How to Boost Agent Productivity Using Dark Data

When we open long-forgotten boxes of memories: photos, souvenirs and snippets of events and vacations, we all go on an emotional rollercoaster ride. We often can't bring ourselves to dispose of them—instead, we accumulate and hoard even more things over the years so that we can preserve them for future generations!

Similar troves of information can be found in companies. Data like emails, chat messages and social media posts are poured into "data lakes" and then left to be forgotten. Companies often fail to see that the data stored on their servers is a treasure chest of untapped "dark data", waiting to resurface and be put to better use.

So, how can you make dark data work for your teams? There are two challenges in unlocking the potential of hidden data: finding dark data in the organisation, and then making sense of the data after uncovering it. Read on to learn how you can use data from customer and employee interactions to help your teams work more efficiently.

What is dark data?

Dark data is information collected by companies during business activities, like conversation data, that is kept but not used for any other purpose. Because it cannot be saved in a typical database format, it is also known as unstructured data. Hence, dark data such as audio recordings are simply saved as files in their original format.

Estimates of dark data as a proportion of enterprise data range from 40% to 90% depending on the industry. This represents both a growing risk as well as an opportunity for technologically savvy companies. However, why do businesses store so much dark data in the first place if they are going to be unused and left alone?

One reason is that businesses have departments who work in silos. Different teams may store the same customer's data format on their own servers or according to their terms. If your sales team, for example, knew that the marketing team saved the country location of website visitors, they would use country data for prospecting.

Complacency is another reason. Software apps and server processes generate real-time data 24/7, so leaving them alone may be easier than hunting down every file created. Plus, storage is now seen as cheap and limitless, so rather than deleting files methodically, it may be wiser to keep them just in case they are needed!

Why should you care about dark data?

If dark data is simply kept on servers and not used, why don't we just leave them alone? Companies may not know exactly what type of data they are storing if they cannot see it. This raises issues of liability and breaches of strict privacy laws such as the General Data Protection Regulation (GDPR) should there be data leaks.

Also, while data storage is seen as inexpensive nowadays, the sheer volume of data that is generated every day can accumulate along with storage costs. Companies may be spending a large proportion of their IT budget on data storage without knowing it. Reducing dark data can lead to cost savings and better use of resources.

Lastly, and most importantly, businesses who do not use dark data miss out on opportunities. Not all unstructured data should be arbitrarily deleted! Instead, software analytics and reports can reveal new streams of value from data that was previously seem as difficult to work with or of little worth. Here's how you can seize opportunities and share insights across your organisation:

a. Identify top performing agents

How do you know which agents provide the best customer service? AI and machine learning can sift through chats, emails and social posts to look for names of agents, keywords like "amazing!" and "you're the best!", and the time taken to resolve them. Once done, you can keep your teams happy and productive by recognising their efforts and sharing their best practices.

b. Analyse consumer sentiment

Instead of bothering customers to manually fill in Net Promoter Score (NPS) surveys all the time, why not use machines to analyse their interaction data and gauge how they really feel? Negative words in social media posts, using ALL CAPS in emails or typing bad language in live chat are all signs that you need a service recovery strategy now!

How to empower your teams with dark data

Agents can use analytics tools to unlock insights from dark data and help them make decisions on the spot. Unstructured data sat unused because manually transcribing call recordings, for example, was impractical. A platform like CINNOX integrates dark data from various sources, ensures data hygiene in one place, and makes it widely available to teams.

1. Assist agents to find information quickly

With dark data, agents have a sense of what it might be, but they don't know exactly where to find it or how much data there is. If your agents are spending time looking for how refunds were treated in different cases, for example, this causes them to lose focus on the task at hand and become unproductive while they search for it.

One way for agents to make sense of dark data is to use enquiry labelling. In the same scenario, past agents would have labelled it "refund" in addition to other descriptive terms such as "annual plan". This helps other agents to filter enquiries about refunds quickly and narrow them down without having to manually search for them.

Also, giving some structure to unstructured data may help companies perform root cause analysis on trending customer issues. A surge in returns for a product, for example, could be due to the product's defective design, poor instructions on how to use it, or a misleading listing description of what the product actually does.

2. Ensure agents spend time with customers

The customer experience may suffer if your agents are not spending enough time with customers. Imagine the awkward pause while agents listen to call recordings to understand what was said or being put on hold many times. This may lead some customers to form a negative impression of your brand or consider your competitors!

AI and machine learning can be used to liberate dark data and make it more accessible to your agents. Audio recordings are time-consuming to digest as you need to listen to them while taking notes. AI-powered audio transcription solves this problem by making calls searchable and easily digestible with a quick read.

3. Help agents deliver personalised experiences

A cookie-cutter approach to customer service will result in an extremely impersonal customer experience. Delivering generic experiences lumps all of your customers together without regard for their likes, dislikes, needs or goals. This can soon alienate your customers, reduce their brand loyalty and, most likely, their business!

Dark data contains nuggets of information that can be mined to personalise customer experiences. Knowing a website visitor's location, for example, can enable your agents to give a friendly greeting in the visitor's language. This extra attention to detail can influence a prospect's decision to jump to another site or stay with you.

In the near future, AI together with machine learning can help agents predict the reason for a customer enquiry and suggest a resolution to the problem. When machines comb through all customer interactions, they may even be able to perform sentiment analysis, pre-empting agents if the customer is very angry or delighted!

4. Encourage agents to work together

When agents don't collaborate because of data silos, your teams are not innovating or solving problems in new ways. They may even burn out! When a customer deals with many agents, each of whom has their own view on the issue, the result may be a frustrating and disconnected experience that is not resolved in the first interaction.

Removing data silos and putting all your data in one place is a way of unifying agents and teams. When everyone is working from one single source of data, like all the chats and emails for a customer, it eliminates any friction such as who owns the data and its accuracy. Plus, it enables teams to confidently make decisions together.

5. Enable knowledge sharing among agents

Your company will miss out on the benefits of the cross-pollination of ideas and best practices if your agents do not get enough face time together. A good part of surprising and delighting customers involves finding creative solutions and fresh perspectives instead of applying a 'one size fits all' approach to problem-solving.

As mentioned earlier, if you want to get the most out of your agents, use real-time data and reports to identify rockstar agents and learn from their successes. Once you learn what they do differently and how they WOW customers, share your findings with other agents so that everyone can benefit from new ways of doing things and earn rave reviews!

Looking at the chat history for your e-commerce website, for example, you notice that your top agent uses conversational marketing to engage prospects. Once the dialogue starts, your agent takes time to understand their pain points, sometimes jumping into a video call to demonstrate product features. The prospect is then escorted through the website right up until a product is added to their shopping cart!