SOURCES
Current product pages and dated announcements support this research snapshot. Inclusion does not imply affiliation.
PRIMARY SOURCE LEDGER
Undated live product page, accessed 2026-11-08 · Checked 8 November 2026
Current name is Microsoft Foundry. The product spans model choice, agent services, organizational grounding, tools, controls and observability. It presents one enterprise surface across the agent lifecycle.
https://azure.microsoft.com/en-us/products/ai-foundry/High for current positioning; exact availability requires feature documentation
2026-06-02 · Checked 8 November 2026
Jay Parikh's official strategy statement connects agent building, enterprise context, runtime operations, governance and improvement. It emphasizes developers, open model choice and human oversight rather than isolated demonstrations.
https://blogs.microsoft.com/blog/2026/06/02/ai-alone-wont-change-your-business-the-system-running-it-will/High
2026-06-02 · Checked 8 November 2026
Describes workload-specific model selection, representative evaluation and continuing operations. Capability, safety, latency and cost are considered together. The article provides a model optimization loop graphic.
https://devblogs.microsoft.com/foundry/build-2026-foundry-models/High
2026-11-05 · Checked 8 November 2026
Separates the Copilot harness public preview from generally available browser automation and hosted-agent resilience. Recovery still requires application checkpoints and safe repeat handling. November Ignite sessions are upcoming, not completed announcements.
https://devblogs.microsoft.com/foundry/whats-new-in-microsoft-foundry-october-2026/High; release-specific
2026-10-22 · Checked 8 November 2026
Foundry costs comprise dependent services and billing meters. Estimate scope must match actual resource scope. Some project attribution remains preview, and estimates are not invoices. Check regional availability and contracted rates.
https://learn.microsoft.com/en-us/azure/foundry/concepts/manage-costsHigh; implementation-specific and changeable
2026-04-22 · Checked 8 November 2026
Launch describes an evolution of Vertex AI, combining model services and agent development with runtime, governance and operational improvement. It names product executives and states future Vertex AI roadmap delivery through Agent Platform.
https://cloud.google.com/blog/products/ai-machine-learning/introducing-gemini-enterprise-agent-platformhigh
Undated living page; accessed 8 November 2026 · Checked 8 November 2026
Documentation maps Vertex AI Platform to Gemini Enterprise Agent Platform and lists renamed components, including Studio, model services, training and runtime. It is a naming reference, not a claim that every feature has identical status.
https://docs.cloud.google.com/gemini-enterprise-agent-platform/vertex-ai-name-changeshigh
Undated living page; accessed 8 November 2026 · Checked 8 November 2026
The visible heading uses Gemini platform. It presents technical-team agent/application building, model choice, tuning, enterprise data, evaluation and MLOps, with documentation, practical blueprints and developer resources.
https://cloud.google.com/products/gemini-enterprise-agent-platformhigh for retrieved text; naming varies across live pages
2026-10-08 · Checked 8 November 2026
Thomas Kurian is identified as Google Cloud CEO. The announcement connects business context, work execution, enterprise controls and cost management. It labels financial-services and legal specialization as preview and additional industries as coming soon.
https://cloud.google.com/blog/products/ai-machine-learning/welcome-to-gemini-at-work-2026high
2026-10-22 · Checked 8 November 2026
Explains the connection between agent sessions, isolation, identity, retained workspaces and governance. The new Agent Compute Engine service is described as preview in coming months, so its announcement must not be presented as current general availability.
https://cloud.google.com/blog/topics/ai-infrastructure/introducing-agent-compute-engine-and-expanded-agent-governance/high
Living notes; latest reviewed entry 2026-11-06 · Checked 8 November 2026
The November 6 entry documents dedicated Memory Bank console management. Other dated entries distinguish availability states and operational changes, illustrating why retention, evaluation, logging and service status must be checked individually.
https://docs.cloud.google.com/gemini-enterprise-agent-platform/release-noteshigh; living documentation should be rechecked before outreach
Undated living page; accessed 8 November 2026 · Checked 8 November 2026
Presents model choice, application/agent development, customization, guardrails and cost optimization as connected capabilities for production AI.
https://aws.amazon.com/bedrock/High for vendor-described positioning; performance and adoption claims are vendor-reported and not independently audited here.
Undated living page; accessed 8 November 2026 · Checked 8 November 2026
Describes an integrated data, analytics and AI experience including Unified Studio, Catalog, Lakehouse and SageMaker AI. Governance and development tools span the lifecycle.
https://aws.amazon.com/sagemaker/High for current vendor-described structure; status varies by capability.
Undated living page; accessed 8 November 2026 · Checked 8 November 2026
Describes modular capabilities for runtime, identity, tool connectivity, observability, evaluations and policy. Explicitly lists building internal/customer agent platforms as a use case.
https://docs.aws.amazon.com/bedrock-agentcore/latest/devguide/what-is-bedrock-agentcore.htmlHigh for documented capabilities; do not infer universal region availability or all-feature GA from overview.
2025-10-13 · Checked 8 November 2026
Establishes the dated transition of the original AgentCore services to GA, including enterprise connectivity and observability. It does not establish the status of every later-added capability.
https://aws.amazon.com/about-aws/whats-new/2025/10/amazon-bedrock-agentcore-available/High; historical announcement.
Framework released 26 January 2023; living overview accessed 8 November 2026 · Checked 8 November 2026
The voluntary AI RMF organises AI risk management around Govern, Map, Measure and Manage. Our evidence inventory is an editorial aid, not an implementation audit or certification.
https://www.nist.gov/itl/ai-risk-management-frameworkHigh for framework purpose; no claim of conformity
Undated living page; accessed 8 November 2026 · Checked 8 November 2026
AWS distinguishes provider infrastructure responsibilities from customer responsibilities, which vary with the chosen services and integrations. Managed services do not remove the need to assign customer-side controls.
https://aws.amazon.com/compliance/shared-responsibility-model/High for the published model; application-specific obligations need specialist review.
