SDLC Transformation with AI

Localization Bootstrap Skill

Agentic skill that ensures a UI project is properly instrumented and setup for localization, and all texts are passed through translation.

Demo: Localization Bootstrap Skill.mp4

Automated Code Review & Fix Implementation with AI Agents

Watch two AI skills work together in the agent chat to handle the full code review lifecycle. /review-pr analyzes a pull request by dispatching specialized analyst agents in parallel — covering code quality, security, accessibility, performance, AEM patterns, and test coverage — then posts a unified review with inline comments on GitHub. /implement-review-findings picks up those findings, applies fixes with test verification, commits to the PR branch, and resolves every review thread. From review to resolution, all from a single chat window.

Git: https://github.com/Adobe-AEM-Sites/aem-cursor-toolkit

Demo: Automated Code Reviews

Remote agentic code review using GH Actions


Presenting two approaches when it comes to AI remote code review using GH Actions and Claude.
One is using the code review skill from the team’s unified AI Native Workspace.
The second is leveraging the AI review kit from the Experience Success skills.
Just create a PR on the repo and the ask the agent ( /review comment) and let it do the magic.

Git: Review Skill,
ESO Repo with the GH Action agent enabled

Demo: Link

AI-driven migration from Commerce Classic to Headless PIM

Migration of Commerce Classic implementations to a Product Information Management using headless AEM, compatible with CIF, developed spec-based with AI. The compatibility tests and clear contracts of CIF offered the right context for AI ( in demo after second 13:20 )

Demo: Celadon - Headless PIM.mp4

LLM Optimizer

UI Native Offsite Opportunities

NEW UI Native for Offsite opportunities Reddit, Youtube, TopCited analysis

Customers: a.com

Demo: UI Native Offsite Opportunities video with sound

AEM Sites Omnichannel

Publishing Visual Content Fragments

Visual Content Fragments can be included in EDS, Cloud Service or any webpage

Demo: Publishing Visual Content Fragments.mov

Figma CF Visualizer Improvements

We’ve significantly improved the reliability and performance of the Figma CF Visualizer agent by making the workflow fully deterministic. Instead of relying on the LLM to decide each step, orchestration is now handled directly in code—eliminating skipped steps and inconsistencies.

Additional enhancements include improved Figma-to-HTML extraction.

Result: up to 10x faster execution and much more predictable outputs.

Demo: Figma CF Visualizer.mov

Content Management Agent


Manages Content Fragments at scale through natural language. Describe what you want to change - update fields, find and replace text, delete or copy - and the agent figures out the scope, shows you what it found and works through it in batches with your approval at every step. A launch is created automatically before destructive changes so nothing is permanent until you're ready. Every operation reports back visually: what changed, what failed, what was skipped.

Demo: CMA launches renderer

AEM Sites Optimizer

Code auto-optimize validation pipeline automation

We're increased our coverage by 28% with our latest finetuning process. We've run the pipeline for Okta and Redtag CS customers and we've managed to have 57% and 83% issue resolution respectively, for the specific a11y issues we're targeting.

We have added the following automations to the A11y autofix pipeline (local) through scripts/skills: Agentic review, Mergic, RV/CM deployment - fully automated

Customers: Okta, Redtag

Demo: A11y autofix demo.mp4

Web Performance Trends is now GA!

  • Web Performance Trends are a set of reports for both Mobile and Desktop that offer visibility of how the URLs are behaving in terms of Good, Needs Improvement or Poor web performance and how they trend for the last 28 days.
  • Also, for each URL we are showing the Web Performance metrics and business metrics in comparison with last week. They can be filtered by path or by type of issue and metric affected.
  • These can now be enabled through the Spacecat Bot Commands.

Demo: demo-wpt-ga.mp4