Configure machine learning models

This article explains how a ServiceNow instance administrator or developer who has access to the Coveo organization linked to their instance can configure Coveo Machine Learning (Coveo ML) models in their query pipelines linked to Coveo for ServiceNow widgets.

Configure Coveo ML models

The Coveo for ServiceNow widgets automatically log Coveo Usage Analytics (Coveo UA) events. Once sufficient usage analytics data has been gathered, this data can be leveraged by Coveo Machine Learning models to provide highly relevant AI-powered recommendations.

You can create and then associate Coveo ML models to your query pipeline to ensure the most relevant content is always shown in your Coveo for ServiceNow widgets.

  1. On the Coveo Administration Console Models (platform-ca | platform-eu | platform-au) page, create an Automatic Relevance Tuning (ART) model.

    ART models use Intelligent Term Detection (ITD) to automatically refine queries with additional contextual information (for example, important keywords from a large textual case description). They also ensure that items which are deemed highly relevant are included in the query result set and have a fairly high ranking score value, even if those items don’t actually match the original query.

  2. Associate the ART model with the desired query pipeline. While doing so, on the Associate a Model subpage, under Advanced Configuration, check the Comply with Intelligent Term Detection (ITD) box.

  3. Create a Query Suggestion (QS) model.

    QS models provide a list of relevant query completion suggestions as the end user is typing in a search box. They can also make the search-as-you-type feature more powerful and reliable.

  4. Associate the QS model with the desired query pipeline.

  5. Optionally, evaluate how well your Coveo ML models are performing before routing a larger part of queries to the query pipeline in which they’re defined. The Model Testing (platform-ca | platform-eu | platform-au) page provides a way to quickly test ART and QS models.

    Alternatively, you can configure an A/B test to test your new query pipeline, which now contains your Coveo ML models, against the default, original query pipeline.

What’s next?

  • You may want to learn how to generate usage analytics reports to help you identify possible content gaps, discover how your Coveo for ServiceNow widgets are being used by your customers and support agents, etc. (see Usage analytics reports).

  • If you require additional specific usage analytics data, you should learn how to log custom usage analytics events through the Coveo JavaScript Search Framework (see Send your own UA events).