Five Sessions, Two Conferences, and a Chance to Give Back

I had the opportunity to speak at both Texas Dreamin’ and Midwest Dreamin’, delivering five sessions across Marketing Cloud, Agentforce, automation, MCP, and AI.

Each session covered a different topic, but they were all connected by a common theme: how we can use new Salesforce capabilities thoughtfully, rather than simply adopting them because they are new.

Texas Dreamin’

Stop Relying on Pardot Sync to Drive Salesforce Automations

The Account Engagement sync is incredibly useful, but over the years I have seen it used as the default solution for almost every automation involving Pardot and Salesforce. That often means extra fields, duplicated configuration, delayed processing, and logic spread across two platforms.

In this session, I showed how External Actions can provide a more direct path from Account Engagement into Salesforce Flow or Apex. Instead of waiting for a field to sync and trigger another automation, marketers can invoke the process they actually need.

My biggest takeaway is that External Actions are still flying under the radar. They are not the right solution for everything, but they can dramatically simplify certain designs and give marketing teams access to much more of the Salesforce platform.

Behind the Scenes: Building a Smarter Marketing Cloud MCP Server

I delivered this session at both conferences because MCP continues to be one of the areas I am most excited about.

It is fairly easy to wrap an API and call it an MCP server. It is much harder to build something that is secure, multi-tenant, understandable to an AI model, and genuinely useful to the people relying on it.

I shared what I learned while building a Marketing Cloud MCP server, including decisions around OAuth, business-unit access, telemetry, least privilege, and tool design.

The lesson that continues to stand out to me is that the protocol itself is not the most valuable part. The real value comes from deciding what complexity the model should not have to figure out. Good MCP tools should reflect user intent, not merely expose a long list of API endpoints.

Midwest Dreamin’

Agent Script: The Hybrid Reasoning Revolution

Agent Script interests me because it addresses one of the biggest tensions in enterprise AI.

Traditional automation is predictable and auditable, but it can be rigid. AI agents are flexible and conversational, but they can also be inconsistent.

Agent Script provides a way to combine the two. The language model can manage interpretation and conversation, while deterministic logic controls important rules, transitions, calculations, and system updates.

My key takeaway is that reliable agents will not come from better prompting alone. They will come from knowing where to let the model reason and where to deliberately constrain it.

AI Uncensored: An Honest and Overdue Conversation

I also joined a panel focused on having a more honest discussion about AI.

There is no shortage of excitement around AI, but there is still a need for practical conversations about what is working, what is failing, and where teams may be quietly creating technical debt in the name of innovation.

We talked about generative AI, agents, implementation patterns, governance, and the lessons admins, architects, and builders are learning in real organizations.

What mattered most to me about this panel was making space for honesty. AI can absolutely improve the way we work, but only when teams are willing to move beyond demos and ask harder questions about trust, maintainability, security, and long-term value.

Why Speaking Matters to Me

The Salesforce community and ecosystem have played a major role for me over the last 8 years.

I’ve learned from people who shared their ideas openly, answered questions, published examples, reviewed my work, and encouraged me to keep growing. I would not be where I am today without that generosity.

That is why speaking matters to me.

Sharing what I have learned is one way I can give back to a community that has given me so much. It is also how the ecosystem continues to thrive: people learn something, test it, improve it, and then share those lessons with the next person.

I am grateful to Texas Dreamin’ and Midwest Dreamin’ for giving me the opportunity to contribute, and to everyone who attended, asked questions, challenged an idea, or continued the conversation afterward.