Error message

Notice: Undefined index: content in clean2014_block_view_alter() (line 631 of /var/domains/

Until now, most people’s only direct experience with the much-hyped concept of Artificial Intelligence (AI) will be with recommendations on consumer websites and services. Whether it’s a similar book to one you just read on Amazon, a song by a related artist on Spotify, or a series you might be tempted to binge-watch on Netflix, so-called ‘AI’ is, for the most part, sophisticated machine learning algorithms used by these services. These machine learning recommendations certainly add to the consumer experience, as well as to retailers’ bottom line. However, when it comes to real utility in the workplace, we’re only just beginning to see the uses machine learning can have.

But first, a couple of definitions. There is a philosophical debate as to whether artificial intelligence is even possible, and it’s therefore important to distinguish what we mean when we use terms like AI and machine learning. All too often, such terms are used interchangeably when in fact they have very different connotations:

  • Machine learning generally refers to a computer program which can ‘learn’ about some activity, measure how it is done and then use this experience to get better at the task. A classic example would be Amazon recommendations. Using a (relatively) simple set of metrics, the algorithm can track what a user has bought before, what they have browsed for and what other people with similar buying patterns have also bought to produce helpful suggestions (that said, many ‘machine learning’ tools are powered by humans too, behind the scenes).
  • Artificial Intelligence refers to the idea of a computer that can think for itself and autonomously make its own decisions and judgement calls about the course of action it will take. At present, no ‘genuine’ artificial intelligence exists, as this article explains.

These definitions taken into account, it’s still exciting to think about the possibilities that machine learning can have in the workplace.

Minimizing monotony

Perhaps the most obvious use of machine learning tools in the workplace is that they can minimize monotonous and repetitive tasks. Take this online supermarket which uses robots to find goods in their warehouses to add to a customer’s delivery. Using a machine learning tool, the firm hopes to improve the potential of the robots. At present, a robot needs to be able to scan a bar code attached to, say, a potato, to ensure it is picking up the right type of produce. However, by using machine learning, and drawing on a database of millions of images, the robots are being ‘trained’ to recognize various fruit, vegetables and other produce, meaning they do not need to find the barcode on unusually-shaped products.

Allowing humans to focus on more complex matters

The ultimate goal of any machine learning tool is that it should allow the person it’s working for to spend more time and energy on creative activities, rather than dealing with repetitive and unskilled tasks. One bank is aiming to do just this with a machine learning webchat assistant.

The tool has been piloted among staff who manage relationships with small businesses. Most of the time, these customer advisors spend time on the human relationship with the customer, giving them advice during short meetings. During these meetings, the customer may want to primarily ask about, say, a new loan, but might also mention that they have lost their bank card. In the past, the customer representative might spend the next ten minutes trawling through the system, working out how to order a new bank card. However, this is a waste of everyone’s time, and uses no real skill on the employee’s part.

So, the machine learning tool can rapidly resolve these kinds of questions. The employee would simply type in a simple natural language question:

  • My customer has lost their card – what steps do they need to take now?
  • My customer has locked their PIN – how do they unlock it?
  • How do I order a card-reader for my customer?

Using machine learning tools in this way means the employee can instead focus on using more critical skills. But the caveat is that this type of technology really only works when the context is very clear, focused, and narrow. To date, machine learning in a broad, open context-the type of context experienced by information workers-has been unattainable.

Organizing information around how humans think

But now, machine learning is able to save information workers a large amount of time and energy by helping organize information in a way that is most useful for humans. Collage – the world’s first topic-driven solution – is designed to do this by putting the human at the center of the digital experience. Collage uses machine learning to monitor how a specific user is interacting with other people at the company, as well as reviewing what documents they have been working on recently and responding accordingly.  Collage uses a host of complementary technologies to create a meaningful context that enables machine learning to be used to simplify a worker’s daily routine. 

Collage draws on’s own machine learning technologies as well as the Microsoft Graph to build up a profile of the user and the content that is most important to them. It is capable of discovering the links between different kinds of content from different systems (including Azure, Salesforce, ZenDesk and many more), and then aggregates any relevant information for a user when they are working on a specific topic.

When an individual receives an email from a colleague or a customer, Collage is able to scan the email and then present related documents, emails and content that the user—and their closest colleagues—have produced in relation to that subject. This all helps put the user at the center of the digital experience, meaning the technology fits around them and saves them time so they can focus on what really matters.

Machine learning beyond the hype

For us, the most exciting thing about machine learning in the workplace is that it really helps users become more productive and more focused. We explain how in our recent blog – The Science behind humanizing the digital experience. We see the potential for machine learning at work to be huge—and much more significant in terms of real, life enhancing impact, compared with recommendations on consumer websites. We’re therefore very excited to see how machine learning will grow to help enhance the employee experience in the coming years. 

To learn more about Collage and the way we believe it can improve the human experience at work, download your free trial today or read how Swiss Re, one of the world’s largest reinsurance companies, embarked on their digital workplace journey with 

David Lavenda
Chief Product Officer