Agent Intelligence in London Release

[gauravchoudhury]

The following article is a consolidation of multiple documentations and release notes with the intent to summarize and help community members seeking information about Agent Intelligence in the London Release.

ServiceNow Agent Intelligence product enables us to use machine-learning algorithms to set field values during record creation. For example, setting the incident category based on the short description. 

One can train predictive models, so they act as an agent to automatically categorize and route work based on our past record handling experience. 

Agent Intelligence enables us to manage high volumes of request at lower costs. Automating the categorization and assignment of requests gives us some benefits, namely: 

• Reduce task resolution times
• Reduce the number of interactions required to resolve tasks
• Reduce the error rates of categorizing and assigning work 

Using predictive models in training solutions: By default the system offers these predictive model solutions for,   

• Incident categorization
• Incident prioritization
• Task assignment 

However, you can also extend Agent intelligence to other processes by creating your own predictive models and training them on your past record data. 

Now in London release, you can use the default Agent Intelligence records as templates to create HR Service Delivery Solutions. 

You can also review additional guidance for solution training, such as tips on conditions you may want to set on the data records you want to train and the recommended number of records for training a good solution. 

One can even, exclude a class from solution training to prevent the model from ever making predictions for a particular output field class. For example, you can exclude a particular incident category from training if you plan to retire or change the category. 

A solution is only good as the record data you use to train it. Hence, operational data should be well categorized and maintained (for ideal results!). 

Now, you can refine and test your machine-learning (ML) solutions iteratively on a non-production instance, and then use update sets to export the changes to your production instance. By following this process, you reduce the risk of having to retrain solutions on your production instance. 

In the Kingston release, the system used the top 25 classes with the highest number of records when it built a solution. In this release, it uses the top 50 classes, increasing the amount of historical data it processes to predict the class confidently. 

As you create separate solution definition records for each predictive model you want to support, you can copy a record and its configuration into a new form by clicking Copy Solution Definition from the context menu of the original record. In London release, your options include Dutch, English, French, German, Japanese, and Spanish, with English as the default. 

Once your Machine Learning (ML) solutions are trained, you can call on the Agent Intelligence API to make a solution prediction. 

When you request a clone of your instance, the system stores your trained solution components as attachment records. Adjust your glide.platform_ml.clone_artifacts system property to preserve these records during the system clone. 

Agent Intelligence is included in the following packages: ITSM Professional, CSM Professional, HR Professional, and HR Enterprise. When you subscribe to any of these packages, you can activate the Agent Intelligence plugins in your production instance using your administration rights. You can also request activation of these plugins on a non-production instance. 

If you are running your instance on an earlier version than Kingston Patch 2, such as Kingston Patch 0 or Kingston Patch 1, when you activate your plugins, you must update your scheduler URL property, so it is compatible with Agent Intelligence. 

To activate Agent Intelligence in your platform, you need these two plugins on your production instance using your administration rights: Agent Intelligence plugin (com.glide.platform_ml) and Agent Intelligence Reports plugin (com.glide.platform_ml_pa).

Other notable mentions… 

The system uses the following business rules for Agent Intelligence, 

Default Task Based Prediction: A business rule that runs before inserting new task records to make a field value prediction based on the solution definition output field and the solution dataset. Use this business rule as a template to create your own prediction logic. This business rule calls the Agent Intelligence API. 

Update Prediction Results: A business rule that runs before closing task records to update the solution statistics with the actual precision and coverage results. 

You can track the coverage and precision of each predictive model using the Solution Statistics dashboard. 

Bonus Feature: A machine learning solution definition can be configured in Agent Intelligence per domain, which ensures that the data used by the solution is domain-specific data. After a solution is trained for a domain, the solution calls for a prediction to be made for resolution, depending on what that domain belongs to. For more information, click here.

Source: https://docs.servicenow.com/

[Gajanan Birada1]

thanks for sharing valuable information about machine learning.

could you please tell about how to activate agent intelligent on personal instance what is the process.

[gauravchoudhury]

For now, Agent Intelligence plugin is not available for PDIs. Agent Intelligence is included in the following packages: ITSM Professional, CSM Professional, HR Professional, and HR Enterprise. When you subscribe to any of these packages, you can activate the Agent Intelligence plugins in your production instance using your administration rights.

Once the above condition is fulfilled, you can activate Agent Intelligence on your production or non-production instance using your administration rights: Agent Intelligence plugin (com.glide.platform_ml) and Agent Intelligence Reports plugin (com.glide.platform_ml_pa).

Only in authorized contracts with ServiceNow (on an Enterprise Level), you could be provided with this. At this point of time, it is not available for all.