numero is presenting a paper at AI-2009, the conference of the British Computer Society Specialist Group on Artificial Intelligence (BCS-SGAI) at Peterhouse College, Cambridge, December 15th-17th 2009. The BCS-SGAI is one of the leading artificial intelligence societies in Europe and AI-2009 is their twenty-ninth annual international conference. This is the leading UK-based conference on Artificial Intelligence and the longest running AI conference in Europe.
The presentation describes how the numero uno software platform uses advanced artificial intelligence technology to analyse inbound messages from customers, in order to meet the demands of modern, multi-channel contact centres. Particular focus is given to how numero uno learns to distinguish different kinds of messages, assign them to the correct business category and help to maximise the efficiency of a team of agents handling inbound messages across multiple channels.
numero uno constructs a complex statistical model of the language used in different kinds of messages. For example, it learns to distinguish the language that people use when writing to complain or to compliment, to apply for an account or for a job. This model enables the system to categorise a new inbound message (regardless of its source including phone, email, written corresponse, SMS etc) and then route it to the agent who is most suited to deal with that type of message, providing customers with the highest level of service available.
Identifying the category that a message belongs also allows numero to correctly prioritise messages so that, for example, serious complaints are sent to the front of the queue. Furthermore, once the system knows what kind of message it is dealing with, it can present the agent tasked with responding with the most relevant information, tailored specifically to that message; this includes the details of an ongoing enquiry, historical customer information, relevant knowledge-base articles and FAQs, suggestions for text to use in the response, and so on.
A statistical model determines how confident numero uno is when categorising 
a message. This can be used to decide whether an automated response should be sent. The software platform also learns from any mistakes it makes, based on feedback from human agents. If a message is assigned to the wrong category, an agent can correct this error; which is then used to fine-tune the statistical model and enhance future performance.
The power of numero uno is such that it can help businesses with large customer service operations to streamline the customer journey for each individual customer – taking account of their lifestyle characteristics, communication preferences and the precise nature of their enquiry/interaction. The importance of such a tailored approach is clear, when considered in the content of the polarisation of consumer needs and attitudes as highlighted within numero’s recent research report into consumer segments.
Simon tees off 2012
