How to Build Your First Custom Thing

Introduction

While our marketplace offers many pre-built Things for common tasks, creating a custom Thing tailored to your unique requirements can provide even greater value. This guide will walk you through the process of building your first custom Thing on our Morrow platform.

Before You Begin

Before creating a custom Thing, you should:

• Clearly define the problem you want to solve
• Identify the inputs and outputs for your Thing
• Gather sample data for training and testing
• Consider the systems your Thing will need to integrate with

Step 1: Define Your Thing's Purpose

Start by creating a clear mission statement for your Thing. For example:

• "This Thing will automatically process customer support tickets, categorize them by urgency and department, and route them to the appropriate team."
• "This Thing will monitor our social media mentions, analyze sentiment, and alert our PR team to potential issues."

The more specific you can be about your Thing's purpose, the better.

Step 2: Choose Your Thing Type

Our platform offers several frameworks to build upon:

• Processing Things: Best for data transformation and analysis
• Communication Things: Ideal for customer interactions
• Monitoring Things: Perfect for ongoing surveillance of systems or data
• Decision Things: Designed for making rule-based choices
• Creative Things: Specialized in generating content or ideas

Select the type that best matches your primary use case.

Step 3: Configure Input Sources

Your Thing needs data to work with. Configure where it will get information from:

• API connections to your existing systems
• Email or messaging platforms
• Document repositories
• Database access
• Web scraping capabilities
• User inputs via forms or chat interfaces

For each input source, you'll need to define authentication methods and data formats.

Step 4: Design Processing Logic

This is where you teach your Thing what to do with the information it receives. Depending on your Thing type, you might:

• Create classification rules for categorizing information
• Design decision trees for handling different scenarios
• Set up pattern recognition parameters
• Establish prioritization frameworks
• Build response templates for different situations

Our visual workflow builder makes this process intuitive, even for users without programming experience.

Step 5: Configure Outputs and Actions

Define what your Thing will do after processing information:

• Send notifications to team members
• Update records in your systems
• Generate reports or documents
• Reply to messages
• Trigger other workflows or processes
• Escalate to human review when needed

For each action, specify the conditions that trigger it and how the output should be formatted.

Step 6: Test and Refine

Before deploying your Thing:

1. Run simulations with historical data
2. Test edge cases and unusual scenarios
3. Have team members review the Thing's decisions
4. Make adjustments to improve accuracy
5. Document the Thing's capabilities and limitations

Plan to continuously monitor and refine your Thing after deployment.

Step 7: Deploy and Scale

Once your Thing is performing well in tests:

1. Roll out to a limited audience first
2. Gather feedback from early users
3. Make necessary adjustments
4. Gradually expand deployment
5. Set up monitoring dashboards
6. Establish maintenance procedures

Real-World Example: Sales Lead Qualification Thing

Let's look at how this process might work for creating a sales lead qualification Thing:

1. Purpose: Automatically qualify incoming sales leads and route them to the appropriate sales representative.

2. Thing Type: Decision Thing

3. Input Sources:

   • Website contact form submissions
   • Email inquiries
   • CRM system integration
   • LinkedIn Lead Gen forms

4. Processing Logic:

   • Score leads based on company size, industry, budget
   • Categorize by product interest
   • Check for existing relationships in CRM
   • Detect urgency based on language

5. Outputs and Actions:

   • High-value leads: Immediate notification to sales team
   • Medium-value leads: Add to nurture sequence
   • Low-value leads: Send automated resources
   • Update CRM with lead score and notes
   • Schedule follow-up tasks

6. Testing: Run against 100 recent leads to compare Thing decisions with actual sales outcomes.

7. Deployment: Roll out to handle 25% of new leads, then gradually increase as performance is verified.

Conclusion

Building a custom Thing is an iterative process that combines your domain expertise with our platform's capabilities. While it requires more initial setup than using a pre-built Thing, the return on investment can be substantial when the Thing is perfectly aligned with your specific business needs.

Ready to build your custom Thing? Start your project on our Create Thing page, or reach out to our solutions team for guidance on complex implementations.