{"id":33241,"date":"2018-11-13T16:45:01","date_gmt":"2018-11-13T21:45:01","guid":{"rendered":"https:\/\/digital.hbs.edu\/platform-rctom\/submission\/can-artificial-intelligence-replace-community-managers\/"},"modified":"2018-11-13T16:45:01","modified_gmt":"2018-11-13T21:45:01","slug":"can-artificial-intelligence-replace-community-managers","status":"publish","type":"hck-submission","link":"https:\/\/d3.harvard.edu\/platform-rctom\/submission\/can-artificial-intelligence-replace-community-managers\/","title":{"rendered":"Can Artificial Intelligence Replace Community Managers?"},"content":{"rendered":"

Artificial Intelligence is the 21st\u00a0<\/sup>century buzzword and we hear things that already make us cringe: self-driving cars, robots replacing humans, articles being written by machines so why not the extinction of newly created jobs such as community managers?<\/p>\n

Even though scary, artificial intelligence opens up an incredible amount of opportunities particularly in the digital marketing industry.<\/p>\n

Currently, European small businesses have missed the rise of social networks[1]<\/a>because of limited human and economic resources, and thus cut themselves off from their customers. In May 2016, I co-founded Artur\u2019In, an AI-powered community manager, to enable small to medium businesses to take full advantage of the power of social networks in a simple, affordable and effective way.<\/p>\n

Artur\u2019In publishes content daily on its customers’ social networks and allows them to monitor and analyze their activities through a dashboard. In addition, the solution sends monthly newsletter campaigns and creates a website and a blog for each client.\u00a0Over 1500 French businesses today beneficiate from the solution at 189$\/month.<\/p>\n

Initially, AI enabled us to solve three main challenges in order to launch Artur\u2019In:<\/p>\n

    \n
  1. Content quality<\/li>\n
  2. Content personalization<\/li>\n
  3. Low price point<\/li>\n<\/ol>\n

    In fact, in order to satisfy our customers, we needed to ensure personalized, different and good quality content at an affordable price.<\/p>\n

    We thus created our proprietary deep-learning algorithm. It first classifies thousands of contents daily by topic and geolocation. The algorithm then ranks contents based on their estimated performance on social media and selects the top 1-3% quality contents.Next step is formatting the contents to generate social media posts: creating a title, writing the post summarizing the article, adding the appropriate hashtags and selecting an appealing image.The posts are then automatically distributed to the social networks of a client based on his industry, the topics he selected and his location.Finally, the algorithm gathers results on each publication and becomes increasingly smarter in its selection and ranking of content.<\/p>\n

    \"\"<\/a><\/p>\n

    Three key technical challenges were solved within 2 years:<\/p>\n