Customer Experience Analytics
Contact Center Automation
Machine Learning for Text
Improve profits by understanding the emotional and rational topics that drive customer loyalty.
The big data revolution has meant that there are nuggets of insight within customer experience data everywhere: CRM data, reviews, complaints, enquiries, surveys, social media etc. The ability to harvest and analyse these data in an automated way to provide predictive, actionable insight is a holy grail for marketers and customer experience professionals.
The existing state of the art when it comes to customer experience analytics is using a combination of text analytics, predictive analytics packages and significant manual input from data scientists.
But firms are getting bogged down, spending too much time and resources cleansing and transforming data. CrowdFlower estimates “data preparation” at 80%. Further, assumptions and errors are an inevitable part of the process where human judgement and skill is required.
PrediCX™ is a speedy way to automate the classification of text data by crucially ‘asking’ users for specific input where it needs it to optimise performance. As a result, it can automatically classify Voice of Customer (VoC) data to a high degree of accuracy with minimal user input.
Our Optimised Learning eliminates this huge amount of human intervention and increases accuracy as it quickly starts to ‘ask’ a user, who doesn’t need to be a data scientist at all, for minimum input.
It’s this amalgamation of automation and promoted human intervention that is a new breed of customer experience analytics and is creating huge opportunities for organisations looking to gain more insight and utilise more of the data they accumulate.
For example, one of the world’s largest airlines sends out tens of thousands of surveys every month to customers of all of their services. They analysed the structured data such as Customer Satisfaction (CSat) scores per touchpoint on the customer journey (cabin cleanliness, check in, meals etc.). However, despite having the state of the art text mining technology they struggled to analyse the free text in a systematic way and really only used verbatims to follow up the structured findings in discrete projects which involved highly-skilled analysts spending a lot of time.
By generating automated predictive insight from survey data, the airline was able to not only pick up the structured data such as multiple choice answers and CSat scores, but also automatically generate the reasons why their customers were happy or unhappy and get some indication on what the predictive factors were such as routes, demographics or customer segments. All in near real time.
Find case studies on how PrediCX™ is being used in CX on our resources page
Train your chatbot effectively and minimise the human input required with our latest machine learning for chatbots.
Chatbots are a way to enhance the customer experience and save time and costs for everyone. Yet the Containment Rates (i.e. the ability to avoid passing off to a human) for chatbots can be suboptimal for many situations. The limiting factor here is the capability to classify customer intent correctly. People ask new queries all the time, and find ways to ask for the same query in different ways. It is also difficult for many companies with chatbots to know which part of knowledge base to build out further, or other key parts of the customer journey.
PrediCX makes it easy and highly efficient to classify the intent of customer queries and also to understand which parts of the knowledge base or other response to automate next (e.g. querying a customer transaction database or asking further diagnostic questions). The training of the chatbot is made the most effective by minimising the human input required to maximise the performance of the machine learning models which drive the classification of the chatbot. And the best thing is that it constantly updates to customers’ requests and keep the performance at the levels required for your business.
Find case studies on how PrediCX™ is being used to train Chatbots on our resources page
Save money and improve customer satisfaction with contact center automation and insight.
When it comes to call center automation and contact center automation, we hear a lot of hype about how AI technology, such as chatbots, is advancing to minimize the amount of contact a customer has, so does that mean that the human customer services agent will become obsolete?
Well, underneath the hype of AI, the reality is that chatbot development has plateaued, and the understanding matured.
Alongside this limitation is the fact that customer interactions themselves are becoming more and more complex as customers learn to resolve more simplistic issues by themselves. There are also more communication channels as customers might try to resolve issues by a website, then chat, then phone.
The current best practice is to deploy rules-based text analytics to make sense of customer interactions. Yet it is not reliable as it is based on keywords and needs to be highly curated in a ‘rear-view-mirror’ fashion and it doesn’t have any quantitative metrics to show what the accuracy is. It also can’t be used for predictive or prescriptive analytics. Whilst it has solved the simplest queries, it cannot organically grow to deal with the next level and so reaches a plateau.
Machine learning is an alternative, often applied to classify text based on the computer ‘learning’ the common patterns to predict and classify without necessarily guessing keywords or using templates. The elegance of this solution is unfortunately mired by the complexity of the task which requires data scientists to spend a great deal of time training, building and testing predictive models, let alone cleansing the data in the first place. Again, this has reached a plateau due to the resource constraints.
PrediCX™ generates machine learning models but when it’s not sure about something it asks a human to help validate in a highly efficient way, thus speeding up the process and increasing accuracy. It is really mimicking and supercharging human judgement on an automated, industrial scale. OL overcomes the disconnects because it doesn’t require any data scientists: The training and interaction can be done by a non-data scientist guided by OL. It is really marrying the best of both the human intelligence and artificial intelligence worlds. This can help to make sense of the disparate data and break down the corporate silos to affect a seamless, structured data stream which translates to better customer experience analytics.
Find case studies on how PrediCX™ is being used in contact centers on our resources page
Save time and money, and reduce errors, by automating your business processes such as product warranty, maintenance, insurance claims, healthcare, credit and fraud/investigations.
Imagine if a system could analyse every single piece of verbatim from surveys, reviews and complaints with a high degree of accuracy and automatically transform it, contextualise and classify it accurately, no matter how complex? Even better, what if when the system’s predictive performance dropped, it would ‘ask’ the user to clarify in a highly efficient and minimalistic way? Sounds like science fiction. The good news is that it is science fact as machine learning is now sophisticated enough to achieve this.
