>Long-form tutorials on Salesforce + AI. Paste-ready configs, real prompts, the gotchas I hit on camera.
CumulusCI creator Jason Lantz names the gap on every Salesforce project: the space between deploying something and actually delivering it. On Revenue Engineer Episode 3, the deploy-vs-deliver test, why it costs 160 hours an engagement, and how to hand it to an AI without shipping a security hole.
Salesforce now gives you two AIs to change your org. On Revenue Engineer Episode 5, Vuk Stajić and Warren Walters put Setup with Agentforce and the CLI side by side, live, on the same prompt: the naming trap, the draft-and-confirm step the native tool gets right, where it still misses your org's context, and the four guardrails that apply either way.
Salesforce MVP Chris Pifer runs AI tooling on enterprise builds with ten-person teams. On Revenue Engineer Episode 4, the seven-phase workflow that captures the speed without the blast radius: risk tolerance first, plan authoring, adversarial review, the lazy agent that fakes a passing test, and Git as the life raft.
Amber Boaz built a 5-part prompting framework with Claude: Goal, Role, Action, Context, Example. On Revenue Engineer Episode 2 she runs it live, from a LinkedIn post to a full Salesforce UAT test plan. Here's the framework and the moves that make it work.
Jack Gimbert demos a live Salesforce build using Claude Code: invokable Apex, a Flow trigger, and an LWC, from user story to deployed feature in 33 minutes. Here's what happened and what it means for practitioners.
One External Client App, four AI clients. Paste-ready configs for Salesforce's hosted MCP servers in Claude Code, Cursor, Codex CLI, and ChatGPT.
A step-by-step walkthrough of the official Salesforce hosted MCP setup, covering the External Client App, OAuth scopes, security flags, and connecting the sobject-all server to Claude on the web. Includes the exact prompts to test read, write, update, and delete from chat.