What I Learned Training 40 Salesforce Architects on Agentforce and AI

Over the past year, I designed and delivered Agentforce and AI training for a cohort of 40+ senior Salesforce architects. These aren’t junior admins learning the platform for the first time — these are experienced architects who have spent years implementing complex Salesforce solutions. They know the platform. They know enterprise implementation patterns. They’ve seen a lot of technology promises come and go.

Getting this group to genuinely understand Agentforce — not just the mechanics, but when and how to apply it well — was one of the more interesting challenges I’ve worked on.

Here’s what I found.

---

The Technical Part Is Easier Than You’d Expect

Experienced Salesforce architects pick up the Agentforce technical model faster than most people assume. Topics and Actions map reasonably well to existing mental models around record types and Flows. The Einstein Trust Layer configuration has parallels to Connected App and OAuth work they’ve done before. Agent Builder’s interface, while new, is learnable.

The technical concepts are not the hard part.

---

The Mindset Shift Is Harder

[Full article coming soon — this post is in development.]

Topics to be covered:

• Why architects trained on deterministic systems struggle with probabilistic AI outputs
• The challenge of designing for “good enough” instead of “correct”
• How to teach prompt engineering to people who think in code
• The most common misconceptions experienced Salesforce architects bring to Agentforce
• What the best-performing architects in the cohort did differently

---

If your team is preparing for an Agentforce implementation and needs architectural training or advisory, let’s talk.