Warm September weather is one of the many reasons to visit San Francisco this month, but for many visitors the draw is Dreamforce, the Salesforce flagship conference held this year from September 15 to 17.
That’s why I’m heading to the Moscone Center where I’m looking forward to three sessions and related presentations that tackle the exact problems my company solves every day: Getting enterprise knowledge into shape so people and AI systems can trust the source material.
I plan on seeing how the rest of the industry is approaching that problem, while also comparing notes with the leaders living through those issues firsthand.
Learn Who Owns AI Accountability When Answers Go Wrong
The first session on my list is the executive roundtable Who’s Accountable When AI Gets It Wrong? Many companies I talk to that have rolled out AI assistants eventually hit the same wall. First, the AI gives a bad answer. Then nobody can agree on who’s responsible for coming up with the right answer. Chaos! Is it the model? The prompt engineer? Or the documentation team? The support team? Assuming they can agree on the right answer, where should it go? A FAQ? The knowledge base? The documentation?
Take a deep breath and relax, models and prompt engineers. You’re usually good. My experience says it’s almost always the documentation or the knowledge base. I’m looking forward to this session, which I hope offers up ways to streamline much-needed accountability for enterprise knowledge.
Compare Notes on Building a Foundation of Trusted Data
I’m also attending back-to-back sessions Data Leaders Roundtable: Trust, Strategy, and AI and Data 360 Roadmap: Activating Trusted Context Everywhere, which approach the same challenge from two angles: Data governance and content structure. My team spends its days doing exactly this work. They do the hard yards of auditing scattered knowledge bases, rebuilding information architecture, and turning stale PDFs and forgotten SharePoint folders into the enterprise knowledge an AI tool can leverage effectively.
Both sessions are also framed as roundtable and community discussions rather than single-presenter sessions, which is exactly why I want to be in the room: I’d rather hear how a dozen leaders are solving this than watch one vendor pitch a solution.
Watch How Governance Teams Unblock Stalled AI Launches
I’ll also spend time in sessions tagged Agentforce and Data 360, plus the Campground demos, where security and governance teams show up because permissions and auditability are holding their AI rollout hostage. This is a transformation problem I hear about often. Or maybe another transformation bottleneck. Even when a company builds a great AI assistant, the team sometimes discovers the thing isn’t ready for launch because of these governance issues.
Why? Sometimes because role-based access was overlooked, or because policy compliance hasn’t been vetted. More unnecessary chaos. I’m curious about what approaches these sessions identify. I hope to see how vendors and practitioners are solving this for Salesforce customers as well as those who are platform agnostic.
Find the Gap Between What Vendors Promise and What Teams Need
The weather might be beautiful outside, but I’ll be inside, walking the show floor because I want to better understand the gap between the AI tools vendors are selling and the groundwork companies still need to complete before solutions work. Surely clean content, a defined lifecycle, and experts who can translate complexity into something both a customer and a model can understand should be a major focus. That gap is where Expert Support lives, and I expect Dreamforce to be a great place to see it up close, at scale, across every industry in one building.
If you’re attending Dreamforce and any of this sounds like your Monday morning, I’d like to meet. Find me at one of these sessions, catch my updates on LinkedIn (@psgustafson) throughout the week, or reach out directly. I look forward to the conversation.
Paul Gustafson is CEO at Expert Support. A long-time Silicon Valley veteran, Paul has helped hundreds of clients optimize their technical communication strategies.
Expert Support helps technology companies explain complex technical knowledge for the people and AI systems that depend on it.