White Papers

Read our in-depth guides to using text analytics

LIVE CHAT: How AI and machine learning is improving live chat

This paper outlines how the latest in AI/machine learning can help optimise
'live chat' channels to improve the speed of resolution, provide more relevant
responses and streamline the chat optimisation process.

AGENT EMPOWERMENT: Replacing CSat pressures with AI

Contact centres are reaching a crossroads: chatbots and automation have helped reduce demand but their impact is plateauing as remaining queries become more complex to automate. On the other hand, CSat and NPS have equally never been under so much pressure as human agents are left to handle evermore complex tasks in a market where customer expectations continue to rise.

This paper sets out the key issues affecting the use of CSat surveys in contact centres and suggests ways that customer agents can be empowered by, and more fairly measured by, the use of AI.


Machine Learning can be used in text analytics to classify text based on
'concepts' rather than using keywords and rules. This paper shows the
differences between these two techniques with a real case study and dataset.

Everyone who uses text analytics directly or indirectly intellectually understands the theoretical differences between
machine learning concepts versus keywords yet there is still somewhat of a disconnect in the market place between what
each technique is good for and the capabilities and limitations. Which one should I use? What difference will it make? There
are also some fanciful claims on proponents of both sides which only add to the confusion.

The objectives of this research were to qualify and quantify the differences between ML concepts versus keywords based around exemplary datasets. Whilst this paper was sponsored by Warwick Analytics, a machine learning company, the research was conducted at arm's length using two distinct datasets and blind controls. The other text analytics companies are market leaders and not named herein. For the purposes of clarification, we will be using the word 'label' to describe each count of the issue for both techniques. This is synonymous with 'tags' in other literature.


Whilst there are a lot of fanciful headlines and hyperbole about the latest algorithm, the reality is that to deploy a machine learning model in an operational environment, it needs to be trained well on relevant data, and if the environment changes, to continue to be trained so that it adapts.

This leaves the machine learning experts in a quandary: How can businesses develop machine learning models which automate processes and contact centers not just today, but reliably ongoing? How can they get continually rich insight from models when the data are changing around them. Is the irony that the data scientist cannot bring a model to life like Pygmalion, but actually needs to constantly be the puppet-master: You can't just build a training set of data and then automate, you need to keep feeding training data to keep the models up to date and prevent model degradation.

AUTOMATED ANALYTICS: Turning Predictive Analytics from a Project into a Product

The field of 'Predictive Analytics' is receiving significant attention lately due in a large part to the rise of 'Big Data'.

However this presents significant practical challenges in terms of extracting timely and valuable insight quickly from disparate, dirty and unstructured datasets, without the need for an army of data scientists.

This paper aims to give a broad overview of the current state of Predictive Analytics, the various common techniques and their applications and limitations. It will also attempt to challenge a few myths along the way.

We will then show some of the latest emerging techniques and how these are able to drive reliable insight and prediction in a timely manner for business users without the need for data scientists, even with disparate, 'dirty' and indeed unstructured datasets.

PRACTICAL PREDICTIONS Predictive Analytics for Everyone

The field of 'Predictive Analytics' is receiving significant attention lately, due in a large to the rise of 'Big Data'. However this presents significant practical challenges in terms of extracting timely and valuable insight quickly from disparate, dirty and unstructured datasets, without the need for an army of data scientists.

Submit your name and email below to receive this free download (there is no limit to the number of downloads you can request):

© 2019 Warwick Analytics. All rights reserved. Registered in England & Wales. Number 07724630. Registered address 35 Kingsland Road, London, E2 8AA. VAT 120435168.