Global Non-profit uses Semantic AI to Improve SharePoint Search
Industries:
Associations & Non-ProfitProducts:
SemaphoreFull Story
Challenge
The organization’s structure was widely distributed. Document repositories were spread across thousands of SharePoint sites throughout the world. Global programs, which are created and implemented in the field, made it difficult to facilitate communication between home and field offices as well as between field offices.
They partnered with Semaphore to create a pilot that supported a two-fold vision
- Central office should be able to easily see new key document types that are being produced in the field.
- Field office staff can use search/dashboard capabilities to view all document types that they are interested in or are associated with their area of responsibility.
The pilot focused on few high-priority document types that Semaphore staff and internal subject matter experts used to demonstrate an improved user experience and knowledge discovery in the global SharePoint Online environment.
Solution
The team began by using Semaphore’s Knowledge Model Management (KMM) capabilities to collaboratively build a model for select document types like case studies, strategies and program monitoring notes, and other reports. Using text mining capabilities and the Lexical Enrichment Side Panel (LEX), sample document types are analyzed and used as part of the model development process. The resultant model, built and validated by internal subject matter experts, accurately identifies the relevant concepts, topics, and relationships associated with each document type.
Semaphore Classification and Language Services (CLS) leverage rules automatically derived from the semantic model to iteratively develop a classification strategy that examines and tags each document with rich metadata. The document structure - i.e., headers, titles, footnotes, are identified and used to create additional rules that enhance the ability to identify and tag the context and meaning found within each document.
Using Semaphore’s SharePoint Online integration pack, the semantic metadata is fed into SharePoint and its native capabilities are leveraged to improve user search and knowledge discovery.
Result
With a full implementation of the pilot, users will have a precise and consistent knowledge discovery experience that allows them to locate all information associated with a particular document type regardless of where it resides within the organization.
Document dashboards provide users with a single access point within the SharePoint environment to view the documents they need as they are being created. For example, if someone loads a document in their own library anywhere in the world it is automatically classified and displays in the dashboard in real-time.
An additional and unanticipated outcome of the project allowed the team to expand the pilot and provide field offices the ability to access their own as well as neighbouring offices documents with minimal effort.
From a project success perspective, the team achieved the objectives of the pilot, which was globally rolled out behind the scenes in a production environment to build a strong business case to the broader organization. Internal stakeholders view the project favorably and have developed a 5-year plan that expands the use of Semaphore to incorporate additional document types, topics, and information (i.e., PII, GDPR - how it will impact the organization and how they will comply).
The combination of Semaphore capabilities and subject matter expertise results in a robust and precise user experience that reduces duplication, alleviates frustration, and provides users with a real-time view of the documents they need.