Overall update summary
2026.2 centres on making trusted master data usable by AI and on removing manual effort from the platform underneath it. The Consumer MCP Server becomes generally available as a governed interface for large language models and AI agents, while Dynamic Parallelism lets multithreaded background processes scale themselves against real-time capacity. Around that, the release extends Faceted Search into asset content, enforces terminology in AI translations through glossaries, tightens the asynchronous translation framework, and moves match ruleset tuning to data owners. In PDX, retailers gain centralized supplier monitoring and a localized Digital Catalog. The New UI becomes part of the base license for all SaaSv2 customers and presents complex relationships in a single structured view.
Consumer MCP Server
Platform | Included in Base | Admin, Developer, Business User | 26.2 highlight
What changed
Agentic AI is increasingly used to reason over enterprise knowledge, but access to trusted master data has usually depended on custom, tightly coupled integrations that limit reuse and scale, and without a standard interaction layer AI agents struggle to discover, understand and safely access master data. In 2026.2 the Consumer MCP Server is generally available. It provides a governed Model Context Protocol interface that lets large language models and AI agents discover and access enterprise master data, building on Data as a Service to expose that data with the semantic context AI-driven interaction needs.
- Decouples AI agents from underlying data implementations while preserving trust, governance and centralized control.
- Lets organizations scale agentic AI use cases and reuse consistent data access patterns.
- Makes enterprise master data more accessible and meaningful for AI-driven interaction.
Additional resources
Dynamic Parallelism for One Queue
Platform | Included in Base | Admin | 26.2 highlight
What changed
One Queue already balances background workloads across integrations, bulk operations, internal processing and scheduled jobs, but parallel processing capacity for multithreaded processes such as large imports has been statically configured — capping throughput when spare capacity exists and forcing manual adjustment for large or time-critical work. In 2026.2, Dynamic Parallelism is added to the One Queue framework so multithreaded background processes automatically scale their processing capacity up or down at runtime based on real-time capacity availability and overall platform demand.
- Improves processing speed for large workloads while maintaining fairness and stability across the platform.
- Maximizes use of available capacity without manual tuning, reducing operational effort.
- Supports more predictable execution of business-critical workloads.
Additional resources
Self-Service UI improvements
Platform | Included in Base | Admin | 26.2 highlight
What changed
The Self-Service UI has offered limited visibility into user activity and relied on manual provisioning of Control Service API credentials, leaving login events, platform-initiated changes, external log access and credential management only partly supported. In 2026.2 the User Activity Log includes login events and changes performed by Stibo Systems and is available through the Control Service API, and administrators can create, scope, renew and revoke Control Service API credentials directly in the Self-Service UI.
- Gives customers a single, consistent audit trail and direct control over API credentials.
- Improves transparency and strengthens security and compliance posture.
- Removes manual provisioning effort and supports integration with external auditing and monitoring tools.
Additional resources
Faceted Search for asset content
Platform | Included in Base — Early Adopter Release | Business User | 26.2 highlight
What changed
Users often recognise an asset by what is inside the document — contracts, certifications, specifications, manuals — and need search to match that content rather than only metadata and attributes. With 2026.2, Faceted Search can match text inside assets such as PDFs, Office documents, HTML, text and common CSV-like formats and return those assets in results, using the same full-text search patterns as today, now extended beyond metadata. This is available in both the New UI and Faceted Search in the Web UI.
- Speeds up “find the right document” workflows.
- Lets users search on document body content without leaving familiar Faceted Search patterns.
- Extends existing full-text search rather than introducing a separate experience.
Additional resources
SMTPS for email communication
Platform | Included in Base | Admin | 26.2 highlight
What changed
The SaaS platform has supported outbound email over SMTP with STARTTLS, which requires protocol negotiation before encryption is enforced and therefore allows a fallback to unencrypted connections. In 2026.2, support for SMTPS (implicit SSL) is added for outbound email. SMTPS enforces TLS encryption immediately when the connection is established and aligns with current industry standards such as RFC 8314, as part of a broader ongoing initiative to strengthen platform security.
- Email is encrypted from the moment a connection is established.
- Reduces the risk of misconfiguration and eliminates unencrypted transport.
- Aligns email communication with modern security and compliance expectations.
