'Machine learning' and 'artificial intelligence' have been important terms in 2017, and eventually these technologies will have a significant impact on the way enterprises do business and communicate. But what benefits will they really bring to communication, and how can businesses start preparing now so they can reap the benefits more quickly when it arrives?
Many businesses are already automating some of their interactions. For example, customer service centres are increasingly supported by chat bots and automated assistants that can help direct calls or answer basic enquiries. Machine learning and AI add to this by eventually making automated customer communications infinitely more intelligent and useful. The holy grail is fully autonomous customer service, based on real-life data, where machines can comprehend interactions and provide intelligent responses, as well as understand intonation and sentiment direct from voice recordings. Escalation of an interaction would only happen because of value, or lack of progress, enabling businesses to save more highly skilled customer service agents for tasks that require human logic, intuition, and empathy - where they can really add value.
The future is bright, although the path ahead may appear tricky to navigate. There is, however, a clear development path to achieving fully automated customer service, and businesses can start planning now to put in place the practices and processes that will allow them to maximise the benefit from machine learning when the technology is enterprise-ready:
1. Make communication contextual: Adding context to business communication is a simple step that vastly improves its effectiveness. ‘Contextual communication' simply means being able to do communication within the context of a task or transaction. An example of this is ‘app-less' voice and video or text communication from within the shopping mechanism of a website, allowing a customer to find information and then interact with the business via that website - importantly from the environment (or context) that they're already operating within. This gives every party more effective engagement, as well as aligning with the way today's consumers want to communicate: it's simple, accessible and instant. For businesses, it provides much more insightful data about how customers interact with businesses, including about their behaviours, attitudes, choices and so on, providing useful information upon which to base future systems, services and products.
2. Classifying communications: In its simplest form, machine learning is effectively pattern recognition, meaning that the more patterns it has to draw on, the more intelligent it can be. For businesses to get the most value from machine learning, it needs access to a database of conversations and business systems so it can learn and understand patterns and categorisations. Businesses can already begin to accelerate this learning by capturing, classifying and tagging communications, including call recordings and automatic transcriptions, as well as their outcomes and sentiments. Logging why a communication was successful, as well as markers for the most productive conversations - and what a failing one looks like - all helps to build up a valuable database. This gives more opportunities for the machine to identify patterns using actionable insights.
3. Combine context and machine learning: For tagging to be really useful, it needs to understand context; there's a big difference in a customer calling to return an item that is faulty versus one that is the wrong size for example. Without that context, machine learning and AI are limited, meaning that its ability to give good or accurate answers and follow the right process flow is also limited.
Context helps because it is an intensive process to develop an AI or machine learning system to do learning for any arbitrary piece of knowledge; machine resources simply don't exist to do that. Layering machine learning onto contextual communications reveals why a customer is there and what that customer's journey was to reach that point; records the outcome combined; works out if the communication was effective or not; and finally provides ways to make it more effective if needed. By linking just the appropriate databases with CRM systems across the business - sales, marketing, contact centre - businesses can provide a really effective way to improve the workflows and processes that underpin customer engagements and experience, and feed that into machine learning databases.
Contextual data about a customer, a transaction or big data trends allow better decisions to be made at the point of communication, and enable a much more intelligent system that can deal with more requests.
4. Focus on the job at hand: Businesses should think strategically about how they will use machine learning and where it'll deliver the most value back to the business. Furthermore, constraining the machine learning to a specific job makes it much more effective at giving good answers. If you know context of a communication's purpose, for example a customer service request generated from part of your FAQ, or a sales enquiry, then you can deploy machine assistance to do that more effectively within the resources you've got.
Machine learning and AI are already starting to show how they can streamline and support business communications. Ultimately, businesses are keen to drive data-driven, personalised user experiences and the technology exists to deliver this. But context is the most important part in getting this right - without it, automation will fail, cause confusion and lead to frustration. The convergence of contextual comms and AI has the potential to be really exciting, freeing up human-to-human interaction time to the areas where greatest value can be added. This is where we'll see fundamental transformations in how the real-time enterprise of the future will communicate - via human or machine, or a mixture of the two - with its employees and customers in context: at the right time, with the right information at their fingertips and in the right application.
Rob Pickering is CEO of technology firm IPCortex
Molybdenum ditelluride is a two-dimensional material that can be easily stacked into multiple layers to create a memory cell
New light-guiding nanoscale device can control and monitor a nanoparticle trapped in a laser beam with high sensitivity
Optical traps are scientific instruments in which a focused laser beam is used to exert an attractive or repulsive force on a microscopic object to hold it in place
Scientists estimate that the exoplanet has already lost up to 35 per cent of its mass over its lifetime
The observations were made using the Atacama Array in the Chilean desert