AI Software and Technology

In this video we’ll look at the tools and technology that can help you to implement some of the AI methods described in our previous activities. Before we can conduct any AI analysis we need to have a place to store all our customer data and, as we have learnt, this data can come from multiple channels. A CDP, which stands for Customer data platform, is a database where all of this can be stored. This includes details about their purchases, preferences, interactions such as live chat conversations, emails or phone calls with customer service, and engagements via social media or review sites. Having all of these interactions in one place provides the foundation for AI driven analysis.

A lot of customer conversation happens on social media and therefore social listening tools such as Netbase, Crimson Hexagon and Hootsuite which collect data about conversations are crucial. These applications feature text analysis, which allows us to understand what people are saying about both our brand and our competitors, and sentiment analysis, which can help identify if the conversation is positive or negative. They use clustering to enlighten us as to what’s trending in the market, giving our brand insight into common topics of discussion and an understanding of what interests our customers. They provide recommendations on what language to use and which products to market to them.

This also allows us to easily identify customer pain points and assess how well the brand is doing through good reporting and benchmarking. Customer experience can also be measured through understanding user behaviour as they interact with your website. Although many brands have deployed web analytics that captures information about where users come from, which pages they look at and hard numbers on the amount of interactions with the site, Tools such as Decibel and Content Square extract deeper insights. They capture detailed information about how the visitors use the site, where on a page they hover or click, which content they actually see and whether it is appealing enough to drive their journey forward.

Using clear visuals including heat maps and even recordings of user actions, they can show you pain points, where they engage and where they don’t. All this can be summarised using machine learning to categorise behaviour and provide recommendations into where improvements need to be made. Speech recognition technology has already found its place at multiple stages in the customer journey. In call centres, software such as IBM Watson Assist and Five9 Intelligent Cloud listen to caller’s voices, prompting them to answer questions that can allow AI behind the scenes to both direct calls to relevant departments, or provide answers on their own. This allows your brand to provide a phone service at all times, not just conventional opening hours.

In retail stores speech recognition is filling spaces where it would be impractical to have a sales assistant. For example, in fitting rooms, virtual assistants built using technology similar to Alexa or Siri can listen and respond to the customer. Need a different size shirt? Just say, and they’ll request a member of staff bring one over. Speech recognition has seen significant improvements and user’s awareness and comfort levels have also increased. Many experts predict that nearly every application and device will have integrated voice technology in some way in the next 5 years. Image recognition is an evolving area in AI specifically when related to improving customer experience.

We have learnt that as brands create more content in the form of images and videos it can be beneficial to understand how users engage based on what they’re seeing. Tools that facilitate this process include Adobe Sensei and Aura, both of which feature image recognition capabilities which allow you to upload your content library to their platforms and gain detailed insights. The algorithms help break images down to basic attributes including whether they contain objects or people and details such as colours, shapes and sizes. These insights are useful for both content categorisation and can provide input into creative thinking. Alternative tools such as the open-source Computer Vision could also be used to accomplish the same analysis in-house.

Although machine learning is integrated into many platforms, for example social listening, it is also a technique that can help analyse customer data without the need for an off the shelf tool. There are machine learning code libraries and packages in programming languages such as Python and R that can be adapted specifically for analysing the customer experience. These can be used to conduct tasks including clustering which help group your customers based on common attributes. If you have in-house data science capabilities it makes sense to explore these options.

If you don’t have in-house capabilities, there are off the shelf tools such as Tensorflow and Periscope that enable you to upload your data and apply relevant pre-built machine learning algorithms based on what you are looking to achieve. For example, calculating the likelihood of an individual completing a specific action such as purchasing a product or deciding what product to recommend someone based on their previous browsing history. Machine learning tools will allow for more effective personalisation and automation. Picking the best tool for your company, based on size and need will allow for improved ROI. We have covered a number of different tools and platforms that use AI techniques.

It is important to be aware of your budget, key priorities and team capability when deciding which tools to invest in versus what can be built in-house.