Advocates and campaigns have a new tool to help them sift through the deluges of words contained in everything from emails to opposition research.
Natural Language Processing (NLP), which is being used by everyone from Facebook to IBM, can give advocates new ways to ingest, search and analyze huge amounts of text – commonly referred to as “unstructured content.” Considered one of the main branches of Artificial Intelligence (AI) technology, NLP can also automate the process of spotting specific issues of interest, trends, and patterns of meaning contained in massive bodies of language.
Unsurprisingly, given their financial resources, banks are among the organizations pioneering the use of NLP to solve some of their most critical content analytics challenges. The French bank Crédit Mutuel, for example, announced recently it will use NLP to help its advisors analyze more than 350,000 daily customer emails, to rapidly assess their contents and determine the level of urgency and provide faster, smarter service.
Back in D.C., it’s worth recalling that during the Obama administration, White House staff sifted through an average of 7,000 letters per day, 70 percent of which were submitted by email.
These 7,000 average daily incoming emails were manually sorted by members of the Office of Presidential Correspondence. Their contents were recognized as an important source of valuable insights for the administration. Cody Keenan, the president’s speechwriter, reported drawing heavily from these thousands and thousands of daily emails when drafting messages and talks for the chief executive.
Even with as large a staff as the White House employs compared to a typical campaign, the time required to search, sort, and analyze the contents of several thousand emails manually, compared to the speed and thoroughness of automated NLP tools, represents a quantitative and qualitative difference.
With a tailored solution that tags keywords, concepts, named entities, organizational affinities, or issues of direct interest to a campaign or incumbent, email messages can be instantaneously ingested, sorted, and analyzed for patterns as well as sentiment at the speed of light. In turn, incoming emails can be routed automatically to the appropriate personnel according to preset rules.
Sentiment analysis capabilities such as those provided by IBM Watson’s Tone Analyzer, for example, can automatically detect and categorize whether the author of an email is happy, angry, or disgusted, allowing campaign managers to quickly identify and address issues of immediate concern. The insights made available through NLP sentiment analysis can add a whole new level of responsiveness to those responsible for effective campaign messaging, potentially produced now at the same speed as the all-important news cycle in today’s 24/7 information world.
What’s needed are additional practical uses of advanced AI solutions to gain a practical advantage. Starting with the application of established NLP tech to rapidly gain valuable insights from large bodies of incoming email content represents one of the most natural places for nearly any campaign or advocacy group to start.
John W. Davis II is founder and CEO of N&C Inc., provider of the Regendus analytics solution.