The role of AI in CX in 2020

2020 could be the year AI takes the world by storm. Large and small companies are adopting it, and both users and customers see this technology’s potential. One respected forecast expects as much as many as 95% of all customer interactions will be through channels supported by artificial intelligence (AI) by 2025.

Many see it as chatbots and virtual assistants, but it’s far more: AI transforms the engine of the whole company. The vast majority of business leaders expect AI to transform their industry. Research by Gartner indicates that “in a few years from now, 89% of businesses will compete mostly on customer experience”, and yet according to IBM, only 4% have plans to be disruptors themselves.

In this blog, we’ll discuss the possibilities of AI technology in Customer Experience. Learn how you can be a disruptor in your industry. We will also include some examples of how Wonderflow’s technology is being used by global brands to stay on top of their customer-centricity game.

All CX professionals know that dealing with consumers generates lots of data. Navigating multiple platforms to make sense of structured and unstructured inputs is a difficult task. CX data is messy, and customer behavior can seem erratic. Until recently, it was regarded as the nightmare scenario for data science. 

And that’s where AI comes in. AI systems understand unstructured information in a way similar to humans. But they not only consume vast amounts of data with far higher speed, they learn from interactions. This intelligence allows them to glue data together and fill in gaps to come up with meaningful, actionable analysis. Allied with human knowledge in establishing the business context, AI in CX allows for smarter, faster decision-making based on real data.

Much of the computing behind AI in CX isn’t new at all. What has changed is the ability to do it on a larger scale and very quickly. AI and Natural Language Processing (NLP) can analyze customer datasets, break them down into individual components, and divine the individual’s intent. The AI program can then make predictive suggestions based on the interaction.

AI systems in CX have three essential building blocks that will enable the whole process. They start by ensuring the researcher has a complete overview of the customer. After this, insights are delivered in real-time. Finally, the results can be applied to the company’s business strategy. Let’s go through these in more detail.Data unification: data unification makes the creation of a single customer view cheap and fast. In technical terms, it’s the process of joining together multiple data sets from different sources and preparing them for analysis by matching, deduplicating, and cleaning the records. It’s the activity that consumes more than 60% of all data scientists’ time, and it’s essential. The aim is to pull in all your data from CRM, call center, web, and retail systems to build a single view of the customer. 

Take the example of deduplicating data. You’ve probably done this yourself in Excel and will know what a tedious but critical task it can be. In deduplication, we are looking for an accurate and scalable way to cluster records (usually from different data sources) that represent the same entity. At its core is the ability to link records, for example, to decide if two records belong to the same customer or if they could be errors.

Take the UK postcode SW1P 1WG which could be written as SW1P1WG or erroneously SW11WG or Swp11wg. Would the fields Zip Code, Postcode, and Address line 3 all represent the same data? Most data scientists will try and treat these using an ETL script or a rule-based tool, depending on the volume of the data. However, with CX projects, much of the data will interact in ways that can’t be computed immediately.

With large datasets, this may mean multiple passes or extensive stretches of computing time as the consequences of changing and mapping are worked through. With AI, an algorithm can be constructed to cluster other attributes such as business name or telephone number to identify duplicates to merge or flag for correction. With large organizations working with legacy systems, it’s necessary to build systems that learn from the data and adapt. By using machine learning-based automation to recommend how attributes and records should be matched, the AI helps organizations work faster with more data. The AI platform enables the connection of large numbers of data sources efficiently, addressing one of the most significant needs in modern diverse, fragmented IT environments.Real-time insights delivery: For any system to make an impact on CX, insights have to be realized in real-time, perhaps even in front of the customer. This is where front end API and reporting systems have a considerable part to play. Even if you can develop insights, being unable to make them meaningful could make your entire investment pointless.

Wonderflow’s dashboard, the Wonderboard, is an excellent example of how AI can quickly deliver insights. Based on the Artificial Intelligence of NLP, our technology automatically mimics the human ability to comprehend texts which can be used to generate automatic insights. 

The results are not just for data scientists. The Wonderboard pulls together data 24/7 and can integrate into established business reporting patterns. After the initial setup, the analysis takes minutes and can be run by anyone. 

