When maintaining multiple similar connections for different environments, such as DEV, STAGE and PROD, Release 26.3 can save you time and clicks:

• Duplicate a connection - copy all settings except credentials. Re-enter passwords and API keys and update environment-specific values. Password variables are kept. The copy stays in the same folder, or in global.

• Delete a connection - completed and aborted releases no longer block deletion. Only active releases, triggers or other active configuration do. Clickable references show where the connection is used.

• Duplicate title warning - duplicate titles make connections hard to tell apart when users select one in a task. Release now warns when another connection of the same type has the same title. You can still save, for backward compatibility, but whoever configures the connection is made aware.

• View connections permission - a new folder-level permission. Users can see connections and their details, but cannot create, edit, delete or test them.

• Connection picker in the Task Drawer - connections are grouped by folder and show their URL, so picking the right one is easier.

Try these out and tell us what you think. What else would you improve in connection management?

👉 Read about the changes in the 26.3 release notes

With Digital.ai Deploy 26.3, we're introducing Deploy on Event (Technical Preview) — a new way to connect Git changes with deployment execution.
🔭 Imagine updating your deployment definition in Git and letting Deploy take it from there.
✅ No manual deployment triggers.
✅ No inbound webhooks.
✅ Less dependency on external CI orchestration.

How does it work?

Configure a Git Poll, select the deployment file you want to monitor, and set a polling schedule. When Deploy detects a change to that file, it prepares and triggers the deployment.

It's a small change in how you initiate deployments, but an important step toward more Git-driven deployment workflows.

🚀 Give it a try!

Whether you're already using GitOps practices or just exploring ways to simplify deployment operations, this is an opportunity to try the capability and help shape what comes next.

👉 Explore Deploy on Event in our documentation

How it works

The Agility data mart is a denormalised, reporting-optimised representation of your Agility operational database. Digital.ai enables and configures the data mart for your environment — you then connect Power BI to it and open the templates directly. Because Power BI runs in your own environment and reads your own data mart, your delivery data never leaves your security boundary. This means the templates work seamlessly across on-premises, private cloud, government cloud, and air-gapped deployments.

Every template is meant as a starting point, not a finished product. Analysts are free to extend visuals, add custom measures, apply their organization's branding, or use any template as a foundation for building their own reports.

Available Now

Coming Soon

Get started

Reach out to your Digital.ai representative to have the Agility data mart enabled for your environment, then connect Power BI and start exploring.

Refer: https://integrations.digital.ai/integration/ac-power-bi-templates-for-agility-1-0/

Help documentation: https://docs.digital.ai/agility/docs/agility/power-bi/power-bi-reports-overview

We'd love to hear what you build — share your customisations and feedback in the comments below!

Deploy on Event (Technical Preview) - Deploy can now watch a Git repository and automatically trigger a deployment whenever a tracked deployment file changes, no manual trigger or webhook needed.

Kubernetes installation updates - Setting up Deploy on Kubernetes gets more flexible in 26.3: you bring your own ingress controller instead of relying on one the installer manages, and PostgreSQL now installs and upgrades more smoothly by default.

Updated plugins: AWS, Azure, CyberArk Conjur, Helm, Jenkins, and Kubernetes.

⚠️ Breaking changes & End-of-life announcements

• Java Security Manager sandbox removed on both JDK 21 and JDK 25 - action required if Deployfile scripts relied on it.

• Apache Derby no longer supported - H2 is now the embedded DB for non-production; migrate production databases per the migration guide.

👉 Digital.ai Deploy Release notes | Download here (requires login)

The Digital.ai Deploy Team

AI Assistant - Context-aware, in-depth knowledge about the product and DevOps best-practices assistant that answers natural-language questions about your Release data - status, risks, blockers, overdue tasks - read-only and scoped to your permissions. Now available on-prem. Supports 3 agents:

• Generic Agent – General questions about releases, tasks, templates, folders, variables.

• Reporting Agent (Tech Preview) – Answers with charts/tables, savable to a dashboard.

• Task Troubleshooting – Explains why a task failed, from the Task drawer.

Agent Orchestration - Now an official plugin (previously community-maintained). Anthropic joins OpenAI, Gemini, and Digital.ai LLM as a supported provider for AI: Prompt/Chat/Agent tasks. MCP Server and Digital.ai LLM connections now support OAuth2. MCP: Call Tool renders a schema-based form instead of raw JSON.

Release MCP Server - Now ships with Release by default and no separate install needed. Used by AI Assistant and MCP clients like Claude Code, Cursor, and Copilot to work with releases, templates, tasks, and variables, honoring your existing permissions.

Other improvements:

• Python 3 for automation - This release extends Python support for Release automation with a standalone API client, a new SDK task, and direct API access from the Python 3 Script (Container) task.

• Timezone support - preferred timezone visible across Release, especially handy for distributed teams

• Live Deployments - Improved cards and external links to get a better overview of what's being deployed where with more details

• Connection management - Quality-of-life improvements to make it easier to find and manage 3rd-party connections

• Improved Kubernetes installation - For customer-managed installations of the Release product on Kubernetes. All the major vendors supported. Now it's easier to spin up a GitOps installation with support for Kustomize and the new xl kube apply command.

• Webhook (Container) - The new Webhook (container) task allows you to make REST calls to services outside the network of the Release server in a secure way

• and more

Updated plugins: Argo CD, Argo CD Container, Conjur, Deploy, GitLab, HashiCorp Vault, Jira, Remote Completion, ServiceNow, Tekton Container - and a new plugin for Terraform Enterprise/Cloud.

👉 See the Release notes for full details, including breaking changes and end-of-life announcements.

👉 Download here (requires login)

The Digital.ai Release Team

26.8 introduces two major features, Advanced Analytics and AI-Powered Root Cause Analysis, along with Pixel 11 Series support and same-day support for iOS 27.

👉 Read the full 26.8 Release Notes

A major OS update can change how your app behaves in ways you may not expect. Validate functionality, accessibility, performance, and hardened applications on real iOS 27 devices.

Watch the Overview

👉 Existing customer and want to upgrade your devices for iOS 27? Reach out to Support.

Requests are now available as a board, and that changes what a review feels like.

Read your intake in one look

Each request is a card in a column based on its status. Open the board and the pipeline is right there where requests are entering, where they are sitting, which stage is carrying more than its share. The questions that used to take a round of filtering get answered by looking.

The review becomes the work

When you decide something moves, you move it: drag a card to another column to update its status, or up within a column to change its ranking. Approved a request? Generate a story, portfolio item, defect, or issue right from the card. Nothing gets left in a notebook to action later.

Stakeholders get a picture, not a status report

"Your request is in review" lands very differently when someone can see it sitting next to everything else you are weighing. Install the board as a panel in a PlanningRoom and run the conversation on the live board — no exported snapshot that was stale before you sent it.

And it bends to how your team works

Set the fields on cards, group cards into rows by Priority, Source, Type and more, and write your board policies so everyone shares one definition of "In Review." Filters work exactly as they do on the list.

Try it on your next review

Open the hamburger menu > Product > Request Board, spend two minutes in the customise options, and run your next review on it. The list view has not changed most teams will want both: the board for the conversation, the list for the deep data work.

Over to you: What statuses does a request move through on your team, and where does it usually get stuck? Post your columns in the comments.