The more I work with AI, the more I find myself talking about everything except AI.

Here’s the surprising thing about preparing your Salesforce org for AI: most of the work has nothing to do with AI.

Over the past year, I’ve had countless conversations about AI with clients, at community events, during conference sessions, and across LinkedIn. No matter where the conversation starts, it almost always finds its way to the same question:

How do we prepare our Salesforce org for AI?

It’s a great question. But the more I thought about it, the more I realized I was giving the same answer every time.

Not, “You need the latest AI feature.”

Not, “You need to implement Agentforce.”

Instead, I found myself talking about data quality. Security. Governance. Naming conventions. Documentation. Metadata. All of the architectural fundamentals we’ve been talking about for years.

That’s when something clicked.

AI isn’t replacing good architecture. It’s exposing bad architecture.

AI didn’t suddenly make good architecture important. It simply made it impossible to ignore.

For years, architects have encouraged organizations to invest in clean data, consistent naming conventions, thoughtful security models, governance, documentation, and scalable configuration. Those conversations weren’t always exciting. They were often the first things cut from a project timeline or pushed into a future phase because they didn’t provide immediate business value.

Today, they’re no longer nice-to-haves. They’re competitive advantages.

An AI agent can only be as effective as the information it’s given. It can’t confidently answer questions using outdated knowledge articles. It can’t reason over inconsistent data. It can’t magically understand metadata that even your administrators struggle to decipher. And it certainly can’t respect security boundaries that were never clearly defined in the first place.

What’s changed isn’t the importance of good architecture. What’s changed is the visibility of poor architecture.

The more I reflected on that idea, the more I realized there was a bigger story to tell. We spend so much time chasing the newest Salesforce features that we rarely stop to appreciate the fundamentals that make those features successful in the first place.

That realization sparked an idea for a new blog series.

Introducing Back to Basics

Rediscovering the Salesforce features and architectural principles that have been quietly preparing us for the age of AI all along.

Rather than focusing on the newest AI capabilities, this series will revisit the foundations that make those capabilities successful. Along the way, we’ll explore four pillars of an AI-ready Salesforce organization:

Metadata & Organization

The structures that help humans and AI understand your org, from naming conventions and field descriptions to documentation and metadata strategy.

Data & Trust

The quality, governance, and consistency of the data that powers every workflow, report, and AI interaction.

Security & Scale

Permission models, access strategies, and architectural decisions that create confidence as your organization and AI capabilities grow.

Architecture & Maintainability

The design choices that keep your org adaptable over time, from configuration and Custom Metadata Types to technical debt and long-term scalability.

These aren’t new ideas. Most have been part of the Salesforce platform for years. But viewed through the lens of AI, they suddenly feel different. Features that once seemed like administrative best practices are becoming strategic advantages.

My hope is that this series helps us look at familiar topics with fresh eyes. Not because the principles have changed, but because the context around them has.

So over the next several weeks, we’re going back to the fundamentals. Not because they’re old, but because they’ve never been more relevant.

Let’s get back to basics.