IBM Cloud Docs
Start getting value from your data

Start getting value from your data

Learn what IBM Watson® Discovery can do to help you find answers, recognize patterns, and gain insights from your data.

Find checklists of the high-level steps to follow to achieve the following goals:

Pinpoint answers

Help customers find answers faster. Analyze content from various connected data sources, pinpoint the most relevant passage or phrase, and return the right information when someone asks for it.

Shows the search bar with search results after it.
Figure 1. Search bar with search results

If pinpointing answers is your goal, complete the steps that are listed in the following table.

Checklist for getting answers
Step Task Related information
checkbox Create a Document Retrieval project. Creating projects
checkbox Add up to 5 collections that connect to external data sources or contain uploaded files. Creating collections
checkbox Run test queries to assess the quality of the initial results. Previewing the default query results
checkbox Take actions to improve your results. For example, you can customize the search bar to enable autocompletion. Improving your query results
checkbox Deploy your search solution. Deploying your project

Extract meaning

Use award-winning natural language processing technology to enrich your data and ensure that the right information is found when someone searches for answers.

Shows a custom machine learning facet.
Figure 2. Machine learning facet for filtering search results with custom enrichment values

If extracting meaning is your goal, complete the steps that are listed in the following table.

Checklist for extracting meaning
Step Task Related information
checkbox Create any project type. Creating projects
checkbox Add collections that connect to external data sources or contain uploaded files. Creating collections
checkbox Chunk large documents into many smaller documents so you can apply more targeted enrichments to the content. Using Smart Document Understanding
checkbox Enhance your data by applying built-in NLU enrichments. Applying prebuilt enrichments
checkbox Identify and promote terms and patterns from your data with special significance to your use case. Adding domain-specific resources
checkbox Submit test queries to assess the results. Testing your project
checkbox Create a facet that surfaces the enriched data from your documents. Facets
checkbox Take actions to improve your results. Improving your query results
checkbox Deploy your solution. Deploying your project

Enhance your chatbot

Delight your customers by fortifying your chatbot with an answer to every question. Discovery is designed to work seemlessly with watsonx Assistant to search and deliver answers from help content that you already own.

Shows a chat bot with emphasize the answer enabled.
Figure 3. Answer finding enabled in the watsonx Assistant web chat

If enhancing your chatbot is your goal, complete the steps that are listed in the following table.

Checklist for enhancing your chatbot
Step Task Related information
checkbox Create a Conversational Search project. Creating projects
checkbox Add a collection that connects to an external data source or contains uploaded files. Creating collections
checkbox Run test queries to assess the quality of the initial results. Previewing the default query results
checkbox Take actions to improve your results. Improving your query results
checkbox Connect your project to a virtual assistant that is built with Watson Assistant. Deploying your project
checkbox From the Watson Assistant user interface, deploy the web chat that is associated with your assistant. Deploying your assistant

For a more detailed look at these steps, take a tutorial that walks you through them. For more information, see Power your assistant with answers from web resources.

Alternatively, you can add a generative language service named NeuralSeek between the Watson Discovery and watsonx Assistant services. For more information, see Use NeuralSeek to return polished answers from existing help content.

Find trends

Uncover patterns, trends, and relationships in structured and unstructured data. Use text analytics to gain insights into social media, e-commerce trends, and user behavior. Or start to address problems by finding their root cause.

Only users of installed deployments (IBM Cloud Pak for Data) or Premium or Enterprise plan-managed deployments can create this type of project.

Shows the content mining application user interface.
Figure 4. Analyzing data with the Content Mining application

If finding trends in your data is your goal, complete the steps that are listed in the following table.

Checklist for finding trends
Step Task Related information
checkbox Create a Content Mining project. Creating projects
checkbox Add a collection that connects to an external data source or contains uploaded files. Creating collections
checkbox Use the built-in Content mining application to analyze your data. Analyzing your data with the content mining application

Analyze contracts

Accelerate the pace at which experts can analyze complex documents.

Only users of installed deployments (IBM Cloud Pak for Data) or Premium or Enterprise plan-managed deployments can create this type of project.

Shows the Contract Info tab for a contract document query result.
Figure 5. Analyzing contracts

If analyzing contracts is your goal, complete the steps that are listed in the following table.

Checklist for analyzing contracts
Step Task Related information
checkbox Create a Document Retrieval for Contracts project. Creating projects
checkbox Add up to 5 collections that connect to external data sources or contain uploaded files. Creating collections
checkbox Run test queries to assess the quality of the initial results. Previewing the default query results
checkbox Take actions to improve your results. Improving your query results
checkbox Analyze the data. Understanding contracts