Select Page

CIO Series: Why Does AI Matter when Migrating Content to Office 365? – Part 1

by | Jan 19, 2020 | Cloud, Governance, Office 365

This post is part one of a two-part series. Find the second part here.

Buzzword Bingo

AI is a hot topic. It is the current buzzword of choice for software companies and integrators across the spectrum of technology.

When something is a hot topic, all kinds of material tends to pop up about it. Why should you use it? How should you use it? Why you need it NOW!

For me, AI is different. It is more than a technology trend. It is a sea change. A change that potentially holds the key to unlock the one thing many organisations, both public and private sector, struggle to achieve with their content – value extraction.

For Keep’s Sake!

Organisations today are awash with content. Swimming in data that has been accumulating since the day the first member of staff put words down on paper. Analogue, digital, structured, unstructured; you name it, organisations have been accumulating it.

For reasons that many organisations can not explain, they love to horde.

The belief that ‘one day’, a document written eight years ago for some long-forgotten reason will be needed drives a constant desire to keep everything ever created. The purpose is rarely a factor. Legality an often overlooked consideration.

The relentless rise of cloud computing has amplified the problem. Storage is, to all intents and purposes, free. Virtual machines are inexpensive, and commoditised legacy systems such as SharePoint can be operated relatively cheaply.

Unless held in a sophisticated ECM solution (which is, of course, maintained at a high cost to the delight of the ECM vendor) this collection of stored information is vast, cheap to maintain, but mostly pointless.

Although available ‘just in case’, the *value* of stored information is limited. Unless content curation is a thing within an organisation, and let’s face it – that is relatively uncommon – all of this information is just ‘there’.

Raiders of the Lost Ark

Let me tell you a story. I consulted recently for a well-known government-funded organisation. World-famous. A household name. They have content. Oodles of content. Like many organisations with heritage, they measure their storage in yards – thousands of yards. Somewhere in the region of 5 or 6 *miles* of stored information.

Vast warehouses maintained at precise temperatures and with sophisticated fire suppression systems are housing millions and millions of pages of data – like the scene at the end of Raiders of the Lost Ark, but without any (let alone ten) commandments.

Unusually, a vast percentage of the data had a purpose. Unfortunately, this purpose was almost unrealisable in a practical sense. When information *was* needed, it took weeks to locate and extract. The ‘index’ (for all but the most recent of content) was a combination of microfiche and yet more paper — nothing short of retro.

A project tasked with looking at the scale of the digitisation project estimated that 1,000 operators would take in the region of 15 years to bring the information into a digital state. It is crucial to recognise that the proposed 15 years was to scan and file only with the filing using the most basic filing approach; implementing a digital version of the current index. A vast scanning ‘factory’ would be required, and a cost estimate deep into nine figures meant that the project was ultimately scrapped.

New Rules for an Old Game

Sure, it will take time and resources to scan miles of documents. Tens of millions of dollars of time and resources. But the filing piece is where the value will be added, where the purpose of the information becomes realisable, where AI will shift the needle.

If we think of ‘adding value’ in the context of information, we are talking about making the data useful. Making the data useful for people means making it findable. Findability is everything.

What makes content findable? Information Architecture (‘IA’) and metadata.

Good IA provides the basic geography of where content should be and enables the construct for taxonomy which allows the use of metadata. Useful metadata allows additional information to be attached to content to provide context beyond the ‘geography’ of location.

Value extraction goes beyond ‘adding value’. Value extraction is the sweet nectar that the wiliest bee will harvest from a flower. Value extraction is about releasing the information you did not know you had. It is about enabling productivity, enabling decision making, and enabling improvement by making use of what you already have. No systems. No new applications, No additional change programme. Just getting more out of what you have been collecting for years – information.

This post is part one of a two-part series. Find the second part here.

Photo by Ravi Roshan on Unsplash