Machine learning plays a pivotal role in modern business success. With machine learning, businesses can optimize their site’s search functionality to deliver more relevant, personalized results, enhancing user experience and driving engagement. At Focalcxm, we have worked on multiple projects using Coveo’s Machine Learning models (Coveo ML). Coveo ML models are a critical part of their service, employing artificial intelligence to boost relevance and provide contextually relevant recommendations. These models leverage user analytics (UA) data to predict and recommend content that is most useful to users. This blog will outline Coveo’s models and provide details on how we have set them up for our customers.
The Coveo ML models are created and trained using algorithmic models, which use UA data to fine-tune their predictions. This allows Coveo to continually improve its search results based on real-time user behavior and preferences. Once a Coveo ML model has been created, it must be associated with a query pipeline to be effective in a search interface. This ensures that the most relevant content is always shown in response to user queries.
Coveo’s ML Models:
Coveo Machine Learning (Coveo ML) utilizes several machine learning models to enhance user experience, each tailored for specific tasks:
- Automatic Relevance Tuning (ART): This model uses machine learning to analyze user behavior and improve the relevance of search results. It takes into account past user interactions and adjusts the ranking of search results accordingly.
- Query Suggestions (QS): The QS model provides users with intelligent query suggestions as they type in the search bar. It helps users form effective queries, aiding them in finding relevant results faster.
- Content Recommendations (CR): The CR model recommends content that is likely to be of interest to the user based on their past behavior and the context of their current session. These could be articles, products, or any other types of content that the user may find useful.
- Dynamic Navigation Experience (DNE): This model optimizes the display of facets and facet values in a search interface based on the user’s context and behavior. It aids users in filtering and navigating search results more effectively.
- Predictive Merchandising (PM): Specifically designed for e-commerce applications, this model analyzes user behavior to predict which products a user is likely to be interested in. It then promotes these products in search results.
Each of these models plays a distinct role in enhancing the user experience, making Coveo’s search and recommendation capabilities more accurate, personalized, and efficient.
ML Model Setup:
Creating the ML models that we want to leverage is straightforward with the Coveo Cloud platform. Here are the steps we used to create a Query Suggestions model
- Navigate to the Machine Learning Tab: On the left-hand side navigation panel, click on the ‘Machine Learning’ tab.
- Create a New Model: Click on the ‘Add Model’ button and select ‘Query Suggestions’.
- Configure Model: Here are the key configurations:
-
- Model Name: Provide a name for your model.
- Associated Query Pipeline: Select the query pipeline that you want to associate with the Query Suggestions
- Data Period: Define the period for which the user analytics data should be considered for model training.
- Save the Model: Once you’ve configured your model, click on the ‘Save’ button to save it. The model will now start learning from the usage analytics data.
Final Thoughts:
Coveo’s machine learning models are straightforward to set up and ultimately empower users to find the results they are looking for. Some considerations:
- The site is ready for machine learning as of Day 1 of the project go-live.
- The more events machine learning has to learn from, the better it will be at providing relevant results.
- Machine learning is best when it learns from 25,000 – 100,000 queries.
- The amount of time it takes to improve the relevance therefore depends on the level of search activity on the site.
The role of AI and Machine Learning in Enterprise Search is not only significant but also transformative. These technologies are changing the way we interact with data, making it more accessible, relevant, and personalized. Coveo’s use of AI and Machine Learning in its Enterprise Search platform uses powerful technology that will continue to shape the future of Enterprise Search.