Andi March 2026 Release: Test, Learn, and Deliver the Full Lifecycle
Every Salesforce team knows what happens after the build. You’ve created the flow, deployed the logic, and now you need to validate it — which means manually setting up test records, checking field values one by one, and hoping nothing fails downstream. Or a new team member builds an Apex class that works but doesn’t follow any of the naming conventions your team spent years establishing. Or a field needs a small update, and the path of least resistance is to delete it and start from scratch.
These are the gaps the March release was built to close.
This month’s update introduces the Andi QA Agent at general availability, Supplementary Context in beta, meaningful expansions to org context intelligence, new modification capabilities for fields and validation rules, and a redesigned analytics dashboard. Together, these changes extend Andi across the full Salesforce delivery lifecycle — from planning and building to testing, refining, and measuring impact.
Here’s what’s new.
Andi QA Agent: End-to-end testing, now generally available
After a successful beta, the Andi QA Agent is now available to all teams — and it fundamentally changes what testing looks like inside a Salesforce org.
Andi can now generate comprehensive test cases directly from a user story, then execute those tests inside your org. No manual test record setup. No prompting Andi to clarify scope. Give it the acceptance criteria you need to test for and it takes it from there — inferring the necessary test scenarios, covering conditions and edge cases, and returning a clear picture of what passed and what didn’t.
Andi offers two testing modes: “hybrid” and “full ui”. Hybrid mode is faster and more consistently reliable, leveraging APIs to trigger and complete the work required. Full UI mode is perfect for those looking for additional proof; Andi logs into Salesforce and mimics the clicks within an agentic browser window, capturing the full test in a video backup file. In both scenarios, every test action is logged via JSON files for full traceability, giving teams a documented record of what was tested and what the results showed.
A few prompts to get started:
- “Generate a comprehensive set of test cases for the following user story: ‘When an opportunity is updated to Closed Won, a new order record should be automatically created and linked to the account.'”
- “Generate and execute all necessary test cases for the ‘Lead Assignment Rule Flow’ to verify that leads from California are correctly assigned to the West Coast sales team.”
Test execution currently supports record-triggered flows, with additional coverage coming in the April release. This capability is included automatically as part of every existing Andi subscription at no additional cost.
Supplementary Context (Beta): Teach Andi how your team builds
February’s Org Context Engine answered the question: what’s in your org? Supplementary Context answers a different question: how does your team uniquely want things built?
Teams can now upload their own internal documents — naming conventions, coding standards, process flows, testing scripts, design guidelines — directly into Andi’s knowledge store via the Supplementary Context UI. Andi automatically references these documents at runtime, so every component it designs and deploys reflects your team’s specific standards and preferences, without needing to be reminded each time.
Documents are organized across three categories — System Design, Standards, and Testing — with 16 category types in total. The knowledge store supports PDF and DOCX files, and access is permission-controlled so Super Admins can assign the right level of access per user.
A few prompts to get started:
- “Using our ‘Apex Best Practices’ document that I uploaded, refactor the AccountTriggerHandler class to include proper exception handling and to match our company’s variable naming standards.”
- “Our ‘Solution Architecture Guide’ PDF is now in the knowledge base. Based on its principles for building scalable automation, design a solution to handle multi-currency contracts on our custom ‘Agreement__c’ object.”
Expanded Org Context: Smarter intelligence across every feature
The Org Context Engine, introduced in February, gave teams a powerful new way to query and understand their org’s metadata. This month, that intelligence has been expanded throughout Andi’s sub-tools — so the awareness that was previously available in dedicated org context queries is now woven into more of Andi’s workflows by default.
This means that when you’re building a validation rule, creating a permission set, or deploying Apex logic, Andi is drawing on the same deep org context — not just when you explicitly ask for it. Metadata detection is more accurate, multi-step planning is more reliable, and Andi’s responses throughout complex workflows are clearer and more consistent.
A few prompts that take advantage of this expanded intelligence:
“Plan and execute the change of the Employee_ID__c field on the Contact object from a Text field to a Number field. As part of the plan, first identify all dependencies in Apex, Flows, and validation rules before proceeding.”
“I want to create a validation rule on the Opportunity object to prevent the Stage from being reverted. Before you create it, analyze all existing validation rules and flows on the Opportunity to ensure this new rule won’t conflict with any current automation.”
Field & Validation Rule Modifications: Update what already exists
A small but high-impact addition to Andi’s capabilities: teams can now ask Andi to update and modify existing custom fields and validation rules, not just create new ones.
For fields, this means changing labels, descriptions, help text, lengths, decimal places, and default values. Teams can toggle required, unique, and external ID attributes, manage picklist values, update formula expressions, and perform safe field type conversions — such as converting a Text field to a Long Text Area — without any risk of data loss.
For validation rules, Andi now distinguishes between create and update intent. When an update is requested, Andi retrieves the current rule definition, applies the requested changes, and redeploys via the Salesforce Metadata API — with improved error handling throughout. As with any new rule, teams can pair this with the Org Context Engine to review what’s already in place before making changes.
What this release means for teams
February gave Andi the ability to understand your org before it builds. March gives it the ability to act on that understanding more completely.
The QA Agent means testing is no longer something that happens outside of Andi — it’s part of the delivery workflow. Supplementary Context means Andi doesn’t just know what’s in your org, it knows how your team works. The modification capabilities mean Andi can maintain and evolve existing components, not just build from scratch. And the expanded org context means every one of these capabilities is informed by the same deep metadata awareness.
For teams using Andi in production, this release makes the full delivery lifecycle — from investigation to deployment to validation — faster, more reliable, and better aligned with how your team actually operates.
