The independent daily for Salesforce admins, developers and architects

Curated by Abhinav (opens in a new tab)

Measure AI's effect on Salesforce delivery with DORA metrics, not output counts (opens in a new tab)

You will learn why output measures such as tickets closed and pull requests merged can mislead when judging AI-assisted development, and which DORA metrics, including the newer rework rate, show whether AI is saving time.

Why read this

The article also covers the human signals to watch alongside them.

  1. Deployment frequency
  2. Lead time to change
  3. Change failure rate
  4. Time to recover
  5. Rework rate
The four DORA metrics plus rework rate, used together to test whether faster delivery costs stability

The article argues that organisations judge AI-assisted development by activity such as lines of code, tickets closed and pull requests merged, which are easy to game and say little about whether the work is valuable. It cites Google’s DORA research finding that rising AI use correlates with greater org instability, and that the rework this creates can consume the time AI was meant to save.

The author recommends returning to the four DORA metrics: deployment frequency, lead time to change, change failure rate and time to recover, together with the rework rate metric DORA added in 2024. Taken together, these show whether faster delivery is coming at the cost of stability. The author also stresses that the numbers show what is happening but not why, so teams should track human signals such as trust in AI output, verification load and user adoption.

The piece closes with a promotional section for Gearset’s AI agent Cam and its DevOps performance dashboard, which tracks DORA metrics from a pipeline. The promotional content is brief and the general guidance stands apart from it.

Read the original at Salesforce Ben (opens in a new tab)

Worth reading?

Using Apex frameworks to enforce code quality and platform limits in the AI era (opens in a new tab)

Readers will learn how to structure Apex libraries with strict interfaces and runtime validations to prevent both human and AI errors.

Why read this

The piece also explains how these patterns automatically optimize SOQL and DML usage to stay within Salesforce governor limits.

The article, written by Piotr Gajek, a Senior Salesforce Developer and Technical Architect, introduces Framework-Driven Development (FDD), an approach developed over a couple of years in his Apex Fluently libraries. Its core idea is to move responsibility for code quality and Salesforce governor limits out of individual developers and AI agents and into shared frameworks and the development process. He argues that LLM output is statistically average and often disconnected from the wider architecture, and that shared limits on an org mean one team’s careless generated code can break another’s solution. He describes the phenomena of “meat proxies” and “slop grenades” as reasons why ownership of code becomes unclear.

The approach rests on three rules: frameworks should be idiot-proof and AI-proof, generated code should be optimized for platform limits, and quality should be built in by design. Idiot-proofing uses limited interfaces that expose only permitted methods, such as the SOQL Lib cache module that blocks toList() because of misleading cached results, and the Trigger Lib contexts that allow only relevant actions per trigger phase. Internal validation throws explanatory exceptions with suggested overrides, and documentation and AI-readable skills act as a final safeguard.

For limits, the author highlights Async Lib, which manages queueable, schedulable and batch jobs to avoid the “Too many queueable jobs” error; DML Lib, which builds a dependency graph and commits records in optimized order through a shared Unit of Work; SOQL Lib, which caches queries in static, org or session scope; and Trigger Lib, which batches parent queries from many handlers. Quality by design is reinforced by phase-specific rules and small, single-responsibility classes. He also suggests administrators could benefit when Apex-based logic replaces Record-Triggered Flows. The conclusion recommends combining libraries, context documents and layered testing, with TDD where useful.

Read the original at Beyond the Cloud (opens in a new tab)

Worth reading?