25 April 2018

The problem with human language

Posted in Industry

Two languages dominate our world in this digital age - Human and Machine.

Like any language, if you don't 'speak' it, you can't understand it. Machines understand a structured language. Structured means fixed, predefined, predictable data that doesn't vary from what is expected. It's how systems store data and communicate with other systems. Organisations are dependent on their systems and yet these systems struggle to understand ‘human’ which is the unstructured language.

The reason this is such an important subject is that, to any organisation, humans are not only customers but suppliers and employees too. It's therefore essential that organisations can understand the human language regardless how varied, emotional and highly descriptive it is.

The interpreters of human to machine language have been humans themselves. As humans, they can read and understand unstructured language and enter it as structured data into systems that their organisations depend on. Only when the data is interpreted into the structured language can systems store and process the data, and they can then do that at 'machine speed'.

If IBM's claim that 80% of the data gathered by organisations today is unstructured is true, then machines had better learn human language quick.

Enter Machine Learning, the essential interpreter that systems need in order to understand the human language. Let's be clear, we're not talking about semi-structured machine generated invoices, it's the human generated angry complaints or complex enquiries. By enabling systems to understand this unstructured data, organisations can better understand what customers, suppliers and employees are saying in near real-time and then rapidly determine and execute the most appropriate outcome.

Organisations will never achieve competitive advantage while they depend on manual effort to interpret human to machine. They need to enrich the capability of their systems with artificial intelligence such as machine learning, natural language processing, neural networks and fuzzy logic to augment their existing human intelligence to deliver better service and operate more efficiently.

Productivity is the key focus and Celaton inSTREAM™ offers a ready to deploy platform that processes the plethora of content and information that flows into organisations every day from customers, suppliers and employees received by email, post, paper, fax, social media and many other electronic data streams. It learns through the natural consequence of processing and by monitoring the actions and decisions of people who are involved in the process. This 'human-in-the-loop' is essential because the software learns from human intervention in dealing with exceptions. With every transaction, confidence improves, continuous optimisation is achieved and the need for human intervention is reduced.

With machine learning platforms like inSTREAM at the 'front door', organisations are enabled with almost unlimited capability to handle an increasingly complex environment. It’s machine learning, but to customers it’s the best knowledge worker they ever hired and it means better customer service, compliance and financial performance.