How to score your first AI quick wins: Social Listening

Posted by Stephanie Organ on 15-Jan-2020 09:58:00

In ai use case, AI in marketing

Marketing is a ripe area for Artificial Intelligence (AI) adoption, with all the data and the insights, but how do you make the jump into the AI pool when you look around and all you see is resistance? Quick win projects are an essential tool for building confidence among your team, particularly when introducing new concepts. That is why Nightingale HQ have created a guide of quick-win projects to help departments and companies ease into AI and build momentum for future, more complex projects. In this edition, we will discuss the benefits of practising social listening and how to pull it off as your first AI win.

Social whispers for self awareness

When it comes to promoting your brand and engaging with your audience, you’re probably already well aware of the need to have a conversation. Nobody wants to see a one-sided story repeatedly plastered over their social feeds or inboxes, they want interactions and engagements that build real relationships. But to be really socially aware, you also need to listen to the things people aren’t saying directly to you. That’s where social listening comes in.

social-listening

The value

Social listening helps you keep on top of your reputation as a brand and catch negative comments, even when they weren’t sent directly to you. On top of this, it can also help you keep on top of keywords and topics that your customers are discussing and reach new customers that are discussing things that you offer. You can track industry trends as inspiration for your next campaign or track the competition and discover the pain points their customers are facing, giving you an opportunity to do this better.

You can even find new influencers to work with and use your listing to better understand how customers interact with each other and with other brands and use this to shape your strategy. You can use it to better understand what your customers really want, which can lead to better customer relationships and increased average spend per customer. Social listening can give you a whole lot of insights to act on, and as an extra, it can save you loads of money on customer service. So how do you get this underway as an AI project?

The project

Social listening tools can be fancy and expensive, but you can also turn it into a quick-win AI project and do it on a fraction of the budget using automation tools like Microsoft Flow or Zapier combined with other tools like Power BI, Dynamics 365, and off the shelf APIs such as Cognitive Services.

The idea is very simple and starts off in much the same way with the automation tool of your choice. You will set up a trigger, let’s say for Twitter, and enter the hashtag or keyword that you want to monitor. You can do this using a template or from scratch, depending on what you want to achieve. You can use a template to send all the information to a google or excel spreadsheet, to a Sharepoint list, email them to yourself, and a whole suite of other options.

If you started from blank, next you’ll need to add a step and choose an action. You can choose a Cognitive Services API to detect things like adult content, sentiment analysis, or even languages that are most commonly being used to talk about your keyword or hashtag.

Once you’ve filled in the details of your action you’ll need a final step to decide where to send the data. If you are using sentiment analysis, you might only want to trigger an action for very negatively scoring tweets, so you would set a condition for tweets coming in below a certain score. The flow will be split in two and you can add separate actions from each option, or from only one. Additionally, you can split one the options with another condition, so high priority negative tweets trigger one action, whereas low priority negative tweets trigger another, etc. Again, the actions you can choose from are endless, so it’s up to you what you decide to build.

Another cool option is to send your data to Power BI, in which case you’ll need to set up a new “streaming dataset” there first. The best way to get all the data, including user information, into Power BI is to select the “Email myself new Tweets about a certain keyword” Twitter template and modify it by adding the action “Power BI - add rows to a dataset". Once you have this all set up to send to Power BI, you can go back and delete the "email" activity.

Once your data is in Power BI you can play around with visualising it and now you can even enable Cognitive Services within Power BI, so you can apply Sentiment Analysis, Key Phrase Extraction, Language Detection, and Image Tagging.

You might not be able to cover all bases using the above methods, but there are a load more tools available to you, like IFTTT which can track mentions across a wider range of social networks, or you could go for some of the top tools listed below:

  • Quick Search
  • Google Alerts
  • Hootsuite
  • Mention
  • Keyhole
  • TweetDeck
  • Followerwonk
  • Awario
  • Social Mention

More quick wins

If it sounded easy enough to conquer your first steps into AI with that marketing quick-win, why not try out some of our projects that cover sales, productivity, FAQ bots and more. Download our free 7 Quick Wins Projects guide.