Two Platforms, One Migration: Moving Semaphore Auto-Classification and SharePoint to the Cloud in Parallel.

About the Client

Headquartered in White Plains, New York, NYPA is the largest state public power organization in the United States, operating hydroelectric, transmission, and other generation assets that supply electricity across New York State. An organization of NYPA’s scale generates a large volume of engineering, regulatory, and operational documentation, spread across many organizational units — content that is only useful if people can actually find it.

Industry Context

Utilities and other large, regulated organizations increasingly rely on automated content classification to keep pace with the technical and regulatory documentation they generate, since manual tagging cannot scale across dozens of sites and libraries. As organizations move both their content platform and their classification infrastructure to the cloud, they face a specific risk: migrating the two independently, or migrating one without the other, can quietly undo years of investment in findability if the taxonomy, rules, and search infrastructure are not rebuilt together on the new platforms.

Current Situation

NYPA’s original Semaphore implementation, delivered by Iknow in 2017, had been running successfully in an on-premises SharePoint 2013 environment for years, with both the Semaphore classification engine and the SharePoint content platform hosted on-premises. As NYPA planned to retire that on-premises footprint in favor of Semaphore Cloud and SharePoint Online, it engaged Iknow again — this time under a multi-year Semaphore Cloud hosting services contract beginning January 2024 — to migrate both platforms together, refreshing the taxonomy along the way.

• Semaphore Cloud: a taxonomy that had fallen behind the business. NYPA’s 2017 taxonomy and classification rules did not reflect organizational units that had emerged since, and needed reconfiguring rather than simply copying onto the new cloud platform.
• Semaphore Cloud: fact extraction remained untapped. NYPA wanted to extract structured facts, such as document summaries and regulatory references, directly from its content — a capability not yet implemented on any platform.
• SharePoint Online: no direct path from the old platform. The search infrastructure, content types, and term-driven pages that made Semaphore’s classification usable had been built for on-premises SharePoint 2013 and needed rebuilding for SharePoint Online.
• SharePoint Online: manual, site-by-site deployment would not scale. Rolling out Semaphore configuration by hand across dozens of SharePoint Online sites would be slow and error-prone.

• Migrate Semaphore auto-categorization from on-premises hosting to Semaphore Cloud, updating the taxonomy and rules to reflect NYPA’s current organization.
• Implement Semaphore fact extraction on the new cloud platform to pull structured information directly from NYPA’s content.
• Rebuild the SharePoint Online infrastructure — content types, modern search, and term-driven pages — needed to put Semaphore’s classification to work for users.
• Automate deployment across SharePoint Online sites to support reliable, repeatable scaling.

How Iknow Structured the Work

Iknow structured the engagement as two coordinated workstreams — a Semaphore Cloud workstream and a SharePoint Online workstream — run in parallel rather than treating the SharePoint migration as a container for the Semaphore work. This reflects Iknow’s automated content classification methodology for de-risking a dual-platform cloud transition through staged proof-of-concept before scaling across the full environment.

Semaphore Cloud Workstream

• Taxonomy and rules refresh. Working with the NYPA client team, Iknow added terminology for new organizational units, refined classification rules, implemented separate development and production taxonomy versions on Semaphore Cloud, and validated results through extensive classification testing.
• Cloud-hosted auto-categorization. Iknow moved Semaphore’s auto-categorization from on-premises hosting to Semaphore Cloud, progressing from an initial demonstration to full production implementation.
• Fact extraction implementation. Iknow configured Semaphore to extract structured facts — including summary paragraphs, lessons learned, author names, reference numbers, and regulatory sections — a capability implemented for the first time on the new cloud platform.

SharePoint Online Workstream

• Content types for streamlined tagging. Iknow created and deployed SharePoint Online Content Types connected to the Semaphore taxonomy in the term store, using them to auto-configure classification for each document library.
• Modern search infrastructure. Iknow configured SharePoint Modern Search Web Parts — managed and crawled properties, result sources, and query strings — so a stored search page could be created quickly for any term in the taxonomy.
• Site finalization and deployment automation. Iknow configured each site page’s search filters, sort options, and result views, then scripted the site-level deployment process in PowerShell, including app installation, content type deployment, and triggering auto-classification across every library.

Stakeholders & Collaboration

Iknow served as prime contractor, working directly with the NYPA client team throughout the taxonomy refresh, testing, and phased rollout across both platforms.

Semaphore Cloud: Refreshing a Seven-Year-Old Taxonomy Without Losing Continuity

NYPA’s business had evolved considerably since 2017, but classification still needed to work reliably for existing content as new terminology was introduced on the new cloud platform. Iknow addressed this by maintaining separate development and production taxonomy versions and running extensive classification testing before cutover, ensuring the refreshed taxonomy performed as well as the original before going live.

SharePoint Online: Scaling Deployment Across Dozens of Sites Without Manual Reconfiguration

Deploying Semaphore site by site — app installation, content types, auto-classification triggers — across dozens of SharePoint Online sites by hand would have been slow and inconsistent. Iknow addressed this by scripting the entire deployment process in PowerShell, turning a manual, error-prone task into a repeatable one that could scale across NYPA’s full environment.

Quantitative Outcomes

• Semaphore Cloud — Taxonomy refreshed: Semaphore taxonomy and classification rules updated for the current business and validated through extensive classification testing.
• Semaphore Cloud — Classification migrated: Semaphore auto-categorization moved from on-premises hosting to full production implementation on Semaphore Cloud.
• Semaphore Cloud — Fact extraction implemented: Multiple fact types — including document summaries, lessons learned, author names, and regulatory references — implemented for the first time on the new cloud platform.
• SharePoint Online — Infrastructure rebuilt: Content types, modern search, and term-driven pages implemented and synchronized with the SharePoint Online term store.
• SharePoint Online — Deployment automated: PowerShell-based deployment scripts successfully used to roll out Semaphore auto-classification across dozens of SharePoint Online sites.
• Engagement scope: Six major task areas completed across both workstreams within a multi-year Semaphore Cloud hosting contract beginning January 2024.

Qualitative Outcomes

Semaphore classification continues to yield excellent results for NYPA on the new cloud platform, accurately classifying content without manual tagging or editing. On the SharePoint Online side, the new search and auto-generated topic pages are popular with users and largely self-maintaining: users add content using standard upload and data-entry features, and it appears immediately in search results and on topic pages. Because the two workstreams were planned and executed together, NYPA’s original 2017 investment in automated classification carried forward successfully into both new platforms rather than being lost in the transition, and the underlying hosting contract continues, with further work under way to extend fact extraction and site functionality.

Timeline to Impact

Within the first year of a multi-year hosting contract, Iknow moved NYPA from a legacy, on-premises classification system spanning Semaphore Server and SharePoint 2013 to a coordinated, cloud-based implementation spanning Semaphore Cloud and SharePoint Online, complete with automated deployment scripting and expanded fact extraction. This preserved and extended seven years of investment in automated findability.

Core Skills

• Automated content classification and taxonomy design
• SharePoint Online architecture and implementation
• Enterprise search configuration
• Deployment automation and scripting

Methods & Frameworks

• Staged proof-of-concept to full production implementation
• Development/production taxonomy version control and validation testing
• Scripted, repeatable multi-site deployment

Technologies & Tools

• Progress Semaphore (Semaphore Cloud)
• Microsoft SharePoint Online and SharePoint Modern Search Web Parts
• PowerShell scripting and SharePoint automation