1. The Evolution from Copilots to Autonomous Agents
The conversation has shifted from AI assistants that help humans to autonomous agents that perform tasks independently. Last year was about copilots; 2026 is about 'Agentforce,' where AI agents are designed to reason and execute complex, multi-step workflows
across sales, service, and marketing. Watch for demos showing agents operating without constant human intervention. This signals a move toward a digital workforce that collaborates with human employees, handling routine processes so people can focus on strategic work. The key is whether these agents are just glorified macros or can truly handle dynamic, real-world business scenarios.
2. Data Cloud's Central Role
AI is only as good as the data it's trained on. Salesforce has rebranded its Data Cloud to Data 360, positioning it as the essential foundation for its entire AI strategy. This platform is designed to unify customer data from all sources to give AI agents the context they need to make smart decisions. Pay attention to announcements about new data connectors, especially for unstructured data like documents and PDFs, and 'zero-copy' integrations that allow Data 360 to access data in external warehouses without moving it. This shows Salesforce knows that clean, accessible data is the biggest barrier to effective enterprise AI.
3. The Focus on Trust and Governance
As AI agents become more autonomous, the risk increases. Salesforce is putting its 'Trusted AI' principles front and center. Look for sessions on the 'Trusted Enterprise AI Harness' and 'Agent Fabric'—these are platforms designed to give IT departments control and visibility over AI agents. Key features to watch for include deterministic guardrails, which enforce specific rules (like requiring identity verification), and advanced security and observability tools. This is Salesforce's attempt to answer the biggest question from CIOs: how can we scale AI without losing control or compromising security?
4. The Shift to Consumption-Based Pricing
How do you charge for a user that isn't human? The traditional per-user, per-month licensing model breaks down when AI agents are doing the work. The signal to watch is the shift toward consumption-based pricing. Look for terms like 'Flex Credits' and per-conversation pricing. This model charges based on usage, such as the number of AI-led interactions or API calls. While it aligns cost with value, it also introduces budget unpredictability. Expect many conversations around AI FinOps and how to manage these new, variable costs.
5. 'Headless' Architecture Goes Mainstream
The 'Headless 360' platform is a critical, if technical, signal. It means all of Salesforce's capabilities—its data, logic, and AI—are accessible via APIs. This allows companies to build custom applications and experiences on top of Salesforce without being locked into its user interface. Essentially, it lets developers take Salesforce 'anywhere.' For enterprise AI, this is huge. It means an AI agent built on another platform could still leverage a customer's unified profile inside Salesforce, creating a more integrated and flexible ecosystem.
6. The Search for Measurable ROI
The initial hype around generative AI is over; now, customers want to see the return on investment. The market is moving past impressive demos to demand hard numbers on efficiency gains, cost savings, and revenue growth. Listen carefully to customer success stories. Are they vague testimonials, or do they feature specific metrics? Sessions focused on 'AI for ROI' and turning AI investment into measurable business value will be packed. The lack or presence of these proof points will determine how quickly enterprises move from experimentation to full-scale deployment.
7. Deep Slack Integration
Salesforce's acquisition of Slack continues to be a core part of its strategy. The integration is moving beyond simple notifications to positioning Slack as the conversational interface where work gets done—by both humans and AI agents. Watch for announcements around Slack AI and how it leverages customer data from Salesforce to automate tasks, summarize conversations, and assist teams. The partnership with Anthropic, making its Claude model the default for Slack, is another key indicator of this strategy. The goal is to embed CRM and AI capabilities directly into the flow of collaborative work.
8. The Rise of Industry-Specific Agents
Generic AI has its limits. The real value comes from AI that understands the specific language, processes, and regulations of an industry. Salesforce has been building out its industry clouds for years, and the next logical step is pre-built AI agents for sectors like financial services, healthcare, and life sciences. Look for announcements about agents tailored for specific roles, like financial advisors or medical case managers, that come with compliance and industry-specific knowledge baked in.
9. A New 'Copilot Studio' for Customization
While pre-built agents are useful, every business has unique needs. Salesforce announced a Copilot Studio, which is its answer to this. It’s a low-code platform that allows companies to customize existing agents or build their own using natural language. This is a critical signal because it empowers businesses to create AI that is tailored to their specific workflows and data. The success of this studio will depend on how easy it is for non-developers to use and how well it integrates with the underlying Data 360 and security guardrails.