Additional resources
Multi Edit Data Container — Node Picker in the Create dialog in Web UI
Platform | Included in Base | Business User | 26.2 highlight
What changed
When the Multi Edit Data Container Create dialog moved to the same value controls as Unfolding Data Containers, reference fields lost the ability to use a Node Picker. With 2026.2, the Globally Configured Reference control shows a Node Picker in the Multi Edit Data Container Create dialog, enabling itself automatically when it detects it is in the Create dialog and the table's reference component is configured with a Node Picker — without adding a new configuration toggle.
- Restores feature parity with the earlier Multi Edit Data Container behavior.
- Lets users select referenced objects via Node Picker when creating new data containers.
- Keeps the shared reference control consistent across contexts.
Additional resources
Elasticsearch — improved reindexing speed
Platform | Included in Base | Admin | 26.2 highlight
What changed
Elasticsearch reindexing has been slow on both large and small systems, because the reindex process occupies shared workers for its entire duration — leaving little room for indexing processes — and certain operations bottleneck on larger datasets. With 2026.2, reindexing no longer occupies shared workers while it is not working, time-consuming work is moved to other processes, and the batch size of the indexing processes spawned by reindexing is increased.
- In many cases reindexing is 2–3 times faster.
- Leaves more capacity available for regular indexing during a reindex.
- Removes bottlenecks that affected larger datasets in particular.
Additional resources
Supplier Management in PDX
PXDC / Manage | Available with the PDX Onboarding license | Admin, Business User | 26.2 highlight
What changed
Supplier onboarding in PDX Onboarding could feel long and fragmented, with limited visibility into supplier channel adoption once work was under way, and retailers needed a faster way to review supplier connections, extract data for analysis and track progress without relying on multiple sources or unclear handoffs. With 2026.2, a centralized self-service space for supplier monitoring is introduced, with clearer progress visibility into onboarding processes.
- Faster onboarding cycles and a clear view of where adoption stands.
- Higher-quality product data, helping retailers get compliant listings sooner.
- Helps suppliers reach “live” faster.
Additional resources
AI-powered translations — glossary support
PXDC / Manage | Available with the Translation Service Integration and ProductGen AI licenses | Admin | 26.2 highlight
What changed
AI-powered translations infer meaning and terminology from context, which works well for general language but can translate customer-specific terms such as brand names, technical vocabulary and standard phrases inconsistently — leaving reviewers to correct recurring issues by hand. With 2026.2, glossary support is introduced for AI translations. Customers can define approved source-to-target term mappings that the AI must follow, configured for asynchronous AI translation services in the Workbench, where they act as hard rules enforcing consistent terminology.
- Improves consistency and reduces rework in AI translation workflows.
- Speeds up review cycles.
- Lets customers scale AI translations with greater confidence while preserving brand, technical and domain-specific language.
Additional resources
Enhancements to the asynchronous translations framework
PXDC / Manage | Available with the Translation Service Integration license | Admin | 26.2 highlight
What changed
The asynchronous translation framework has included all reference types and object names in translation exports by default, even when they are not language-dependent, making translation files larger than necessary and increasing processing time and cost without improving translation value. With 2026.2, the framework applies consistent inclusion, contextual and exclusion rules across product attributes, reference metadata and object names, all managed centrally through the asynchronous translation service configuration.
- Translation files are smaller, faster to process and better aligned with user intent.
- Reduces translation cost and turnaround time.
- Translators receive clearer, more relevant context without unnecessary noise.
Additional resources
Match tuning by data owners
CXDC, BPDC, SDC / Manage | Included in Base — Limited Availability | Admin | 26.2 highlight
What changed
2026.1 introduced a shared view of match accuracy and data-quality issues so super users and data owners could review potential matches together, but rule ownership remained mainly with super users, keeping tuning slow and dependent on specialists. With 2026.2, data owners can create, manage and activate match rulesets in collaboration with super users, directly in the Match Tuning app.
- Faster tuning with fewer errors and less manual review.
- Clearer ownership of match rules.
- Supports incremental rollout of tuning decisions.
Additional resources
Relevant import and export configurations in List Processing
CXDC, BPDC, SDC / Manage | Included in Base | Business User | 26.2 highlight
What changed
Users creating or exporting a list have been shown every import and export configuration, making the right one hard to find. With 2026.2, List Processing limits import and export configuration choices to those relevant for each list object type, based on global Web UI configuration.
- Users see only the options appropriate for their lists.
- Reduces clutter and mistakes.
- Makes the correct configuration faster to find.
