BPM Analytics

The impact of AI and automation on social customer care

Customer service is no longer limited to chatbots, phone calls, or emails. It’s gone social. In 2023 the number of social media users worldwide touched 4.9 billion. As per a Twitter report, 64% of active Twitter users say they would prefer to message the brand’s Twitter support handle rather than call them, and 75% of those who do so, expect a quick response. Social media is now a bona fide service channel and the one with the maximum public exposure. And with regulatory authorities having a presence on social media, brands can be at the receiving end from all sides. These numbers are only expected to rise.

81% of customers on social media will only recommend the brand if they receive a response. Developing a customer service channel on social media will be imperative for most businesses. The public nature of the medium poses a unique challenge as each customer query is also a marketing and PR scenario, not just a service request. Other related challenges include the varied demographics and, therefore, customised responses across multiple social media channels, quick response requirements, the need for skilled staff, and complex response workflow. 

Developing a framework for social media customer service will streamline the process. The framework should include a detailed strategy for training and creating a dedicated and combined response and resolution team, analysing and designing an optimum workflow, automating using AI solutions, and creating templates for platform-specific responses. The workflow for social media customer servicing has the following steps: identification of support request posts, verification of posts, allocation to the pertinent team, contact with customer and gathering of information, providing resolution via the social media platform or other channels, and intelligent analytics to study and improve efficiency.

AI and automation bring in the capability for quick response and resolution and insights for further improvement. One of the major hurdles is identifying the service request posts associated with a brand. Negative reviews and support-related posts can be across social media, from Twitter or LinkedIn to a Subreddit. The post might not even have the correct brand handle. An AI-enabled social media monitoring tool can track posts and messages related to the brand on all the chosen platforms, which are then sifted through a trained NLP (Natural Language Processing) classification tool that separates the posts that require customer service attention. NLP-based intent identification can further allocate the requests to the appropriate support team.

40% of customers who post a query or issue on social media expect a reply within an hour. AI-based automated response solutions bring in speed, competence, and customisation. Generative AI models trained on the organisation’s knowledge base and with real-time access to the knowledge base can provide customised, natural, precise, and up-to-date responses to customer queries. An additional AI verification layer for fact-checking will eliminate invalid responses. The solution can consolidate information and provide step-by-step instructions. It can be trained to analyse the tone of the customer post and accordingly adjust the response to facilitate a more positive interaction.

Certain service requests will need a human touch, especially when the post has gone viral or is made by an influencer or a person with a huge following on the platform. It is better to have such posts and messages handled by a highly trained and experienced collaborative team of PR and customer service executives. If a post, especially one that is a complaint, is not dealt with finesse, it could cause much damage to the brand.

Analysing customer service KPIs to measure and improve performance is vital. Some important KPIs for customer service are average response time, first call or contact resolution (FCR), average time taken to solve, effectiveness, customer engagement during the interaction, and brand sentiment. AI automated tools bring in a huge improvement in response time. Providing AI-enabled chatbots on social media platforms like Facebook allows the customer to get immediate support. A high FCR indicates efficient service, and this score should improve drastically with generative AI tools trained on the knowledge base. Likes and retweets for a customer service post indicate a highly positive customer sentiment, and analysing this data provides insights into what the business is doing right. AI sentiment analysis tools can navigate social media platforms to measure customer engagement and brand sentiment on the customer support handles and get a true picture of whether and how much customer service has improved.

Managing customer service through social media is a financial opportunity. In the long run, handling service via social media is at least six times cheaper than doing it through call centres. AI-enabled customer service through social media will soon be the norm.

This article was first published on The Evolving Enterprises