AI Personalization in Pitchbox - A Comprehensive Guide

AI Personalization in Pitchbox enables users to scale the labor-intensive aspect of personalization within their workflows by extracting specific information from websites, articles, or webpages utilizing AI. This feature streamlines the process of gathering targeted insights, making campaigns more efficient and personalized.

Connecting AI Tools for Personalization

Before diving into AI Personalization, it's essential to connect Pitchbox to one of the supported Large Language Model (LLM) tools - OpenAI or Anthropic. Here's how you can set this up:

1. Navigate to the Integrations Page: Find this in your Pitchbox settings.
2. API Key Entry: Enter your API key for either OpenAI or Anthropic, depending on which service you choose to use.
3. Choose Your AI Model: Select the AI model that best suits your needs. More sophisticated models tend to yield better results but come at a higher cost.

It's important to note that the costs associated with AI Personalization are borne by the user. Pitchbox does not charge additional fees for using this feature or for processing the AI tasks. This ensures transparency and allows you to manage your budget according to the AI model you select and the volume of personalization you intend to perform.

Getting Started with AI Prompts

To utilize AI for personalization, you must first set up AI fields. These fields can be configured in two ways:

1. Campaign Level: Personalization fields are tailored to individual campaigns. To set these up, navigate to Campaign Settings > Personalization Fields. When creating a new field, choose "AI Powered Single Line" or "AI Powered Multi Line" data type.
2. Account Level: For broader applications, personalization fields can be set up at the account level and are called "Custom AI Fields". These fields are accessible across multiple projects and campaigns, offering a standardized approach to personalization. Access account-level fields by clicking on the menu icon in the main menu and selecting "Custom Fields."

Once you define your AI field, the next step is to draft your AI Prompt.

Crafting Your AI Prompt

Creating an AI prompt involves specifying the information you wish to extract. Here's how to compose a prompt that delivers targeted results:

1. Be Specific: Clearly state what information you need. The more specific your prompt, the more relevant the output will be.
2. Use Available Parameters: Tune your prompt by including parameters that reference specific parts of the content you're interested in. This allows for more precise targeting of the information you wish to extract. For example:

1. Include [H1] to focus on the main heading of a webpage.
2. Use [Title] to refer to the title of the webpage.
3. Reference [Link and Anchors] to analyze hyperlinks and their text.

Specify Desired Response Format

Indicate the response format you expect to guide the AI in providing the most relevant output. Here are a few examples:

• Yes/No Format: To check if a webpage includes a specific feature, such as a "Write for Us" link, refine your prompt with details on where to look: "Respond with 'Yes' if there is a 'Write for Us' link in the [footer] section, otherwise 'No'." This tells the AI exactly where to search for the information, making its response more accurate.
• Numerical Rating: When asking the AI to rate something, like content relevance, clarify your scale and what the numbers signify: "On a scale of 1 to 10, rate the content's relevance to [your topic], where 1 signifies not relevant, and 10 means highly relevant." This not only guides the AI in providing a numerical rating but also explains the significance of the lower and higher ends of the scale.

Available Parameters for Reference in Your Prompt

You can enhance your prompts by referencing specific content and link fields. Here's what's available to you:

Content Fields:

• [H1], [H2], [H3]: Headings of different levels
• [Title]: The title of the webpage or article
• [Author]: The author of the content
• [Article Date]: The publication date of the article
• [Link and Anchors]: Hyperlinks and their anchor texts
• [body]: The main content of the webpage
• [meta]: Meta tags within the webpage, including descriptions, keywords, author information.

Open Graph (OG) Tags:

• You can reference any Open Graph (OG) tag (e.g., [og:title], [og:description]), as long as it exists on the page. This allows for the extraction of specific information designed for social media sharing, directly from the webpage's metadata.

Link Fields:

• [url]: The URL of the link
• [rel]: The relationship of the linked document (follow, nofollow, sponsored)
• [anchor]: The visible text part of the link
• [section]: The section of the webpage where the link is located (footer, header, body)
• [type]: The nature of the link (external, internal)

Understanding Validation and Feedback

After submitting your prompt, it undergoes a validation process to ensure it's set up for success. This process helps ensure that your prompt is clear, effectively utilizes available parameters, and aligns with Pitchbox's capabilities. If improvements are needed, you'll receive guidance on how to refine your prompt:

• Scoring: Your prompt is evaluated based on its clarity, effective use of parameters, and compliance with Pitchbox's functionality. A score between 0% and 100% will indicate how well your prompt meets these criteria.
• Suggestions: Should your prompt score be below average, feedback will be provided, highlighting areas for enhancement and offering concrete suggestions. This may include tips on making your prompt more straightforward or leveraging specific parameters to deliver better results.

This tailored feedback ensures that you can adjust and improve your prompts, enhancing the relevance and effectiveness of your personalized outreach efforts.

How to Utilize AI Personalization in Your Campaign

Now that AI Fields are configured and AI tools connected, it's time to personalize opportunities at scale. 

AI Personalization can play a critical role at specific stages of your workflow:

1. Inspection: Utilize AI Personalization to quickly determine whether opportunities are a good fit for your campaign. At this stage, we recommend running AI on specific fields and quickly filtering or deleting results that don't match your prospecting criteria. 
2. Personalization: Leverage AI at this stage to personalize opportunities for outreach. This ensures that each communication is tailored to your prospects, increasing your response rate. 

Steps to Personalize Your Campaign

After setting up AI Fields, you have two methods to initiate AI Personalization for your opportunities:

• Personalize All AI Fields: Personalize all AI fields for all or selected opportunities. 
• Choose Specific AI Fields: For more targeted updates, you may select specific AI fields for personalization. This is particularly useful if certain fields require additional refinement of the prompt and you want to go back and re-personalize specific fields. 

Performance and Processing Time

AI Personalization processing time varies, generally taking a few seconds per opportunity. For bulk personalization, the total time may extend to a minute or more. Pitchbox indicates which opportunities have been personalized as the process completes.

Monitoring AI Personalization with AI Status Indicators

Keep track of personalization progress and errors through the AI Status column. A red indicator signifies an error has occurred during personalization for an opportunity. Detailed error messages are accessible by clicking on the indicator, aiding in troubleshooting and correction.