Fill the technology gap
At the Search & Content Analytics practice of Accenture Applied Intelligence, we focus on building end-to-end, AI-driven search and analytics solutions that help power the most intelligent enterprises.
Over the years and through the experience of hundreds of projects, we have seen various technical challenges that available commercial or open source platforms cannot solve. To address these issues, we have developed a collection of search engine independent technology assets to fill in the gaps in clients' search, analytics, and natural language processing (NLP) projects.
These assets were created by our engineers during live implementations and have evolved through numerous subsequent projects where they were deployed and optimized, with substantial input from our clients. By eliminating the need to reinvent the wheel for some repeatable use cases, such as enterprise search, intelligent document understanding, storage optimization analytics, etc., our technology assets can accelerate project timelines and increase speed-to-value.
Accelerate results with the Applied Intelligence Platform (AIP)
These technology assets are available standalone or as options on the Applied Intelligence Platform (AIP).
AIP is a comprehensive and scalable solution that allows organizations to get actionable insights and business outcomes, quickly, with a competitive flexible commercial model. Custom solutions can also be created to meet your organization’s needs. Each of AIP’s flexible and agile analytics apps manages the complete end-to-end process, providing organizations with immediate access to the tools needed to make data-driven and intelligent decisions. Together, AIP and Aspire framework bring a powerful technology stack that modernizes the acquisition, enrichment, analysis, and visualization of enterprise unstructured and structured data.
- Create an end-to-end search and analytics system in days, not weeks, and bring immediate business value
- Deployable in the cloud or on-premises
- Provide a risk-free approach to migrating between search platforms, such as replacing the Google Search Appliance, migrating from FAST ESP to SharePoint, moving from Solr to Elasticsearch, etc.
- Fill in the gaps in open source and commercial search engine platforms using search engine independent, complementary technology assets
Our technology assets work with a range of search engines, including SharePoint Search, Azure Search, Google Cloud Search, Amazon CloudSearch, Solr, Elasticsearch, and others. Each component can be deployed individually or together to complement and optimize your organization's search architecture.
- Aspire Content Processing: a content processing framework that is now integrated into the Applied Intelligence Platform (AIP) and designed specifically for unstructured and semi-structured data. Aspire Content Processing is commonly used in our consulting engagements; it provides optimal functionality, a wide range of ready-made processing components, a Hadoop implementation, and distributed processing capabilities.
- Content Connectors: search engine independent connectors with built-in early-binding security and metadata capture capabilities.
- Staging Repository: an intermediate repository where content can be placed after it has been extracted from a source. This staging repository allows for more efficient content reprocessing without having to reach back to the content source for every processing iteration.
- Query Processing Language (QPL): a query parsing and business rules engine, enabling sophisticated query-side processing to be set up and maintained efficiently. QPL is often deployed with the Search API Server, but also available separately in some implementations, depending on the client's requirements.
- Search API Server: allows new endpoints to be configured in seconds. These endpoints are backed by scripts that can simply pass the incoming queries to a search engine, perform query manipulation using QPL to increase relevancy, or perform other actions such as database lookups or updates. Results are then amalgamated into a single response returned to the Search UI.
- Search UI: an end-user search interface with full source code and can be customized for specific requirements. The use of the API Server and QPL allows the Search UI to be search engine agnostic.
- Admin UI: pluggable admin interface for installation, administration, server management, and health check.
- In addition to the above assets that’ve been deployed in hundreds of client projects, we continue to innovate in the NLP/NLU area with the development of Saga Natural Language Understanding Framework. A new R&D initiative, Saga enables enterprises to create and maintain powerful, flexible, tested, and scalable enterprise language models for user interaction and document understanding. It incorporates many language modeling techniques and machine learning into a single, user-friendly semantic framework to handle a wide variety of natural language use cases.
As we have expanded the range of our technology assets over the years, we have also developed an optimized reference architecture for creating scalable and customizable search, analytics, and natural language processing applications. Built around our technology assets, open source search engines, and other complementary technologies, this proven reference architecture has enabled our clients to create working systems more quickly and gain business value sooner.
The diagram below shows the reference architecture for a browser-based search application.
- In this example, the search application needs to access a number of disparate content sources (e.g., content management systems, text documents, e-mails, image repositories, social media sites, etc.)
- Connectors acquire data from external sources. Aspire Content Processing (or some other content publishing engine) can then do the heavy lifting to prepare the content for indexing by the search engine.
- In some deployments, the Staging Repository can be a buffer between front-end load and back-end publishing from the data sources.
- In this diagram, the Search API Server and QPL are deployed together to enable sophisticated query-side processing, execute scripts in "sequential parallel” fashion, and serve relevant results to the Search UI. In some cases, QPL is available separately based on the client's implementation needs.
- Search UI provides basic or custom templates for most search use cases, including e-commerce, corporate-wide search, data warehouse analytics, media & publishing, recruiting and many others.
- Admin UI is a central, customizable dashboard that allows system admins to holistically configure, manage and monitor all system components.
- Saga NLU provides automated pipeline construction, state-of-the-art handling of language ambiguity, integrated machine learning, and business-friendly user interfaces for creating and maintaining NLU models.
- When application requirements call for text mining, machine learning, semantic analysis or quality metrics, it may be necessary to deploy Aspire in a big data array as part of a Big Data Framework (Hadoop). What may not seem like a big data job (millions of documents) can quickly become one when advanced text mining is required (billions of words, phrases, and semantic relationships).
Implementation and support
We provide a full range of consulting, implementation and support services for applications deploying our technology assets. When appropriate to the solution architecture, these technology assets facilitate the efficient delivery and support of custom search solutions, and our clients benefit from increased reliability. We propose using these assets only where there are clear and specific benefits to the application. After a thorough assessment and consultation with our technical team, clients can make the decision whether to leverage our technology assets in their projects.
Contact us to see how these technology assets can help optimize your search, analytics and natural language processing applications.