This feature uses AI to analyze relevant customer and event data to dynamically set a field. It allows you to update custom fields based on the AI output.
🛑 Warning: This feature is experimental and you should assume that it will make mistakes.
Examples:
Not all of these ideas are great, but it will help you understand how it works so you can come up with solutions to your unique situations.
You can have AI help set your customer type automatically, or determine which surface type, or attempt to qualify new leads as spam or not. You can chain automations together to take further action based on the results of setting a custom field.
set customer type:
- create filter; name: 'recently created'; field: lead create date, last X hours, value 1
- create automation: name: 'set customer type'; filter to 'recently created'; action: 'set field', choose 'contact customer type', new value: 'Residential', check 'ai analyze'
- AI columns: name, phone, email, address, notes
- AI question: what customer type should I categorize this customer as? choices: Commercial, Residential, School
- activate and save. watch it for a while and makes sure it functions well.
hide spam with a 'bad fit' status:
- create an optional field; name: 'is spam' type: checkbox
- create filter; name: 'recently created' field: lead create date, last X hours, value 1
- create automation, 'filter spam', set filter to 'recently created', action: 'set field', choose 'is spam', new value: no, check 'AI analyze'
- AL columns: name, phone, email, notes
- AI question: is this spam?
- activate and save. watch it for a while and makes sure it functions well. If it does, you can create a 2nd automation, create 2nd filter on the 'is spam' field when it's 'yes' and then have the automation action be to 'set status' and mark it to some status like 'bad fit' or something
Prerequisites:
- Setup an optional field
- Proper integration with the AI processing engine as provided in your service plan (currently included free)
- Familiarity with your lead data and custom field types in the system.
Step-by-Step Instructions:
1. Create an automation, setup a filter, and choose action as 'set field' and choose a field to set. We suggest starting with a new optional field.
2. Enable the AI Analyze Option for Automations
• Check “Use AI to Analyze”.
• The 'new value' fields will be used as the backup/default if the AI fails to work properly
3. Configure AI Input Settings
• In the AI configuration section, select which columns to include as context. You can choose from fields such as name, address, email, etc. Only include ones relevant to the decision making.
• Enter instructions into the AI Question text box. For instance:
– “Is this a school? Output should be ‘yes’ or ‘no’ only. ”
• Provide the context data (e.g., organization, name, address, email).
4. Activate the Automation
• Review all settings and save the automation configuration.
• Activate the automation so new leads will be processed by the AI analyze step and the custom field updated automatically.
NOTE: Automations may require staff approval and on creation or modification.
Feature Options & Parameters:
- AI Analyze Text Box – Enter the sample instructions and training examples for the AI.
- Column Selector – Choose which lead data fields to include in the AI input.
- Output Format – Automatically enforced output based on the chosen field type.
- Fallback Value – Set a default value to be used if the AI output does not pass validation.
- Logging Details – The system logs output to track AI failures and default usage.
Use Cases:
- Automatically qualify leads from sources like cobalt by running an AI check—e.g., deciding if a lead belongs to an “inflatable bounce house” company (returns “yes”/“no”).
- Extract structured data such as phone numbers by setting up specific AI instructions to retrieve only the intended contact info.
Troubleshooting & FAQs:
Q: What happens if the AI does not return the expected “yes” or “no”?
A: While we validate the output to make sure it aligns with the field type being set, we don't guarantee it will follow your instructions exactly.
Q: Can I adjust the AI instructions after activation?
A: No, but you you edit the automation and add a new action, then delete the old action. Manual approval may be required after changing.
Q: What types of custom fields are supported?
A: The AI output is adaptable to multiple field types based on your configuration. Ensure your selected field type is reflected in the AI instructions.
Q: The AI does a poor job.
A: The AI can only make decisions based on the data provided to it and general knowledge it contains. If the data provided isn't clear or useful, it won't come up with the right answers. We suggest trying to automate easy stuff that is repetitive.