Company segmentation can be tricky. When someone else defines your ICP, your targeting can miss the mark. Additionally, creating a well-rounded ICP requires significant manual effort. Tapistro lets you,
Customize how data is categorized and classified, instead of relying on strict databases or outdated categories.
Use AI-powered insights and real-time updates to keep your targeting on-point.
You use the AI Agents for accounts, persons, and signals to define requirements that are used later to enrich the gathered information. AI Agents can include static information such as location and designation or include dynamic or even fuzzy criteria such as natural language criteria. This lets you refine the targeting and positioning of your communication and sales efforts.
In Tapistro UI, the AI Agents option in the left sidebar provides a list the existing agents. You can modify or delete the existing AI agents in the Account, Person, or Signal tab.
AI Agents for Accounts
Account in Tapistro is a company, firm, or an organization. You create this type of AI agent to define your requirements about which organizations you want to target or include in your outreach or GTM efforts.
AI Agents for Persons
You create and use this type of attribute to define your requirements about people and roles that you want to target or include in your outreach or GTM efforts.
AI Agents for Signals
You create and use this AI Agent to define which signals to capture.
To create a new Signal AI Agent, follow these steps:
Log into your Tapistro account and select AI Agents.
Open the Signal tab and select New AI Agents.
Provide a name for this AI agent that helps you identify it uniquely when using it in a journey.
Value type is the type of the result or generated outcome. You can select one of the following values:
Boolean: Most suitable selection when you want to capture a yes or no answer as a signal.
Infer: We infer the type of the returned values and capture the signal accordingly.
Number: Use this when you are looking to capture only numbers from a source of information. For example, use it when you want to capture the revenue of a company from its website and that too as a number.
String: Use it when a free-form text value is captured as a signal. We recommend that you select String for best results. For example, use it when you want to capture the revenue of a company from its website but aren't sure if they've written the value as a number or as a string.
Evaluation Config Type are criteria that you define to evaluate the signals. Select from the following values:
Classification: Use it when you want to do categorize people or account and do segment-based outreach to them. Use the Name field to provide one of the possible values that you want matched and describe it in the Description field. You can add multiple sets of Name and Description values that must be used to classify the information processed. Our AI model uses your description to classify the captured signal. For example, our model can classify which industry does a particular company belong to or what persona does an individual belong to. Another example is that you can use classification in your journey to identify if an account is a high value account or not and define a specific action plan (in your journey orchestration) based on the classification.
Content: Specify if you want our AI model to generate an Email PS that is used to personalize your email communications when your journey is created. Having a personalized trivia or a friendly note in the email that your target people relate to can help improve conversion.
Expression: Currently, contact us to create an expression that helps you derive values from the attributes and may be aggregate these values. For example, you can use expressions to find out how many people have a particular role in an account and aggregate them.
Questions: Use this option to find AI agents that fulfil open-ended questions. Our AI model processes your pre-defined question to find and capture the relevant answer as an attribute. If you want one of the specific values to be your captured attribute, then use Classification above. If you want inferred answers to a question to be your captured attribute, then use Questions option. We strongly recommend that you add only one question to receive accurate results.