Technology

Pureinsights Discovery Platform™

Search applications today (in fact any application) revolve around the emergence of modern, cloud-based architectures. This has been led by a design philosophy emphasizing loosely coupled services that are invoked when needed.  The key advantage to this approach is flexible and unconstrained access to computing resources.  These services can be in-house developed custom services, or your choice of best-in-class third party services for Artificial Intelligence (AI) powered functions like Machine Learning (ML) and Natural Language Processing (NLP).  Pureinsights has embraced this approach in the creation of a new technology platform we use to build outstanding search experiences for our clients: the Pureinsights Discovery Platform (PDP).

PDP is designed to enable the creation of a Google-like search experience, built from best-in-class components and services. It uses a modern cloud-based architecture, and incorporates data connectors, content processing, AI services, search engines and knowledge graphs. Used together these tools can provide powerful features such as: Direct Answers, Featured Snippets, ‘People Also Ask’ and Knowledge Panels. These features go way beyond what a typical enterprise search engine provides, enabling organizations to give their people the search experience they now expect. PDP enables developers to ingest massive amounts of content, then understand and enrich that unstructured content to find the key answers that users are likely to seek.  At the same time PDP provides tools to understand the intent of a user’s query so that we can deliver the most relevant, personalized and actionable search results.

Get started with the Pureinsights Discovery Platform™ today

Want to learn more about how you can make your search work like Google?

Schedule a personalized 1:1 and speak to an expert about your requirements or to learn more about the Pureinsights Discovery Platform.

A platform for building an advanced ‘Google-like’ search experience:

Pureinsights Discovery Platform™ Overview

The first piece of any search project is to gain a deeper understanding of the data being searched. People expect to be connected to those content sources that will deliver relevant answers to their search queries. But this content often resides in disparate repositories such as databases, file systems, collaborative platforms, third party applications and websites.  These sources might be in the cloud, in a private cloud or on-premise. PDP enables you to build connectors to any data source, then aggregate content from multiple sources. Data (raw data, documents, metadata, etc) is ingested scalably and efficiently, while honoring access controls. The connector then monitors the data source for additions, updates, and deletions and processes them as they occur.

Most actionable business content is unstructured and not in the best format for indexing with a search engine or ingesting into a knowledge graph. The best content is human generated (documents, presentations, social media posts, emails). These things are not generally created with search in mind. Poor quality data, especially metadata, can have a very detrimental impact on search performance. So, the next critical step in the creation of an excellent search application is to optimize that data so it can be used to answer questions effectively.  We can use our content processing pipelines as we ingest the data to clean and enrich it.

We generally recommend staging all the data you consume in a place where it can be analyzed and improved.  PDP provides a scalable staging repository for this purpose.  This staging repository has several functions. It acts as a holding area for data providing fast access when needed: for example, when publishing to a target application or re-indexing. It also makes it possible to undertake batch content processing for continuous quality improvement and testing of new content processing services, the results of which can be used to improve search engine and knowledge graph performance. Other interested applications (e.g., sales, customer, product) can also connect to the staging repository to leverage and augment data enabling an efficient ‘connect once use many’ approach.

Once the original data is staged, we can iteratively process it, so it is regularly cleaned, filtered, normalized, and enriched. We can call out to cloud-based AI services for language identification, entity extraction, metadata extraction, tagging and classification.

Processed, cleansed, and enhanced data is published to an enterprise search engine and/or knowledge graph. We call this hydration. Our platform is independent of search and knowledge graph technology, and we have built hydrators to industry leading products using our toolkit. The enriched data enables advanced search features such as featured snippets, direct answers, and knowledge panels.

The goal of any search application is to serve the users’ needs quickly and efficiently. PDP provides the tools necessary to build user experiences that meet the diverse needs of their communities. To fully close the loop, we need to establish intent, run the search and present results in a way that meets every user’s individual needs.  The platform includes a powerful Search API that developers can use to create a fully personalized search experience.  Sophisticated query parsing, Natural Language Processing (NLP) and other AI services are deployed to help decipher the user’s intent. Security is included in this API to ensure users are served only results they are allowed to see, which is crucial in the enterprise.

PDP also includes a complete React based Search User Interface that customers can deploy with minimal development effort. This UI includes Question Answering, FAQs, Extractive Answers, Knowledge Panels and all the key pieces to make your search “work like Google.”

Get started with the Pureinsights Discovery Platform™ today

Want to learn more about how you can make your search work like Google? Schedule a personalized 1:1 and speak to an expert about your requirements.

The first piece of any search project is to gain a deeper understanding of the data being searched. People expect to be connected to those content sources that will deliver relevant answers to their search queries. But this content often resides in disparate repositories such as databases, file systems, collaborative platforms, third party applications and websites.  These sources might be in the cloud, in a private cloud or on-premise. PDP enables you to build connectors to any data source, then aggregate content from multiple sources. Data (raw data, documents, metadata, etc) is ingested scalably and efficiently, while honoring access controls. The connector then monitors the data source for additions, updates, and deletions and processes them as they occur.

Most actionable business content is unstructured and not in the best format for indexing with a search engine or ingesting into a knowledge graph. The best content is human generated (documents, presentations, social media posts, emails). These things are not generally created with search in mind. Poor quality data, especially metadata, can have a very detrimental impact on search performance. So, the next critical step in the creation of an excellent search application is to optimize that data so it can be used to answer questions effectively.  We can use our content processing pipelines as we ingest the data to clean and enrich it.

We generally recommend staging all the data you consume in a place where it can be analyzed and improved.  PDP provides a scalable staging repository for this purpose.  This staging repository has several functions. It acts as a holding area for data providing fast access when needed: for example, when publishing to a target application or re-indexing. It also makes it possible to undertake batch content processing for continuous quality improvement and testing of new content processing services, the results of which can be used to improve search engine and knowledge graph performance. Other interested applications (e.g., sales, customer, product) can also connect to the staging repository to leverage and augment data enabling an efficient ‘connect once use many’ approach.

Once the original data is staged, we can iteratively process it, so it is regularly cleaned, filtered, normalized, and enriched. We can call out to cloud-based AI services for language identification, entity extraction, metadata extraction, tagging and classification.

Processed, cleansed, and enhanced data is published to an enterprise search engine and/or knowledge graph. We call this hydration. Our platform is independent of search and knowledge graph technology, and we have built hydrators to industry leading products using our toolkit. The enriched data enables advanced search features such as featured snippets, direct answers, and knowledge panels.

The goal of any search application is to serve the users’ needs quickly and efficiently. PDP provides the tools necessary to build user experiences that meet the diverse needs of their communities. To fully close the loop, we need to establish intent, run the search and present results in a way that meets every user’s individual needs.  The platform includes a powerful Search API that developers can use to create a fully personalized search experience.  Sophisticated query parsing, Natural Language Processing (NLP) and other AI services are deployed to help decipher the user’s intent. Security is included in this API to ensure users are served only results they are allowed to see, which is crucial in the enterprise.

PDP also includes a complete React based Search User Interface that customers can deploy with minimal development effort. This UI includes Question Answering, FAQs, Extractive Answers, Knowledge Panels and all the key pieces to make your search “work like Google.”