Debunking the AI myth in customer service – what to watch out for

As featured on Compare the Cloud.

AI is transforming the world as we know it. Its applications in the e-commerce space have led to a massive shift in consumer perceptions about bot-generated content. Almost every e-commerce site now offers some sort of personal recommendation based on customer preferences. This is AI in a relatively simple form. An increasing number of businesses are also investing in chatbots to bridge the gap between the customer and salesperson in a digital world - capturing and qualifying interest based on their online behaviour.

While these applications of AI are now common place, bot-generated responses are still a long way from making an impact in the customer service sector. We believe some of the reasons for this are:

What if the bots get it wrong? Brands are fearful about the level of accuracy and more importantly, the relevancy of answers provided by bots. An example is Facebook which scaled back its AI ambition after its chat bots hit 70% failure rate.

What if my customers feel undervalued? Fear that bots might remove any scope of personalisation from the interaction. Customers feel comforted, reassured, and engaged by a human interaction. A bot service might not always be intuitive enough to pick on the tone of voice, style of language, and tailor the dialogue accordingly.

The Limitless SmartCrowd platform is all about democratising AI offering our proprietary AI service as plug-and-play for businesses looking to automate their customer service operations. Our SmartAI has been carefully developed to mitigate the well-known risks associated with automation. Here is how we do it:

1. AI and humans working together

AI by itself doesn't work; SmartAI is a blended solution which offers the benefits of automation without losing out on the human touch by being a component of the SmartCrowd (crowd sourcing) platform. SmartCrowd enables enterprises to build their own crowd of expert customers who provide highly personalised and relevant responses to customer queries. This is called CrowdService. SmartAI constantly learns from a stream of relevant conversations, lending a more personalised, human feel to the automated suggested responses. What's more, these suggested answers are being used to train and onboard new customer experts on the Limitless platform.

2. Improved accuracy, faster learning, and reduced costs

Our process collects x100 more data points than other methods allowing it to predict answers much more reliably and learn faster - reducing operational costs.

As the table above shows, a normal call centre handles a million contacts. A trained and knowledgeable third party like a Quality and Compliance Officer must check every agent interaction for accuracy, before the data is used to train the AI. Separately, the system must rely on customers indicating they found the "AI-generated suggested answers" to be helpful, to train the AI further.

Our SmartAI however, saves time and money through 2 levels of quality control - firstly, the end-customer provides feedback on the answer. Secondly, fellow Ambassadors independently peer review every answer provided - this setting can be configured to drive as many peer reviews as required, depending on the task complexity. This means the AI can learn faster from a wider set of data points and predict answers with a higher level of accuracy.

3. Unique and personalised

Our AI IP is unique and based on real client requirements. Algorithms facilitate advanced routing, personalization, rewards and anti-gaming features on our platform. SmartAI is platform-agnostic, and easily slots into your CRM and knowledge management systems, so you can be up and running in no time.

4. Flexibility and agility

When you deploy SmartAI along with CrowdService, it will complement your existing workforce and not replace it. Our blended solution is designed to sit right next to your traditional model, helping improve quality of service and boosting productivity. So why not let the AI handle transactional queries, leaving your crowd of customer experts to deal with those requiring the human touch. This gives your trained customer service agents more time to deal with the most complex and sensitive enquiries.

Furthermore, the pay-as-you-go business model ensures you are only paying for resolved cases - no overheads resourcing or training.