Data for good: Sentiment AI to support corporate social responsibilityBy Noel Peatfield on August 1, 2018
More and more consumers are making a conscious effort to judge brands based on their ethical beliefs and their stance on certain issues. With this in mind, how much of a role does AI-powered sentiment analysis have to play? In #DataForGood blog number three, Noel Peatfield explains...
According to a recent survey carried out in the UK by Morgan Stanley, there is an increasing demand for ethically sourced goods. Consumer priorities such as value and customer care include good ethics, a retail trend that is shown to be on the increase.
A report by Edelman Earned Brand studied 14,000 consumers in 14 countries to find out how ‘belief-driven buyers would choose or avoid a brand based on their perception of the brand’s stance on a controversial issue.’ It found that 57% are buying or avoiding purchases based on the brand’s position on a social or political issue – which is 30% more than three years ago.
Millennials are shown to be the most likely demographic to buy based on their shared beliefs with a brand, and are also the most prolific social media users. The Edelman study also highlighted that belief-driven buyers won’t buy from a brand that stays silent on an issue it had an obligation to address, with 67% buying from a brand for the first time because of its position on a corporate social responsibility (CSR) issue.
As retail continues to move online, so has the conversation, with customers now more knowledgeable than ever about ethical practices, sustainability and CSR. While reports from surveys can reflect opinions from previous months, data can be captured from social media in real time.
In a study carried out by Juniper Research, the current spend on AI for retail will grow from an estimated $2bn to $7.3bn in 2022 – of which 54% will be spent on sentiment analytics and 30% on automated marketing. Both of these disciplines can work together to better understand and respond to the shared CSR issues being discussed on social media.
Most brands and retailers are already using social media analytics to respond quickly to individual cases, but with over 25 billion Twitter engagements every day, NLP and AI can recognise how large groups of users are collectively feeling about subjects surrounding brands, products and services.
Using social media data, sentiment analysis is being used to automatically categorise groups of posts by determining whether the writer’s attitudes were negative, positive or neutral towards a specific subject. However, with AI, the range of emotional intelligence for machines is widening to include nuances of negativity and positivity such as joy, thankfulness, anger and sadness.
The emotive data inherent in conversations around CSR subjects can be understood by NLP applications, but there will be challenges ahead to meet the AI adoption rate for sentiment analysis forecasted by Juniper Research. The context in which sentiment is expressed through words and understood by machines is key to assess the why, how and what of the sentiment.
The advancement of NLP can be clearly seen with conversational AI interfaces like Alexa and the three times winner of the Loebner prize, Mitsuku. A quick chat with the chatbot Mitsuku demonstrates the level at which AI can already understand and respond to words, and how quickly this technology is developing.
Amazon Echo utilises AI-powered assistant Alexa
Social sentiment analysis needs a system to follow up on its data, and businesses assess sentiment on CSR issues in order to respond with appropriate communications. Social media community manager at The Co-op, Sophie Newton, explains in an insightful blog that they are using social listening to make more informed policy decisions and to create content to clarify their messages.
AI is already at the forefront of NLP and sentiment analytics, and of course, it’s playing an increasingly significant role in marketing automation, too. Artificial intelligence is ideal to deliver on the model of social listening and marketing response when it comes to CSR policies and messaging. Improving the speed and accuracy of this process with AI can lead to meeting the increasing expectations of retail customers.