Receive AI-based predictions and recommendations in processes based on your data. Save time by automating decisions based on machine learning driven insights.
You can use our Predictive Model Qualifier to figure out if your use case is ready for AI.
Examples of Use
- Determine if an expense should be automatically approved
- Predict the likelihood of a sales opportunity to close within the next quarter
- Categorize and route an email based on text in the subject and body
- Recommend a compensation bonus amount based on employee performance data
- Suggest the best leader for a project based on project criteria
Each predictive model is based on a data table, so you will first need to have your data in a data table before following these steps to create a predictive model:
- Select Predictive Models on the left
- Select Create Predictive Model
- Name the model and select the data table
- Select the field to target for the prediction
- Remove fields that should not be included in the analysis
- Select Create Model
The predictive model may require over an hour to initially train, depending on the amount of data in the model. You will receive an email when it completes.
You can now use the predictive model within processes by adding the Field: Make a prediction action.
- The accuracy of predictive models increases with more data. 300 rows in your data table is a good minimum to aim for. On the Predictive Models page, you can see the accuracy of each.
- The field to target can be a Choose One, Integer, Decimal, or True/False type.
- When choosing fields to include in the analysis, remove fields that are not directly relevant to the prediction, fields that do not get values determined until later in the process after the prediction, as well as fields that could bias the model in dangerous ways (race, gender, age, etc.).
- When adding the predictive model to a process, you can start out by simply including the prediction value and prediction confidence in task instructions and keep all decisions with a person. As the prediction accuracy increases, you can set conditions based on the confidence to determine when the decision can be made automatically.
- The predictive model is based on a logistic regression machine learning algorithm, an example of supervised learning that takes a number of variables as input, and predicts the most likely value of a single output variable based on trends that it has learned from historical examples.
- Binary Classification, Multi-Class Classification, Regression, and Natural Language Processing are all supported within the predictive model.
- Our natural language classification is based on a bag-of-words approach using a logistic regression model. The algorithm automatically determines the set of terms that are most predictive of a given category, and calculates weights to apply to each term. Natural language features are considered along with all other input fields that are provided to the predictive model.