A challenge in the public sector is handling unlimited requests with limited resources. If your organization has maxed-out requests, you can save time and resources by managing your inbound requests with natural language processing (NLP).
What is natural language processing?
Picture this: you have an event coming up, and lots of questions are coming in. There are requests from members, donors, event attendees, journalists, and the general public. Some people want to know what’s on the event schedule. Others are curious about your giving levels. Others tried to buy a ticket online, but aren’t sure their transactions went through. Some ask if it’s possible to add a ticket to an existing order, while others need to cancel. As you can see and probably know firsthand, this scenario can easily turn into an overwhelming inbox.
Natural language processing is an excellent tool to help you manage the situation. First, let’s define a few key terms:
- Natural language processing, commonly called NLP, is one branch of data science that specifically works with text data. NLP allows us to transform language into data a computer can understand, then train a computer (or model) to take in text and return something that is useful to us, like the topic of an email.
- Data science is an interdisciplinary field that can cover many topics. For our purposes, think of it as a mix of computer science, statistics, and linguistics used for solving problems.
- Data refers to any kind of information that can be collected, accessed, and analyzed. While NLP makes use of text, data can also refer to numbers, audio, and images.
How can this help us put on a smooth event by managing your inbox? Let’s dig in!
How natural language processing can help your inbox
Let’s say an email or contact form like this comes in:
Hi. I need to change the name of one of my guests. Mary can’t come, so I invited Joanne.
If you were to tag this message by hand, you might use “Event Team” to signify who will handle it, “Guest List” to identify the main topic, and “Urgent” because the event is less than 12 hours away. You can pick out these themes quickly in part because you’ve seen lots of these requests before. While computers can’t understand language like we can, they are much faster at processing information. With NLP, you can train a computer to tag new text based on how you tagged texts in the past. You can also ask a computer to pick out new themes from a group of emails or other text.
In the time it takes for you to read one email, a computer can process and tag thousands. Once you have trained a model to tag or theme messages with accuracy, you can put other rules in place. For example, you might forward messages to a teammate or folder when certain criteria are met. You could also store the most pertinent information from all emails in a single spreadsheet.
The details of your implementation will depend on your specific needs and how inbound requests make it to your organization. Custom work will likely be needed to set up a business intelligence stack where you can export and analyze messages. If you use Gmail, for example, you might do analysis in BigQuery, extracting raw email data and then applying a model to it. If you are drowning in emails, developing a system with NLP will be well worth the investment, saving you time and simplifying your outgoing communications.
How to know if NLP can help your organization
NLP has a robust set of capabilities for analyzing and modeling text. If you’re not sure whether NLP as described above may help your organization, here are some questions to ask yourself:
- Is my organization able to respond to incoming requests, particularly urgent requests, quickly?
- Does someone on my team spend a lot of time reading and synthesizing information?
- If we had unlimited time and resources, what else would we read to get a better sense of how we’re serving our community?
- Where can we improve our operations by categorizing long-form text into predefined or unknown topics?
Data science is well-known as the “brain” behind some of the most ubiquitous technologies, such as movie and product recommendations. NLP provides the foundation for working with text. Other high-impact use cases include social listening tools, content personalization, sentiment analysis, chatbots, and text generation.
The public sector can use NLP to manage incoming requests and more to increase efficiency and further important causes. When you have more data than you can easily synthesize, data science provides excellent solutions. Taking time to automate the tedious parts of your operations now will free you and your team up for more complex tasks in the future. If you can ease your workload with automated text tagging, theming, or analysis, consider using NLP.