Client story: Regulatory authority
Reducing time spent on non-value-adding activities by 80%
Allowing employees to allocate more time to valuable tasks by automating email routing using Natural Language Processing
A regulatory authority wanted to automate time-consuming routines for their employees. Together with 2021.AI, the regulatory authority have applied Natural Language Processing (NLP) to reduce time spent by the front desk on emails, improving the organization’s workflow without compromising how well they handle emails.
Download the case
Download the PDF-version of “Reducing time spent on non-value-adding activities by 80%” client story
The challenge
Processing emails is very time-consuming. For this regulatory authority, they estimated that the front desk routed an average of 100 emails per day, which delayed the organization’s workflow.
The lengthy process of reading, writing, and tagging emails can be significantly simplified through automatic processing, allowing employees to focus on more valuable tasks.
Developing an AI model that can learn the rules of routing emails may significantly reduce the time spent handling this manually. By applying an AI model to redirect emails to their proper recipients, the authority can reduce errors that may be made through manual routing.
··········
There is great potential to deploy new and disruptive technologies, such as AI, to save one of our most valuable commodities, time. Increasing efficiency and productivity bring not only value to public organizations, but value to employees’ wellbeing, benefitting society as a whole.
The solution
This project aims to build an email routing model that can accurately predict an incoming email’s recipient. It is a sophisticated and modern AI solution by which NLP learns how to classify emails into different groups. In this case, five main inboxes have been identified, each dealing with a different topic.
The model is trained on historical data to enhance its ability to predict an email’s destination. An essential aspect of this model is that its underlying algorithm learns the numerical structures of a text rather than its linguistic ones. It first removes linguistic noise and reduces text features, preserving only the useful information using the following methods:
- Lowercasing: Making all uppercase characters lowercase
- Tokenization: Splitting sentences into single words/tokens
- Removing punctuation
- Removing stopwords: Deleting frequent words, such as ‘this’
- Removing numeric characters
- Stemming or lemmatization: Extracting the roots of verbs
The text is then transformed into a term frequency-inverse document frequency (TF-IDF) numerical format, where each column relates to a specific word or sentence, reflecting its importance. Different mathematical algorithms are then trained and the model that performs best is selected to route front desk emails. Its benefits include increased productivity, faster case response times, and an increase in customer service quality.
The results
By implementing AI, front desk employees have been able to increase efficiency and productivity, as emails are automatically routed without compromising the accuracy and how well they are handled.The lengthy process of manually routing 100 emails per day has been reduced significantly, with 84% of emails automatically routed. Our AI model reduces the time spent by employees on emails, freeing up time to work on more important and value-creating tasks.
Project highlights
- The model deals with up to 100 emails a day, freeing large amounts of employee time
- Our AI model sorts and routes 84% of incoming emails with 75% accuracy
- By implementing an AI model, the authority has reduced 80% of the time spent on emails
About the company
The regulatory authority are an ambitious and modern authority. They employ around 1,000 people working to create the best framework for companies’ growth and development throughout the nation.