Everyone is talking about generative AI, and we already know that the quality of its results depends on the quality of the data used. Creating high-quality datasets is not a trivial task, but this is what we do best at Explorium. Here are a few examples of how we use GPT3 in our product, and how we keep our datasets fresh and high-quality.
New transformation step: GPT
A new transformation type is now available, leveraging chatGPT by open.ai.
This transformation enables applying extremely powerful manipulation and intelligence on text. You can feed the step with free text instructions, describing what you would like to do, which parameters to use (data columns) and how you would like to get the output.
You can think of it as a super complex SQL step without the need to know SQL.
Here are a few examples of how to use the GPT step:
- Classify companies using a closed list of categories
- Extract information from google results – to determine whether a restaurant serves alcohol, offers any form of delivery, and specifically offers 3rd party delivery
New! Professional profile: Contact & workplace
Get a 3600 view on employees by accessing contact information sourced from the professional profiles created by individuals and the firmographic information on their workplace. The data bundle includes employee’s email addresses, phone numbers, social platform URLs, education, work experience, and more. The individual’s workplace information includes: industry category, revenue, company size, HQ location, and more.
Sourced from multiple partners and sources, utilizing Explorium’s Entity Resolution methodologies
Significant improvement in coverage and precision
Fresh data that is updated frequently: varies between daily updates to quarterly updates
Coverage improvement: Company trends by workforce demographics
Expect to see up to 34% better coverage on SMBs in this bundle providing visibility into company trends based on employee statistics and aggregations by quarter.
Data is sourced from professional profiles of employees and displays rate of change trends, providing benchmarks used for company evaluations.
Historical data available from 2015 onwards.
Number of professional profiles: selected quarter
Number of professional profiles: YoY
Number of professional profiles: QoQ
3 million new businesses were added to this bundle, providing access to the contact information of a company’s employees. Expect higher coverage on SMBs.
Employee email address
Email address validity
Employee LinkedIn URLs
Expect a more stable experience while using this live bundle due to a seamless change behind the scenes. You can now access insights and results about companies from leading search engines at a greater scale and improved performance.