- Input – the data flowing into the central system from outside
- Process – the action of manipulating the data into a more useful form
- Output – presenting the data as information in a user-friendly way
Tuesday, 20 August 2013
Decision-as-a-Service: Applying Analytics at the point of Capture (part 1 of 2)
Many organisations have recognised the need to apply some kind of document and data capture technology on the journey to the digital enterprise. After all, you cannot realise the vision of the paperless office if you are still pushing paper around.
Leading organisations are now embarking on the next phase of their digital journey. For those that have laid the initial foundations, the transition to the digital enterprise is less about paper and more about data.
“Now that we are capturing and routing all inbound documents electronically – how do we make use of the actual content?”
From Digital Mailroom, to Customer Dialogue Management
The description of capturing and routing all inbound documents is of course the digital mailroom concept. Such systems point to proven ROI – speeding downstream business processes through automated classification and routing, thereby increasing the overall effectiveness of an organisation. Things get done quicker and smarter, with less effort, less paper and less cost. But what comes next?
Looking at the bigger picture, all processes, communications, interactions and decisions, be that in a consumer or business capacity, have one thing in common – they share 3 core components – an input element, a processing element and an output element.
In other words, the information flow moves from data, to insight, to action.
In this sense, the Digital Mailroom, acting as the input element, represents only the start of the value chain and therefore just the beginning of any end-to-end business improvement initiative. Traditionally, the processing element “what is the value of this content, and therefore what do I have to do with it?”, or more simply, the decision point – has been a job for humans, but will increasingly become a job for machines as more advanced technologies emerge.
Organisations across the globe have realised the benefit of handling incoming content (paper, email, web, PDF) through one multi-channel content capture platform – standardising the capture process. But what about the response process?
Leading organisations are one step ahead and have recognised the strategic importance of the Digital Mailroom and its ability to drive customer-centric activities. The system is set to evolve from a purely admin-centric operation, stationed on the organisation periphery, to one placed at the heart of business operations driving high-value, high-impact customer dialogue.
Decision-as-a-Service: Automated Decision Management
The concept of a fully automated input-process-output chain of events – touchless processing with minimal, or zero human interaction, is human-computer interaction (HCI) at its best and straight out of science fiction. Ask a question, receive an answer – in real-time. Sure, it is a futuristic one, but an increasingly realistic one when you begin to combine advanced technologies in recognition, analytics and linguistics.
Imagine a state-of-the-art system where incoming content (requests, enquiries, complaints, questions) is analysed with human-like interpretation to drive automated interactions, decisions and actions (answers, offers, recommendations, approvals) in near real-time or even real-time itself.
Such a concept gets close to a real-time conversation platform between businesses, suppliers, customers, employees and artificial agents, with a myriad of one-to-one interactions in a futuristic, always-on, hyper-connected society. No waiting, no down-time.
Setting the Scene
As a customer of a large mobile phone provider, you would like to email customer services with a specific complaint – you are receiving spam text messages and you are requesting they put a stop to it within 2 days, otherwise you will switch provider.
In today’s environment, not only is this not possible due to a lack of multi-channel capabilities by the provider (they cannot handle email), it is also reliant on a human processor who has to understand the nature of your request (requires human logic), and who has a number of other tasks to handle (delayed response). Further, due to a lack of integration with your historical customer data the agent is unable to appease you when it matters most (loyalty programmes and special offers) and the net result is that you remain an unhappy customer (defecting to the competition).
The business case stacks up when you multiply the issue across your customer base. Not only does the mobile phone provider have a potential customer attrition problem, it is also missing out on leveraging customer insight, in real-time, for up-sell / cross-sell opportunities. Reason enough to put forward a case for customer experience optimisation and the subsequent implementation of more advanced technological capabilities. The outcome: happier customers, healthier balance sheet.
Integrating Multi-Channel Input with Multi-Channel Output…and a whole lot in between
The idea of an automated business decision platform – driven by an integrated input-process-output chain of events sounds simple in practice, but in reality is rather complex. The truth is that there are a number of enterprise software applications implicated in this story, not least CRM, ERP and Database Management systems holding a plethora of historical data, which may, or may or may not need to be drawn upon at the “decision-point” in real-time.
In addition, the output element – or outbound communications component – is fragmented between the fast-growing marketing automation space and declining print media space. Traditional document composition systems will need to mature into multi-channel output systems (including mobile and voice capabilities) capable of creating meaningful content on-demand, as opposed to heavy-duty document production.
And last but not least, the glue that joins it all together, semantic analytics, acting as the central brain and decision-engine, remains a super-technology in early adoption. Even then we should not discredit human intervention – after all, we are still some way off true artificial intelligence, meaning intelligent systems will still require a form of human touch. It is likely that decision management platforms will initially surface as decision-support engines, running as a centralised enterprise service. Consequently, they will automate much of the business process, but still require human guidance.
Part 2 of this article will explore barriers to adoption, and key recommendations to move this concept towards reality.