV
  • View team configurations
  • Access onboarding tools[3]

​
Job Type Onboarding Template

Use this template to collect job type and business unit information from customers: Job Types, BU, Tags, and Zips Template[4]

​
Call Debugger

Review how the AI selected business units and job types for specific calls:
  • View confidence scores for each selection
  • See which descriptions the AI referenced
  • Identify patterns in misclassifications[10]

​
Troubleshooting Common Issues

​
Issue: AI is selecting the wrong job type consistently

Solutions:
1
Review the description of the correct job type and add more specific keywords.
2
Review the description of the incorrect job type and remove overlapping language.
3
Reduce the total number of active job types if possible.
4
Check if you need to add time-based conditions to the descriptions.

​
Issue: Customers are frustrated by clarifying questions

If the AI is asking “Just to confirm, this is for [job type]?” too often, it usually means:[2]
  • Multiple job types have very similar descriptions
  • The descriptions don’t include the specific keywords customers are using
  • The total number of job types is too high
Fix: Make descriptions more distinct and focused on customer language.

​
Issue: Jobs booking into wrong business unit

This often happens when:[8]
  • Business unit descriptions are too vague
  • Booking window business unit mappings don’t match job type associations
  • The job type and business unit were selected independently without validation
Fix: Ensure job types are clearly associated with their correct business units in both descriptions and ServiceTitan configuration.

​
Additional Resources

  • Responder Onboarding Guide: For complete setup instructions[3]
  • Booking Windows Guide: To configure scheduling logic[11]
  • Business Information Configuration: How to set up company details[12]
Last modified on February 19, 2026