What changed
July 19, 2026
Manage access by department and group
Administrators can give people and Access Groups the right role at the organization, department, or Functional Group level, then use Access Inspector to understand why someone has access.
Dashboards and reports feel faster
Common dashboard and reporting views now reuse data more efficiently, preload pages as you navigate, and avoid unnecessary refreshes while keeping workspace information current.
A smoother path to your first agent
Website consultation, account setup, email verification, agent creation, validation, and deployment now provide clearer progress and a more continuous guided experience.
More reliable live updates and agent editing
Live status stays connected more consistently as you move between pages, and agent source opens as readable code without stale or incorrectly encoded editor content.
Bug fixes
We fixed issues that affected account setup, plan price displays, audit time filters, page responsiveness, and several slow or stale loading states.
July 15, 2026
A simpler way to work with agents
New accounts start in Business mode with clearer agent names, statuses, and pages. Studio mode remains available when you need source and runtime detail.
More tools in the connector catalog
Browse hosted connectors by category, choose between a provider’s available servers, and connect another instance without leaving the catalog.
Feedback now turns into agent recommendations
Rate completed runs and chat replies, explain what went wrong, and review suggested improvements before sending a change to Agent Builder.
Help is available inside the app
Open Help from anywhere in Iron Gorilla to find relevant articles or send a support request with useful page and diagnostic context.
Plan limits and upgrade paths are clearer
Agent, audit, routing, environment, and identity limits now explain the available plan and whether you can upgrade or need an administrator.
Authorized administrators can pause AI work
Preview the impact of a pause, stop AI work for an organization, department, group, or agent, and resume it when the issue is resolved.
Voice conversations with agents
Agents configured for voice can hold governed web conversations with consent, mute, restart, transcript, approval, and fallback controls.
Agents can use scoped memory and autonomous work
Where enabled, agents can recall approved context within their assigned scope and work through bounded multi-step tasks with governed tools.
Review risky changes with replays and narratives
Where enabled, reviewers can compare a proposed recovery in simulation and approve a cited governance summary before it is exported or shared.
Grok 4.5 model routing
Organizations with xAI enabled can select Grok 4.5 as a preferred or fallback model and review its context, pricing, and trace details.