Launch year: 2016
Country: USA
Funds raised: $12m

Planck: Underwriting made easier through AI

Concept description

Planck: Underwriting made easier through AI

Using just the prospective insured’s business name and address, Planck scans all public information available (including government databases) to find all relevant pieces of information about this business. Examples of relevant information returned by Planck include whether or not the building has received fire code violations, comments on social media about bouncers, whether or not the place serves alcoholic beverages, presence of smoke detectors etc. This information would usually be requested through a lengthy insurance application questionnaire, if at all.

Planck uses a combination of AI-related technologies to return the most relevant information possible to its clients (mostly insurance carriers or brokers) such as:

  • Image processing: being able to retrieve information about for instance exit signs or smoke detectors from images
  • Natural Language Processing: understanding meaning behind for instance web reviews (who is doing what in this business)
  • Satellite and geolocalisation analysis, such as assessing distance from fire department
  • Unstructured web analysis: making sense of information in any type of web format

Planck focuses on the small commercial Property & Casualty space, such as bars, restaurants, artisan contractors and other “mom and pop” shops.  and advertises 90% accuracy and coverage.


Value proposition

  • Lowers costs of “pre-underwriting”: automatizes the internet search when receiving an insurance submission. No more than 2 questions needed, anytime
  • Prevents fraud: Easy to verify 2 answers as opposed to multiple ones
  • Broadens data insights: Planck retrieves and analyzes a lot more information than what a typical insurance submission would contain
  • For brokers: improve and streamline insurance submissions. Planck’s API can be directly integrated into a broker’s existing workflow

Illustration (Video)

PHP Code Snippets Powered By :
Scroll to Top