The secret behind the latest advancements is a new machine-learning technology known as optimised learning. Optimised learning is a speedy way to automate the classification of complex textual data that ‘asks’ users for specific input for where it needs to optimise performance. As a result, it can automatically classify VoC data to a high degree of accuracy with minimal user input. It works better when used with another innovative technology called automated information retrieval (AIR), which automatically generates and selects features from unstructured data which can be used to train machine-learning models.
Conventional machine learning (with or without text analytics) needs to ‘train’ itself on a dataset parsed and curated by a skilled human. The human will also need to generate the ‘features’ within the dataset for the predictive models to be built on. For example, these could be keywords, phrases or other more complex signals within the data. Indeed, the same data can generate several models — consider, for example, multiple ‘topics’ within a single piece of text or conversation.
Clearly, this becomes complex very quickly and requires skill, time and a degree of ‘art’ and experimentation (perhaps the data scientist is more data artist). There is a lot of guesswork, even with experience, and the more features and data that are explored for this training, the more likely the predictive model will perform well. However, the amount of cleansing, parsing and general tinkering is unknown and indeed if the dataset are random then it is a lot of hard work for zero result.
Optimised learning eliminates this huge amount of human intervention and increases accuracy as it quickly starts to ‘ask’ a user, who does not need to be a data scientist at all, for minimum input. With respect to ‘asking’, what the technology actually does is to point the human user to the records and topics that it calculates will reduce the overall uncertainty for the minimal input. By way of analogy, when a child learns, it learns to ask questions that underlie several facts and concepts rather than just clarifying each one. In this way, it can be said that the child learns to understand. (Note that this analogy is used to show the utility rather than anything more — the analogies between human and artificial intelligence are contentious enough and this paper is not intended to stoke the flames of that issue.) Now the predictive models are being generated, and improved in a highly efficient manner by a non-data scientist armed only with domain knowledge (e.g. it could be an agent in a call centre).
This amalgamation of automation and promoted human intervention is creating huge opportunities for organisations looking to gain more insight and utilise more of the data they accumulate. It is also easy to see how new use cases are possible that were not previously feasible due to the time lag and expense of resources. One such case is an early warning classification system that flags up topics and issues and quantifies as they arise.
Achieve accurate automated filtering of content with minimal human input.
Inappropriate content is becoming one of the biggest concerns for many businesses online both from a reputational and customer experience point of view, and also from a legal risk perspective.
However, preventing and removing inappropriate content is hard because it can be subjective in some parts, and ever-changing.
Currently content moderation is performed by a mixture of humans and machine learning, although because of the above reasons, this is both costly and sub-optimal.
PrediCX™ can act as an effective dynamic Content Moderation filter by generating and updating machine learning models to ensure the appropriate levels of performance and risk to prevent the bad, enable the good and quarantine the risky content as efficiently with human labor as possible.
Improve customer conversion by optimizing search relevance.
Spending online marketing budgets on attracting customers can be expensive. Therefore it is crucial that customers are able to search easily to find your products and services, as well as advertising in the most relevant and effective places.
Whilst this sounds obvious, it is far from straightforward sometimes with all the potential noise and complexity from web textual data. Also, different individuals search for the same thing in many different ways and styles, and of course with shorthand or imperfect spelling and grammar.
PrediCX™ can help supercharge the labeling process and model training to dramatically improve and enhance search to bring more customers to you, more effectively and more economically.
PrediCX automatically labels and classifies your text data just as a human might.
Don’t miss out on valuable Sentiment Analysis. Understanding context, emotion, and sentiment in consumer data is one of the most powerful tools for text analysis of customer data. However, it is easily lost in the huge amounts of text data contained in reviews, complaints, surveys and social media. Furthermore, traditional rules-based and keywords based text analysis lack the capability to accurately classify sentiment and context.
PrediCX is a super charged labelling engine that automatically classifies data where there’s a lot of text. PrediCX will label concepts rather than rules/keywords, automating Sentiment Analysis so you know exactly what your customers are thinking in near-real time.
Our Sentiment Analysis tool will figure out common patterns, which are potentially a lot more than single keywords, and takes into consideration sentence positioning, context through associated qualifiers, and whether they are nouns or verbs, just as a human might.
Transform your complaint handling from an expensive and reactive experience into an opportunity for improved customer experience.
Using our latest machine learning for Complaint Handling software PrediCX, complaints can be turned into great opportunities for learning and/or converting negative feedback into positive customer advocacy.
Text-based complaints can be long and complex, containing many topics and themes. PrediCX uses the latest in machine learning and Artificial Intelligence (AI) to automatically classify key concepts in the text data, enabling:
– Accurate classification of incoming queries, matching them with the ‘next best response’, be it a resolution, a request for further clarification of the query, or re-direction to an agent where necessary
– Automation of insight and mapping of ‘hotspots’, speeding up future resolution
– Early warning of issues flagged in complaint handling, allowing proactive resolution
– Non-data scientists to extract clear and actionable insight
© 2017 Warwick Analytics. All rights reserved.
Registered in England & Wales. Number 07724630. Registered address 35 Kingsland Road, London, E2 8AA. VAT 120435168.