.png){kind=logo}
Customizing Claybot to Categorize Industries According to Your Definitions
Is it possible to use the Claybot to categorize the industry based on our own set of definition?
Is it possible to use the Claybot to categorize the industry based on our own set of definition?
Yes, you can use Clay to categorize industries based on your own custom definitions. There are two main approaches: 1. Using the Claygent action, you can create custom industry categories and automatically classify prospects according to your unique definitions. You can list as many categories as needed - from simple classifications like B2B vs. B2C to detailed lists with 50+ categories. 2. Alternatively, you can use the "Categorize Items Based On Keywords" feature, which allows you to define categories and their corresponding keywords. The system will then match items to appropriate categories based on keyword matching. Both methods help overcome the limitations of standard industry classifications by letting you create segments that are meaningful for your specific business needs.
For example the categoriztion should follow the description in my pic below. And is there any complication if mylist is super long than this?
Yes, you can use Clay to categorize industries based on your own detailed descriptions. There are two main approaches: 1. Using the Claygent action, you can list as many custom industry categories as needed and instruct Claygent to classify prospects according to your unique definitions. This works for both simple classifications and detailed lists with 50+ categories. 2. Alternatively, use the "Categorize Items Based On Keywords" feature, where you define categories and their corresponding keywords. The system matches items to categories based on keyword matching. For very long lists, there are no specific complications mentioned in the documentation. Clay is designed to handle extensive categorization needs, from simple B2B vs. B2C classifications to detailed industry segmentation.
so basically i just need to write very long prompts inside my clai ai agent right?
Yes, you can write detailed prompts in your Clay AI agent to categorize industries based on your own definitions, even with very long lists. Using the Claygent action, you can list as many custom industry categories as needed and instruct it to classify prospects according to your unique definitions. This works well for both simple classifications and detailed lists with 50+ categories. Clay's AI capabilities are specifically designed to help categorize information into buckets, allowing you to segment leads into various categories (like B2B vs. B2C companies, or categories by industry size). There's no mentioned limitation on prompt length, so you can include your comprehensive industry definitions in the prompt to ensure accurate categorization.
Yes, you can use the AI Prompts to get the desired results
Cool, is the prompt as simply as the screenshot above? I just say categorize it as A, when you found the website (or other inputs) that fits the description A, etc etc?
Yes, you can create a prompt that simply describes each industry category and instructs the AI to categorize based on those descriptions. For best results, structure your prompt using the S.P.I.C.E. framework: ** Sections: Organize your prompt with clear sections (#VARIABLES#, #CONTEXT#, #INSTRUCTIONS#, #EXAMPLES#) ** Instructions: Clearly explain how to categorize items based on your descriptions ** *Context: Provide background on why you're categorizing industries this way When setting up your categorization, you have two main options: 1. Use the "Categorize Items Based On Keywords" feature where you define categories and their corresponding keywords 2. Create a custom AI prompt that includes your detailed category descriptions For very long lists of categories, both approaches can handle extensive categorization needs without complications.
You need to craft the Propmpt and also provide an example in the Prompt as well. Run it for one row and check if you are getting the desired results. You can tweak and twist on one rwo before running on all rows
To craft an effective AI prompt for industry categorization in Clay: 1. Start with a system prompt that provides context and outlines your goal for categorization. 2. Use the Claygent action to list your custom industry categories and instruct it to classify prospects according to your unique definitions. 3. Structure your prompt with: • Clear step-by-step instructions • Specific output format requirements • Your detailed industry definitions 4. Include examples of how to categorize different industries based on your definitions to help the AI understand your criteria. 5. To test on a single row first: • Configure your AI prompt • Run it on just one row to check results • Refine the prompt as needed before applying to all rows For help creating your prompt, click the "Help me" button in the Configuration section of the "Use AI" panel and describe your desired outcome.
Got it. To follow-up slightly, it's possible to ask the prompt to classify the industry based on the website domain right?
Yes, you can use AI prompts in Clay to classify industries based on website domains. You can create custom prompts that analyze domain information to categorize companies according to your specific industry definitions. Clay offers several approaches for industry classification: 1. Using AI prompts to extract and categorize industry information from website domains 2. Creating specialized prompts that can identify specific types of companies (like SaaS) based on their web presence Clay also provides standard industry classification data points that can be enriched from domains, including: ** Company Industry ** Company Industry - NAICS ** Company Industry - SIC ** Company Industry - Sector You can customize your classification criteria to match your specific needs rather than relying solely on standard industry categories.
Sources: We Made 24 Prompts So You Don’t Have To [Automated Outbound] - Clay University Data points: AI Prompt | Clay.com Data points: Domain - Website | Clay.com 11 AI Prompts to Automate Prospect Research with Claygent [Automated Outbound] - Clay University Data points: Company Industry - SIC | Clay.com
can i feed website, company linkedin url, company name and ask the prompt to classify industry from there? without scraping the website text or company about us text from linkedin? can the ai read the website data and linkedin url? Or i need to explicitly feed the data into the prompt?
Yes, you can classify industries in Clay by providing just the website domain, company LinkedIn URL, and company name without explicitly feeding website or LinkedIn content into the prompt. Claygent, our AI web scraper, can analyze these inputs directly to categorize companies according to your custom industry definitions. When you provide a domain or LinkedIn URL, Claygent can visit these pages and extract relevant information to make classification decisions. For example, you can use Claygent to identify specific types of companies (like SaaS) based on their web presence, even with minimal input data. When setting up your enrichment, simply select Claygent from the Tools section under AI in your workspace, and create a prompt that instructs it to categorize industries based on your custom definitions using the provided URLs.
what would happen if the website domain and the found linkedin url returns conflicting information? Would this mess up the way clay ai classify my industry?
When website domains and LinkedIn URLs provide conflicting information, Clay's AI will analyze both sources and make the best determination based on available data. Clay's AI tools like Claygent are designed to handle such conflicts by evaluating multiple data points. When using AI prompts for industry classification, you can specify which source should take precedence in case of conflicts. For example, you can instruct Claygent to prioritize LinkedIn information over website data, or vice versa, depending on which you consider more reliable for your use case. Clay's AI is particularly effective at identifying specific industry types (like SaaS companies) even when information might be ambiguous or conflicting across different sources. To minimize classification issues, consider testing your prompt on a single row first to verify results before running it on your entire dataset.