Additional resources
Digital Catalog browsing experience
PXDC / Share | Available with the PDX Digital Catalog license | Business User | 26.2 highlight
What changed
The Digital Catalog UI supported only English, with limited search and navigation and a basic product page, which slowed product identification and held back adoption. In 2026.2 the catalog adds UI localization in German, French, Spanish, Italian and English, improved search and more intuitive navigation with faceted filtering, and a webshop-like product detail page with an image carousel, enhanced content and an export option.
- Delivers a modern, user-friendly catalog experience.
- Accelerates product discovery.
- Supports global rollout to retail partners.
Additional resources
New UI included by default for SaaSv2 customers
PXDC / Manage | Included in Base | Admin | 26.2 highlight
What changed
Access to the New UI required a manual enablement step — a $0 order form submitted before a customer could try or use it — which added friction and delay even though the New UI is the successor to the Web UI and the future frontend for the Stibo Systems platform. Starting with 2026.2, the New UI is included in the base license for all SaaSv2 customers. No order forms or license flags are needed, and customers control which users have access through standard privilege management.
- Removes customer friction and internal operational overhead.
- Enables faster migration to and exploration of the New UI.
- Confirms the New UI as the default forward path for the Stibo Systems platform.
Additional resources
Enhanced overview for complex relationship scenarios
PXDC / Manage | Included in Base | Business User | 26.2 highlight
What changed
The New UI works well for simple relationship scenarios, but as data models grow more complex, users need product data and related context understood together — and when that context is spread across views, maintaining a holistic overview becomes harder and more effortful. In 2026.2, relationship data is presented together with relevant product data and key information from related objects in a single structured view, enabling a 360-degree overview for complex scenarios while still supporting simpler use cases without added complexity.
- Better overview and understanding of complex relationship scenarios, with less navigation and clearer context.
- Users work more efficiently and confidently.
- Improves data quality and reduces risk when managing relationship-heavy data.
Additional resources
- Youtube chapter: Watch
FAQ
2026.2 makes the Consumer MCP Server generally available for governed AI access to enterprise master data and adds Dynamic Parallelism so multithreaded background processes scale at runtime. It extends Faceted Search into asset content, adds glossary support and inclusion rules to AI and asynchronous translations, opens match ruleset tuning to data owners, and filters import and export configurations in List Processing. In PDX it adds centralized supplier monitoring and a localized Digital Catalog, and it makes the New UI part of the base license for SaaSv2 customers with a new single-view overview for complex relationships.
Administrators are the primary audience for most platform updates — Dynamic Parallelism, Self-Service UI, SMTPS, Elasticsearch reindexing, translations and match tuning. Business users are targeted by Faceted Search for asset content, Multi Edit Data Containers, List Processing, the Digital Catalog and the New UI relationship overview. The Consumer MCP Server targets administrators, developers and business users.
Manage covers supplier management, translations, match tuning, List Processing and both New UI updates. Share covers the Digital Catalog browsing experience. The remaining updates — Consumer MCP Server, Dynamic Parallelism, Self-Service UI, Faceted Search, SMTPS, Multi Edit Data Containers and Elasticsearch reindexing — are platform-level.
Most updates are included in base, including the Consumer MCP Server, Dynamic Parallelism, Self-Service UI improvements, SMTPS, Multi Edit Data Containers, Elasticsearch reindexing, List Processing, match tuning and both New UI updates. Supplier Management requires the PDX Onboarding license. Glossary support requires the Translation Service Integration and ProductGen AI licenses, and the asynchronous translations enhancements require the Translation Service Integration license. The Digital Catalog browsing experience requires the PDX Digital Catalog license.
Yes. Faceted Search for asset content is an Early Adopter Release — contact your Stibo Systems representative. Match Tuning by Data Owners is included in base but in Limited Availability; email simplifymatching@stibosystems.com to take part.
Links to the 2026.2 update notes, overview video, YouTube chapters and overview deck are listed under Supporting assets. [Insert / confirm]
Action required: SaaS customers should use the SaaS Self-Service UI to create an Update Preview Sandbox and/or update one of their existing environments.
Implementation notes: Customers on Continuous Updates should have their production environment updated by September 18. The enhancements to the asynchronous translations framework are delivered behind a configuration property and can be adopted at your own pace; existing functionality remains available for now and will be phased out later, on a date not yet defined. Match Tuning by Data Owners is an early release — contact simplifymatching@stibosystems.com to take part. Faceted Search for asset content is an Early Adopter Release; contact your Stibo Systems representative.
Upgrade considerations: 2026.2 is only available on SaaS. On-premises customers should contact their Stibo Systems representative to start migrating their solution to SaaS.