1. Home
  2. Industries
  3. Publishing RFP Case Study

CASE STUDY

How a large US publisher accelerated the RFP process
while gaining key market insights

Customer

US-based media conglomerate

Goal

Automate parts of the inbound email process for RFPs and extract key information from email requests and their attachments

Challenges

  • This large media company owns more than 40 publications and works with different agencies to sell advertising spaces
  • They receive approximately 100 requests per week, corresponding to about 5,000 messages per year and double as many attachments
  • These Requests for Proposals (RFPs) need to be classified depending on criteria like vertical market or target audience before being directed to the appropriate department
  • The RFPs are highly unstructured documents that come per email, generally with an attachment which can be a verbose word document or a succinct powerpoint presentation with bullet points
  • The fact that RFPs do not follow regular punctuation rules, that extraction targets are very diverse and with little context to learn from made previous attemps to automate this process unsuccessful

Solution

Leveraging the capabilities of SemanticPro Extract & Analyze, a RFP-focused solution has been developed that automatically extracts, reviews and analyzes key data from requests with a high level of precision. This solution is able to handle short texts without punctuation even with little to no context, as well as different document types like Word, Excel or Powerpoint. The solution has been successfully trained to recognize very diverse extraction targets like “Campaign name”, “Client” and “Agency”, as well as very specific vocabulary, like “type of ad products“.

By using SemantiPro Extract & Analyze, the company was able to automatically:

  • Extract a dozen attributes from RFPs, including due date, size of opportunity, vertical market, campaign timeframe, type of campaign, etc.

  • Resolve the problem of missing punctuation by infering the expected punctuation based on features like words location and capitalization

  • Perform these extractions with an accuracy higher than with other machine learning-based approaches

  • Filter documents below a certain confidence threshold that necessitate human review

  • Export the extracted items into Excel documents, or JSON formatted files for importing into Tableau or other Business Intelligence tools

    Results

    • Shorter review time of RFPs
    • Quicker response to requests
    • Higher deal close rate
    • Higher responsiveness

    The Cortical.io Difference

    • The solution extracts key data from requests with a high level of precision
    • The solution is able to handle short texts without punctuation even with little to no context
    • The solution is also delivering critical insights about trends and market shifts
    • The business users have full control over improving the process by verifying, correcting, and adding extraction targets
    • No need for AI experts nor data scientists
    See Contract Intelligence In Action - free demo