It’s often said that email is a ‘flawed system that has never been bettered’, which may explain why so many of us spend so much time chained to our email client software applications and alert systems.
Depending on the industry and role, the average employee may send and receive somewhere between 100 to 200 emails per day. Add that burden to newer communications channels that exist across chat, instant messaging, voice and ‘tickets’ stemming from enterprise software platforms and we get inevitable overload, incoherence and chatter.
But there’s a mismatch here. We now live in an age of autonomous intelligence where automation is no discretionary line item in any strategically built IT stack architecture. Shouldn’t we be able to use machine-based compute and analytics engines to communicate better, faster and with less white noise?
The answer is yes and the solution is communications mining.
Say hello to communications mining
A close cousin to process mining and task mining, communications mining is the act of analysing unstructured communications to discover ‘data points’ which logically belong to automation opportunities that can be created to form process and workflow accelerators.
As a base layer, communications mining makes use of technologies including optical character recognition, document capture and intelligent document processing. Combining text-based communications mining with voice requires the use of natural language processing technology to digitise speech into written form so that all communications channels can be corralled and coalesced.
When we have a central pool of communications streams to work with, we are then able to look for patterns that relate to processes and start to build a picture of how the business is being run.
As communications mining starts to paint the digital picture of what’s happening behind human workflows, it enables us to see where the most company resources are being used to achieve different tasks.
Machine learning learning
Training a Machine Learning system to be able to understand the contextual and semantic meaning related to each element being analysed in a communications mining process is an ongoing and continual process. Often referred to as ML in a process of active learning as part of an artificial intelligence engine, communications mining builds an intelligence ‘model’ that can be graded for its ability to predict as it grows and develops.
Best thought of as a cumulatively applied advantage, communications mining can shoulder the heavy lifting in an employee’s daily communication stream. Sometimes also referred to as conversational data intelligence, at its crux, communications mining enables us to take unstructured information (at various labels of unstructuredness) as we convert, transform and manage it into a machine-readable form.
Extremely well-suited to application use cases across ERP, CRM, HCM and other core enterprise software platform use cases, communications mining uses both sentiment and semantic intent analytics to extract meaning from messages and understand what part of the total conversation process they belong to inside any given workflow.
As communications mining develops to become progressively better trained over time, it can be applied to analyse human interactions that straddle increasingly mission-critical parts of an organisation’s operational fabric. Prudently applied, communications mining can help a business to pinpoint inefficiencies, to identify bottlenecks and to exploit core efficiencies and competencies and bring them to the fore for commercial advantage.
As part of wider robotic process automation strategy, communications mining works alongside task mining and higher-level process mining to enable an organisation to create bots that will further drive efficiency and make automation not just a way operating, but also a way of innovating across the business.