8 Mar, 2021 • 5 min readArtificial Intelligence Natural Language Processing Featured Machine Learning Communications Mining
What is Communications Mining?
You have probably heard of Process Mining? You may even have heard of Task mining? So why is it important that you know what Communication Mining is?
The average employee in the enterprise sends or receives 126 emails each day, in the enterprise up to 40% of an employees day is spent just in Outlook.
Imagine being able to mine the useful information in even a handful of these emails, finding unwritten processes, data points and uncovering insights at scale, this is what Communication Mining enables.
Couple this with customer communications across email, online chat, and even voice, and suddenly mammoth amounts of previously unstructured and untouched data becomes structured and it’s insights accessible, this is Communication Mining.
The glue that holds our processes and procedures together is the communications we send internally between employees, and externally between our customers and employees.
These communications range from email, tickets, to instant messages, online chats, notes in our CRM/ERP systems, the telephone conversations we have. In short, anywhere we converse in natural language.
The Communications Mining Process
Mining this unstructured text based data for insights is the next frontier in Digital and Business Transformation. Businesses have spent decades, documenting, and refining the “do” of their processes without really understanding the drivers of why the process needed to be completed in the first place, who is asking, and more importantly “why”.
That’s where Communications Mining starts to bring large and incremental value to the enterprise, helping you understand what people are saying, how frequently they are saying it and in what context. When these newly mined communication insights are coupled with process understanding, a whole new avenue of transformational opportunity presents itself.
- Firstly, common themes and semantic intents can be grouped together to provide a holistic view across any communication channel (including Voice when speech to text translation is used.) This is a first for the enterprise, historically siloed insight capabilities would have looked at channel specific communications, almost entirely focused on customer communications, and nearly always from the Voice channel.
- Secondly, making use of Artificial Intelligence (AI), takes the heavy lifting out of categorization of all these communications, and provides untold value in the volume and quality of the analytics, this is all within weeks, not months or years. Training can be real-time and should be underpinned by a critical approach in Machine Learning - “Active learning.” This creates a dynamic feedback loop between man & machine. Meaning that as a new communication enters the business, the AI instantly starts making predictions across both historical and new in-coming comms. The agent can simply label new concepts on the fly and continue with their day job, The system is constantly learning, re-training, re:inferring in the background. Training the AI to categorize like your best agent, means something else (AI in this instance) does the heavy lifting, post training; and you can focus on the improvements you need to make in your operations vs. drowning in unstructured data.
- Thirdly, using the same analytic capability, you can prove your improvements have actually worked through the use of the same data that highlighted your improvement areas in the first place. You should see an instant reduction in these queries/requests entering the business “if” you have made the correct change, or see data to support the funneling of these requests into your automations. The key here is having data to prove the value you created.
A history of Communications Mining
Of course Communications Mining on it’s own is only part of the solution to business wide transformation. It covers the text based aspects of an organization, however on it’s own it produces actionable insights, but is not an Automation tool itself, like Robotic Process Automation (RPA), or a Chat Bot.
Communications Mining is also not the only analytical tool you may need in your organization. Both Process Mining & Task Mining are complementary tools that will assist with your understanding of the “Do”.
When Communications Mining, Task Mining, and Process Mining are linked in with Automation Technologies, and the right methodologies, the concept of Hyper-Automation becomes a reality.
Why Communications Mining is important
Communications mining is important because traditional means of looking at processes generally fail to capture the reason/customer intent for any actions needing to be completed.
This is a key element historically missing from most transformation programs, which will therefore also fail to accurately capture any changes (positive or negative) made post transformation, either through sentiment shifts, volume fluctuations or reduced customer/agent effort.
When looking to understand if your organization has a requirement for Communications Mining, ask yourself these 5 things
- Do we currently analyze communications across our business as a means to improve processes for our employees and customers?
- Do we currently understand at a data level, who received, when and actioned the requests and the reason/intent for these communications?
- Do we currently truly understand what our customers are asking of us?
- Do we have a current understanding of all those undocumented processes that happen via emails across the organization currently?
- Have we directly fed our Automation capabilities with structured outputs from unstructured data sources such as emails?
It's highly likely that your answers “no” to at least 3 out of 5 of the above, and so extremely probable that Communications Mining will add value to your organization, either creating new customer insight, process awareness and improvement opportunity and of course areas of interest from an automation perspective.
Who is responsible for Communications Mining?
One of the challenges with Communications Mining to date has been internal ownership of the capability. Should it sit in Customer Experience, Data & Analytics, Marketing, Automation, Process Improvement, Operations or some other department in the Enterprise.
Unfortunately, this isn’t a one size fits all, but the growing trend is that this type of capability sits in close quarters to Customer Experience, Process Improvement or Automation. Much will depend on your organization, and who benefits most from the value that communication mining creates as to where it is best deployed in the organization.
The Outcomes and Business Value of Communications Mining
So what should you expect to achieve through Communications Mining? It’s always difficult to put a value on knowledge however there are a number of indicators and metrics that are being used by the earlier adopters to showcase the value the technology is bringing.
- Categorize the query types, into buckets of Process Improvement, Automation or Elimination, this is your potential opportunity for contact removal, which can be calculated as a value.
One early adopter, found an opportunity in over 40% of the emails that they received to either improve a process, add additional content in their digital self-serve, or provide a level of email automation all of which lead to a potential 25% reduction in the manned effort across their email function.
- Analyze all data points associated with your communications, not just what was said, but when it was said, by who, and even the sentiment on the response?
The insights associated with this meta-data, can be invaluable when looking at your internal staffing quality (are any particular staff members taking more responses to resolve a query), resource profiles (when are queries being received, what are the response times and are our staff shift patterns optimal).
Various users have reported benefits in their ability to spot staff training and quality assurance needs, leading to better quality outcomes for customers in the future.
- Similar to the outputs of Process Mining, the data associated with the path your “processes” takes will lead to far more variation than you ever expect.
This data driven approach over the top of your communications will tell you far more about your work than you are ever likely to achieve through manual reviews.
It is difficult to put a value on the unknown, but understanding undocumented, sometimes unknown and un-official processes, has created incremental value, beyond expectations in those earlier users.
The Future of Communications Mining
Communications Mining is going to be so important to the data driven enterprise. Understanding what is happening, outside, as well as inside your processes through this technology which will create both new insights and alternative gateway to process transformation, uniquely concentrating on both customer and employee communications to showcase where change is needed.
The early pioneers are already experiencing monumental levels of business value creation, and gaining insights that would have taken potentially years to obtain in just a few weeks though the use of AI.
Communications Mining explained in 90 seconds