Business Context: AI in CX is all about understanding the customer journey. However, for each business, this is unique. It’s highly unlikely your customer journey will be the same as your industry peers or competitors. The business context of CX means that each step of the journey needs to be captured and measured often, to understand revenue or cost impacts.

This means that the AI must understand the business impact of every interaction. Simply labeling them “email” or “inbound call” and attaching an average cost isn’t enough, because the touchpoints interrelate. Costs are dependent on the customer, and the journey they take shapes the touchpoints and vice versa. Untangling this is hard but necessary as these holistically affect KPIs.

We’re only just beginning to see how AI can transform the CX landscape. Our early indications are that the customers love the impact and that for organizations, it can be internally transformative. In this section, we’ll talk about some of the ways AI-enabled CX can change your business.

Chatbots, Virtual assistants and personalization

Probably, the most obvious way AI supports CX is through the rollout of Chatbots and Virtual Assistants. When people think of chatbots, they often think of text-based interaction. Still, developments in AI speech-to-text recognition and NLP have opened the door to chatbots with voice-activated functionality. These can be launched on online and mobile platforms and also within contact centers. The purpose is to act as a gatekeeper and deal with more straightforward requests by using simple keyword recognition to address the customer to informational content that can help them, like FAQs or other particular forms and resources.

The next level of sophistication is when the same technology is allied with NLP to determine customer intent- for example, being able to detect frustration to escalate an interaction to a human service agent. Similarly, customers that express delight can be directed to an agent for upselling opportunities. By integrating the chatbot with a recommendation engine, customers can benefit from personalized recommendations.Contact Center Analytics

Most inbound contact centers generate a massive amount of data. This can be audio logs, tickbox form outputs, or speech-to-text transcripts. In the majority of cases, this is used retrospectively for simple contact center metrics like service level or agent performance. 

However, speech analytics allows us to analyze the sentiment in phone calls effectively. Were customers angry? Which product features keep on getting mentioned in angry calls? Did the customer raise their voice? Algorithms can be used to track the sentiment in transcripts, which can be used to see the effect of product updates or to find faults before customers get to leave a review. Detect Emerging Customer Experience Issues

Most CX professionals will focus on a small number of metrics obsessively. Most are mature and well understood, even if they have flaws- NPS or CES, for example. Usually, this kind of feedback is collected at the end of the customer’s journey, and by its nature, it can be easily skewed by especially poor aspects of performance. It doesn’t give you quick information, and it tells you little about how the customer feels. For that, we need to look at unstructured data. Unfortunately, this is held in a large number of spaces across the enterprise- from the contact center to retail to sales forces.

AI-enabled customer journey analytics can find every single relationship in the data that exists without explicitly being told to look for it. Similarly, although sales teams are the frontline of understanding how new products are doing, there are several ways AI can be used to detect issues in customer experience very early on. 

For example, being able to know that “great price but feature Z loads slowly” in a four-star review detects the customer is happy with the price aspect but that “slowly” indicates product issues and the critical “but” tells us that this is not a primary issue yet. Other markers can be created to identify channels or customer groups, which could be having particular issues to head off poor reviews or excessive load on customer service teams. AI allows you to close the loop and act on issues as they arise.

Find Insights Across Customer Journeys

We mentioned above that one of the critical aspects of using AI is to pull together various platforms to build a single view of the customer. The single view allows you to take action to improve not just the customer experience itself but, through consistent measurement, organizational responsiveness to your customer’s needs.

For example, when an offer is launched through email, you can track responses in retail, contact center, and web. This allows you to focus on the journey rather than the touchpoints and, with some experience, allows effective prediction of intent for future offers. This, in turn, means a higher return on marketing investment, lower customer service costs, higher rates of cross-selling, and upselling.

AI-enabled CX can pull data from various sources to deliver insights in a format that is accessible to anyone. And more, smarter data is the beginning of a virtuous circle: empowering analysts to ask better questions about customers.

AI-driven customer feedback analysis, therefore, can improve customer experience dramatically. Read more about how Wonderflow helped a Global 500 consumer electronics company to obtain and understand relevant and actionable insights, consequently building improved customer experience. https://www.wonderflow.co/use-cases/customer-feedback-faster