Error message

Notice: Undefined index: content in clean2014_block_view_alter() (line 631 of /var/domains/
Right click to download images

After an explosion of hype about ‘Big Data’ three or four years ago, people have gradually become aware of certain limitations to the concept. Despite the promise, getting any value from Big Data is still hard work, and certainly isn’t as simple as turning on an analytics machine which miraculously starts crunching numbers.

In short, one of the clearest Big Data limitations is that it’s, unfortunately, nowhere near as ‘easy’ as it first seemed. Indeed, Gartner announced in 2013 that Big Data had surpassed the peak of ‘inflated expectations’ on their Hype Cycle curve which projects how new tech will be adopted, and was descending to the ‘trough of disillusionment’. 

Nonetheless, four years on we have a soberer view, and the concept of Big Data is arguably moving toward what Gartner calls the ‘plateau of productivity’—i.e. actually being useful. One area this is happening more and more is the ever-growing number of enterprise apps that claim that, besides doing project management, CRM, document storage or whatever else, they can also use Big Data to help with all manner of business decisions.

From Microsoft’s ‘Intelligent Search & Discovery’ to Salesforce’s Einstein, tech companies are bolting AI-enhanced features to their tools which intelligently scan all the data that a company holds, before presenting this to the user.

Big Data limitations must be addressed first

There’s no doubt that Big Data has potential, and it’s exciting to see how companies are attempting to innovate and help end users. Nonetheless, there are certain Big Data limitations that these tools seem to pass over in their desire to offer people colorful charts, ‘real time’ graphs and projections. 

  • Correlations – not relevancy

One of the most important Big Data limitations that seems to arise time and again in these new AI-enhanced tools is that they aim to present people with information that they might be interested in. The way such tools work is to track the activities that employees are working on. If they find two people who seem to be working on similar tasks, the applications ‘surface’ this information to each employee.

Sounds good, yet this approach has a serious flaw. Because the machine cannot understand the context of the data it crunches, it will present people with apparently ‘related’ content which, in fact, is nothing of the sort. Read about this in my recent CMS Wire article.

  • Distracting people with data

Many of these AI-enhanced tools present their Big Data features as a kind of virtual assistant, showing users timely information that helps them make better decisions. This is universally presented as a good thing…but is it really what most people need?

For instance, say you were a salesperson, and on every page of your CRM you were presented with a chart telling you how many sales you’d made this quarter. Perhaps it would provide recommendations of leads it thinks you might want to follow up with. Maybe it would prod you with information on how far you were from your targets.

Now, this sounds good in theory, but in reality? Besides being patronizing, this sort of tech presents a huge, constant distraction (a problem we’ve covered before). People don’t need more notifications, messages, directions…they need the information to help complete the tasks on which they are now focused.

  • Limited value in the ‘real world’

A final problem with this technology is that these AI-enhanced apps do not recognize a fundamental reality of modern enterprise workers: they need to use multiple apps to complete their work.

Take Office 365’s Intelligent Search & Discovery, for example. It’s great if everyone in your organization exclusively uses SharePoint Online to store content, OneDrive for Business for personal files and Yammer for conversations. In this case, the tool is really useful.

But the reality is that businesses use a much broader range of technologies (than Microsoft) for all kinds of interlinked tasks; Companies routinely use Salesforce for CRM, Outlook for email, ZenDesk for ticket support—in addition to Office 365’s many apps as well. This is important because AI on one of these systems can only tell you about the links between content on that system; and that’s not enough. Because what’s important are the links between content and people across all systems.

Beyond Big Data limitations

So, while Big Data is an exciting concept, most tech providers are simply ‘bolting on’ AI-enhanced analytics tools to their products without addressing the Big Data limitations highlighted above.

Of course, there’s no denying that there is some potential value in this kind of information, and it can potentially help with decision-making and saving time. Nonetheless, to really fit around what knowledge workers need, what we require is a holistic view of big data and a way to combine the data in a way that delivers to each knowledge worker what they need, when they need it…without the overload.

We built (with our unique Collage technology) to bring knowledge workers information in context from all their different applications. Collage appears as a sidebar in the employee’s Outlook interface, so that when they select an email, uses AI to search for related content from across email, SharePoint, their company CRM and many other apps. then uses its unique Enterprise Graph to connect information between many sources by topic, such as customer, product, service, or project.  The Graph is able to surface ‘what’s most important now,’ by taking into account Big Data, and then prioritizing each data piece according to what the worker currently working on; for example by gauging the affinity of a colleague who updated a CRM record and how recently was the notification posted.  By using the current email the worker is viewing as a context anchor, can uniquely present the most relevant related information to the worker when they really need it.

Thanks to the power of the Graph, information workers who use are spared the need to switch contexts while rummaging for information. Rather, the most information is brought to them exactly when they need it.

See why with Collage makes sense for your information workers. Request a demo today.


David Lavenda
Chief Product Officer