Clay Icon

Best Practices for Writing a Prompt and Using Clay to Filter Ideal Client Profiles

Β·
Β·

Hi there...can someone please share with me how best to write a Prompt and use Clay together to accomplish the following challenge, I'd be extremely appreciative. -When running a search to find the correct Ideal client profile such as VP of Engineering or CTO for a company that is 1,000-5,000 employees...the problem is, this lead is working multiple jobs...many times 2-3 jobs. I'm receiving lead data whereby (in this example), they are the VP of Engineering or CTO for one of the other companies they're working for...but those companies have 50 employees. So, they are NOT a good ICP, but they show up cause I asked for CTO title of 5,000 employee company, but he's a CTO AND currently works at both companies. I'm having this happen many, many times and it's killing me cause leads are booking meetings....but they are they're not a good ICP. They're the CTO for a 100 man company and he also works as a Senior Engineer at the 5,000 man company. If you could help or point me in a direction on this...that'd be great. πŸ™πŸ˜Š

  • Avatar of Clay T.
    Clay T.
    APP
    Β·
    Β·

    Hi Adam, happy to help!

  • Avatar of Clay T.
    Clay T.
    APP
    Β·
    Β·

    There are probably multiple ways to approach this problem but one that comes to mind is you could first create a company table that meets the "company aspect" of your ICP (i.e. 1,000-5,000 employees) and make sure that you are satisfied with that list of companies.You could then run a 'Find People' search from the company table (via the button in the top right of your table) and indicate the seniority level of the people you want to find from those companies. This process should first guarantee that all your companies match your requirements and then only find individuals based on that. Does that make sense?

  • Avatar of Adam
    Adam
    Β·
    Β·

    Anjali D. hoping this works...time will tell, looks like it though. Appreciate your helpπŸ™Œ

  • Avatar of Greg H.
    Greg H.
    Β·
    Β·

    Adam if you're starting from Clay's LinkedIn Enrichment, I would start with a column that pulls the experience objects from the LinkedIn enrichment where the end_date property is null (all of the companies that list a "Present" end date) using this formula:

    JSON.stringify({{Enrich Person from LinkedIn Profile}}?.experience?.filter(item => !item.end_date))

    Then I'd feed the value of that column into GPT 3.5 with the following prompt.

    Review the following JSON data representing a LinkedIn user's work experience, and return ONLY a comma-separated list of the values of each `url` property where the `title` property equals a job title that is some variant of "CTO" or "VP of Engineering". DO NOT INCLUDE ANY OTHER COMMENTARY OR FORMATTING SUCH AS SINGLE OR DOUBLE QUOTATION MARKS - ONLY THE COMMA SEPARATED LIST. If you are unable to make a determination, or if there are no matching criteria, RETURN ONLY AN EMPTY STRING. Here is the JSON to review:
    ```
    {{Present Companies}}
    ```

    from there you can prompt Claygent with:

    1.) Visit each of the following list of LinkedIn company profile URLs
    2.) For each company page, navigate to the "About" tab, and  determine whether or not the "Company Size" section contains the value "1,001-5,000 employees".
    3.) If ANY of the companies on the list satisfies that criteria, return `true`. Otherwise, return `false`. If you are unable to make a determination, also return `false`.
    
    Here is the list of LinkedIn company URLs: 
    ```
    {{CTO/Engineering URLs Response}}
    ```

    Any rows with a true value from that last column are matches.

  • Avatar of Adam
    Adam
    Β·
    Β·

    Greg H. you’re amazing! Thank you SO much!! πŸ™Œ πŸ˜„