Artificial Intelligence Bizagi, Machine Learning, Process Automation

Artificial intelligence and process automation are intrinsically linked by the goal of empowering businesses to make smarter, faster decisions.

As organizations strive to digitally transform, many are exploring how AI could enhance the power of process automation programs by intelligently managing workflows according to continuous data insights.

To get an expert view of Bizagi’s AI capabilities, I recently interviewed my colleague Andrea to hear from her what Bizagi 11.1 is all about.

[Interview with Whole Product Manager – Andrea Dominquez ]

Andrea Dominguez Alarcon Bizagi

TB: So what’s exciting about Bizagi 11.1?

AD: Along with other exciting new features like live processes, our latest software release (Bizagi 11.1) introduces Artificial Intelligence capabilities. These features are the first step towards our vision for the future of AI technology in process automation.

TB: Everyone seems to be talking about AI, what’s Bizagi’s approach?

AD: Bizagi’s Artificial Intelligence allows you to leverage Machine Learning capabilities directly in your Bizagi processes. You can do this by setting processes to rely on a predictive analysis service of the reliable data that you have in a Bizagi Dataset.

 

 

Through predictive analysis, you can train Machine Learning models to carry out experiments that rely on algorithms and predict an outcome based on all historical data (with a given degree of accuracy depending on the experiment and the amount / quality of data).

You can easily connect your Bizagi processes to such models and rely on the predictive analytics to recommend the best business decision.

Artificial Intelligence Bizagi

Importantly, it’s possible to configure all of this without necessarily being an expert data scientist. But for those that understand the terminology, the Machine Learning algorithms that Bizagi Artificial Intelligence uses include:

  • Decision tree (C45)
  • Decision tree (ID3)
  • Linear SVM Classifier
  • Multiple Linear Regression
  • Logistic Regression

Bizagi will choose the best algorithm for your specific use, though you may choose to explicitly select which algorithm you wish to employ. For instance, when predicting attributes / variables of a categorical type, the Logistic Regression algorithm is often used. Similarly, Multiple Linear Regression would be commonly applied to continuous attributes / variables.

TB: Can you give some examples of how these features could be used?

AD: Many organizations are hunting for ways to make use of Artificial Intelligence in their own businesses, so let’s look at a few examples:

  • A financial services company is looking to use AI as part of a credit card application process. When a customer requests a loan, based on the amount of credit, customer type and credit type, they want to predict beforehand if the process requires additional documentation or not (i.e. predicting a categorical variable).
  • An insurance organization is looking at using AI as part of its underwriting process. When insurance is to be underwritten, based on the insurance amount, customer type and estimated price, they want to predict an estimated duration to approve the insurance underwriting (i.e. predicting a continuous variable).
  • A helpdesk team is looking to use AI as part of its ticket process. Based on the severity, operating system, ticket type the team would like to predict if the ticket would be likely to be solved by first-level support, or if it needs to be escalated to second-level support or third-level support (i.e. predicting a categorical variable)

Rather than using fixed business rules, the intelligence of these processes will continuously learn from input data and recorded outcomes, improving the accuracy of Bizagi’s predictions.

TB: People often mention IoT along with AI, do the two link here?

AD: Bizagi’s digital platform is renowned for its integration capabilities, and along with connecting to business technologies like SAP and SharePoint, connecting devices and “things” is a huge part of this.

The key here is that real-time inputs from such devices, whether it is automated data inputs from an oil well, or human input into a mobile device in a warehouse / hospital – all of that data is immediately fed into the datasets that the artificial intelligence is drawing on. It goes without saying that the system can only be intelligent when it has the right data to make the right decision, so the link with IoT is a clear one.

If you would like to learn more about our latest Artificial Intelligence and Machine learning features, please contact us, or watch our full platform demo.