UiPath AI center use cases: Part 1

Benefits of AI-Enabled RPA with UiPath AI Center

By Nisarg Kadam, UiPath MVP | Senior Consultant at WonderBotz

UiPath AI Center has always surprised me with incredible out-of-the-box Machine Learning models. The AI Center doesn’t just make the models easy to deploy but also enables hassle-free management while continuously improving the machine learning models. Users can simply drag and drop ML models into a workflow and begin building strong cognitive automation. The best part of using UiPath AI Center is that the developer of this process doesn’t have to be a data scientist. All you have to do is go through the instructional documents and begin paving your way into the world of AI-enabled RPA.

In this blog, we will:

  • Focus on UiPath AI Center Text Classification Machine Learning models
  • Review Language Translation Machine Learning models
  • Glance at use cases and industry value of these models
  • Cover Text Classification which now also supports multi-language and increases the scope of ML in RPA. AI+RPA together can go beyond the expectations of RPA evolution
  • Look at how to leverage these out-of-the-box packages to grow your business with AI-Enabled RPA
  • Review use cases that we can automate with the help of UiPath’s AI Center



Before we start with the models, let us look at these three resources:


Talk to a member of our team about how our UiPath MVP’s can help you implement Ui Path AI Center in your organization.

Book a Meeting

English Text Classification for ML Models

English Text Classification is the preview version of a generic, re-trainable model for English Classification. This ML Package must be retrained if deployed without training first because the deployment will fail with an error stating that the model is not trained. This model has a deep learning architecture for language classification. It is based on RoBERTa, a self-supervised method for pre-training natural language processing systems. A GPU can be used both at serving time and training time. A GPU delivers ~5-10x improvement in speed.

Input to this model is a String: “I need new keyboard as current is malfunctioning.”

Output of this model is a category predicted: {“prediction”:”Hardware”,”confidence”:0.9422}

Let us look at possible use cases where English Text Classification will fit in:



Most organizations and industries rely on the latest news and updates about market information, such as the stock market. Keeping up with the latest news related to a particular topic and focusing on only specific domain news is always tedious.

To solve this issue, we can use UiPath’s AI Center English Text Classification model and train the model with categorical data to predict the category of news from its content.

For example, “Apple is launching iPhone 12 without charging port.” will be categorized under the technology section.

For a doctor, Technology and Bollywood news might not be relevant, therefore, categorizing and classifying news into certain categories and displays is bound to improve search speed and reduce the time spent in looking for specific domain-wise news.



We can leverage the English Text Classification model to identify a specific category of a service incident raised based on the email body. This will help reduce time consumption for standardization and make it easy for a non-technical person to raise service issues.

We can also leverage the API automation of Service Now, which will help reduce the amount of time taken for incident creation on the portal substantially thus making it easy with only a single natural language email.

The L1 support team verifies ticket details and based on the issue and priority assigns the issue to a specific resolver group. With the help of such an ML model, we can reduce the hectic and repetitive work of the L1 team for classification and ticket assignment work. It also improves the speed of processing issues and addressing clients faster which in turn improves the client experience with services and helps in client retention management.



We can also use English Text Classification to categorize web search results in order to improve personalize the client’s experience of search results.

For example, if a dancer is searching with the keyword “salsa”, he should get only search results related to salsa dance, not the recipe, and when a chef searches “salsa”, the search result should be about a salsa recipe, not the dance.

English Text Classification can be used to classify and categorize the search results we get from search engines to optimize the client’s web search results and improve the overall SEO experience.


Nowadays, the world is growing with technology. Everyone is connected via social media such as WhatsApp, Facebook, Twitter, LinkedIn, Instagram, and many others. Naturally, this leads to the generation of a large dataset on a daily basis. But, can we use this dataset for a better purpose?

We can categorize specific keywords that indicate panic situations or emergencies and during the monitoring of social media posts using English Text Classification, we can identify and immediately help in such cases. We can also handle a panic situation before it creates chaos.

This is highly sought after in the market where running such ML models on continuous live data might help many people in emergencies or panic scenarios.

For example, a person stuck on a countryside road without a vehicle, scared about their safety, might post on Facebook for help. With this algorithm, we can immediately identify such a post and the person’s location in order to send help without wasting time.



Most prominent auction houses have a digital business where people raise auction estimate requests for items they find in their backyard or old houses. If they think the items could be precious, they get them checked by the auction house regarding their value. However, these auction houses receive many price estimation requests daily, making it difficult for them to reply on time with the correct estimated value to the client.

In such a process, auction houses sometimes lose a valuable customer who doesn’t get a timely response.

To tackle this situation, we can use English Text Classification to help auction houses categorize estimate requests based on priority and make the whole process more efficient. This process improves business outcomes and customer satisfaction with fast response.


With the right trained data, we can identify fraudulent emails and alert the users before they click on any links or make transactions.

In this era of e-commerce, people prefer online payments over paying in cash, and in such scenarios, they tend to save their payment details in cache memory. Hackers can send click baits or phishing emails to get all of your saved credentials in their hands with a click, resulting in a significant loss or leading to additional hacking scenarios. To avoid clicking on any suspicious emails or links, we can use the English Text Classification model to train the data and warn the user to not click on such suspicious links.

This might result in the reduction of fraud and hacking cases across the globe.

UiPath AI center icon for WonderBotz
Learn more about how UiPath Ai Center works here

UiPath AI Center use case – Language Translation  

AI Center consists of four different language translator out-of-the-box models:

  • English to French
  • English to German
  • English to Russian
  • German to English

Input to this model is a string of the first language: “My name is Jon.”

Output of this model is a string of translated language: “Mein Name ist Jon.”

Let us look at the possible use cases where Language Translation will fit in:


As businesses become global, the first challenge every marketing and salesperson faces is communicating with international clients.

Suppose an Indian sales guy, Sumit, wants to strike a deal with a French Businessman, who hypothetically does not understand English.

Now, Sumit is having difficulty communicating with the French Businessman which is impacting his business.

In such a situation, we can use UiPath’s English to French ML model to convey Sumit’s message to the Businessman and help Sumit close an international deal without learning French. How cool is that?

Similarly, we can resolve many such issues of international communications with language translation models.


Most people have to travel around the globe as part of their jobs, to see family, or just as a hobby. The most common issue for an international traveler is communication and understanding written information.

For example, an English person in a German country has to sign some medical form for an emergency, but he doesn’t understand the local language and cannot read the documentation. However, with the help of UiPath Document Understanding and AI Center, we can create a model which will help him extract written information from the medical form and understand the rules and norms in English. Thus, it makes it easy for him to tackle such emergencies and survive in a foreign country.

We can think about many such use cases that can make an international traveler’s life easier such as processing information shared at historical places, reading a restaurant menu or driving instructions, and understanding company documents and business emails.



Many countries have old books about the culture and history written in their native language. We can leverage UiPath language translator OOB Models to help translate these books into any specific language and gain knowledge without learning that language. This model increases the scope of sharing science, history, culture, political and economic information with everyone across the globe.

Conclusion for UiPath AI Center Use Cases

I hope this blog was able to acquaint you with the world of AI + RPA and shed some light upon its possible use cases for your organization to automate the right and first processes with the help of AI Center Machine Learning Model.


Case Studies