We're thrilled to announce a game-changing addition to our platform: Web QA. This groundbreaking feature is designed to redefine the way our users explore, discover, and leverage data on our platform.
What is Web QA?
Web QA is a custom enrichment service that empowers our users to tap into an extensive pool of readily accessible data on the web. This tool allows you to ask any question and get precise answers right within your account on our platform, making your data discovery and enrichment process simpler, quicker, and more efficient.
Imagine a scenario where you need to find answers to questions about a specific company - things like such as its commitment to sustainability, the date of its latest data breach, or perhaps, the number of patents hit holds. Instead of conducting numerous web searches and wading through multiple sites, Web QA fetches the answer for you with a straightforward approach.
Just input your question into the Web QA tool, and it will search the web for you, fetch the top results, and use state-of-the-art AI to extract the answer for you. Plus, you have the option to define the format of your answers as Text, Boolean, Number, or Year to better suit your needs.
For our expert users, we offer the ability to create a custom prompt template. This feature will be particularly useful when you need the answer format to be a Text-type but also to follow a specific set of rules.
How does it work?
You will head to the Enrichment tab in Tactic and add a new “Web QA” enrichment.
You will configure the enrichment with a Label, Answer Format, and Question template.
The “Label” will be used to display the answers in the Table View of the app, the “Question template” will be used to search the web for search results, the results will be fed into GPT which will be used to extract the answer in the specified “Answer format.”
The “Question template” will be a free form text field which allows the user to ask any question they like including a template variable “{company}” for the name of the company we are researching for them. When we run the enrichment we will fill in the name of each company in their CRM for each search. We will take the templated question and run a web search, get the top three results, and feed them into GPT to extract the answer.
The “Answer format” must be set as Text, Boolean, Number, or Year. We will use our typing system to extract the answer from GPT in the specified format.
There is also an option for expert users to create a custom prompt template which will be fed into GPT with the search results. This can be useful when the user wants the answer format to be a Text-type but also follow a specific set of rules.
After you save the enrichment and run the project the results will show up in the Account table. All results will include a url Source which will tell you where we got the answer from.
