Tool Configuration allows your Serova assistant to take action, not just reply with text.
Tools give your assistant the ability to fetch data, schedule tasks, trigger workflows, send structured messages, and automate processes.
Each tool type can be customised into tool instances — named, configured versions of a tool that your assistant can call when needed.
This page explains:
Tools are actions your assistant can perform.
They move the assistant beyond “question and answer” and allow it to carry out real tasks.
Tools enable your assistant to:
A tool instance is a personalised, named version of a tool.
Tool instances are named, configured versions of base tools that your assistant can call directly.
For example, instead of having a generic “send message” tool, you might create:
Each instance has its own configuration (channels, templates, settings), and the AI knows exactly when to use each one once you define this in prompts.
Additional example tool instances:
pricing_lookupavailability_checkmaintenance_logmember_verificationEach instance includes:
Your assistant calls tool instances only when instructed in prompts — tools do not work automatically.
Serova currently offers three tool types:
Retrieve information from an external REST API.
Use Get Data when the assistant needs to:
Examples of Get Data tool instances:
fetch_pricingcheck_availabilitylookup_membercheck_order_statusSchedule tasks or reminders.
Use Scheduler when your assistant needs to:
Examples:
booking_reminderlead_followuprenewal_notificationSend a structured, template-based message.
Use Send Response for:
Examples:
support_acknowledgementproduct_info_messagemaintenance_confirmationadmin_alertCreating a tool does not enable it automatically.
You must tell your assistant when to use each tool by writing clear instructions in your prompts.
Example prompt instructions:
If a user asks about pricing, call the `fetch_pricing` tool with the product or service name.
If the user asks whether a facility or item is available, call the `check_availability` tool.
If the user reports a problem or needs support, call the `support_request` tool before responding.
This ensures the assistant triggers the correct tool at the right time.
Use tools to automate:
If you only need the assistant to answer questions with static information, tools may not be required.
Follow these steps:
After linking tools to prompts, test to ensure the assistant behaves correctly.
For advanced configurations, use Serova’s public documentation:
🔗 Prompt Structure Guide
prompt-structure
These guides explain:
If you prefer Serova to assist with tool creation or prompt configuration:
We can help with: