Predictive intelligence

[Raghavendra K]

Hi Community,

i want to know where the exact use cases of the Predictive intelligence methods like Classification, similarity, clustering, regression.

i have idea how to implement the these all methods, but want to know use cases not like what is use of these methods , like where we can visible results and how to utilize in real time.

Thanks, 

Raghavendra.

[Yashsvi]

Hi @Raghavendra K,

Here are the use cases along with the solutions for each predictive intelligence method in ServiceNow:

1. Classification

Use Case: Automatically categorizing incidents into predefined categories to streamline the incident management process.

Solution:

1. Data Preparation: Ensure you have a dataset of incidents with labeled categories.

2. Create a Classification Solution:

   - Navigate to Predictive Intelligence > Solution Management > Classification

   - Click on New to create a new classification solution.

   - Select the Incident table and configure input fields (e.g., short description, description) and the output field (e.g., category).

3. Train the Model: Train the model using historical incident data with known categories.

4. Enable Predictions:

   - Integrate the model into the incident creation process.

   - Use business rules or flow designer to apply the model and automatically categorize new incidents.

 2. Similarity

Use Case: Finding similar incidents to suggest potential solutions.

Solution:

1. Data Preparation: Ensure you have a dataset of incidents with meaningful fields for similarity comparison.

2. Create a Similarity Solution:

   - Navigate to Predictive Intelligence > Solution Management> Similarity

   - Click on New to create a new similarity solution.

   - Select the Incident table and configure similarity fields (e.g., short description, description).

3. Configure Similarity Fields: Choose the fields that should be considered for similarity comparison.

4. Enable Similarity Matching:

   - Integrate the similarity model into the incident resolution process.

   - Use business rules or flow designer to provide suggestions of similar incidents to agents when a new incident is logged.

3. Clustering

Use Case: Grouping similar incidents to identify common issues.

Solution:

1. Data Preparation: Gather a dataset with fields that will be used for clustering.

2. Create a Clustering Solution:

   - Navigate to Predictive Intelligence > Solution Management > Clustering

   - Click on New to create a new clustering solution.

   - Select the Incident table and configure clustering fields (e.g., category, priority, description).

3. Configure Clustering Fields: Select the fields that will be used for clustering.

4. Run Clustering:

   - Execute the clustering solution to group similar incidents together.

5. Analyze Clusters: Use the resulting clusters to identify patterns and common issues, and take corrective actions.

4. Regression

Use Case: Predicting the time it will take to resolve an incident.

Solution:

1. Data Preparation: Collect historical data with numerical outcomes (e.g., resolution time).

2. Create a Regression Solution:

   - Navigate to Predictive Intelligence > Solution Management > Regression

   - Click on New to create a new regression solution.

   - Select the Incident table and configure input fields (e.g., category, priority, description) and the output field (e.g., resolution time).

3. Configure Regression Fields: Choose the input fields and the numerical output field to be predicted.

4. Train the Model: Train the regression model using historical data.

5. Enable Predictions:

   - Integrate the regression model into the incident management process.

   - Use business rules or flow designer to apply the model and predict resolution times for new incidents.

Thank you, please make helpful if you accept the solution.

[Raghavendra Ra2]

Hi @Yashsvi can you please connect with this number(6304547451-whatsapp) wil discuss the implementation