INDUSTRY EVENTS | 15-17 September 2026
dreamforce
2026 key takeaways
Salesforce Dreamforce 2026 marked a huge shift away from static software management toward an autonomous, agent-driven model. Here is a breakdown of the three biggest announcements and what they mean for you as a Marketing Cloud Next user.
Published: 23 September 2026 | Derek Bell
Photo by Kellen Riggin on Unsplash
Complete the Design for Automation mini-assessment today to identify strengths and gaps in your marketing automation strategy and processes.
Campaign Agent & Adaptive Campaign Audiences
Salesforce unveiled Campaign Agent as its main advancement for digital marketers, heading to general availability in Marketing Cloud Next Advanced by October 2026.
Salesforce promises that this release changes how you build and launch campaigns moving forward.
Historically, you had to map out every customer path by hand. You spent hours building complex logic in Journey Builder, setting up audience segments, triggers, branching rules, and entry limits to anticipate every move a customer might make.
Campaign Agent swaps this legacy path building for a goal-driven model. You set the campaign target, share customer context, set brand guardrails, and state your target outcomes. The AI then handles execution.
Here is what powers this change:
- Campaign Assembly: Builds the campaign structure straight from your objectives
- Adaptive Campaign Audiences: Turns campaigns into always-on engines, using AI to dynamically update target lists in real time so you don’t need static list pulls or SQL queries
- Agentic Content: Generates multi-channel asset variations and pulls brand-approved content from headless CMS tools like Contentful
- Campaign Prioritisation: Evaluates real-time customer context to pick the most relevant campaign for an individual at any micro-moment
This changes your day-to-day work. You’ll spend far less time drawing complex journey maps or sorting data by hand. Instead, you can focus on overall data structure, platform integration, and setting the right guardrails so AI agents operate safely.
Job-ready signal agents: Meet Piper & Hunter
Dreamforce 2026 also introduced job-ready agents built to transform signal-driven engagement: Piper and Hunter.
The best inbound and outbound marketers currently rely on static tools like web forms, lead scoring, and engaging email nurture flows. Piper and Hunter promise to turn real-time customer signals into instant AI action without waiting for batch processing.
- Piper (Inbound): Engages web visitors instantly, learns their intent, and adapts the conversation in real time. Rather than making prospective customers fill out static form fields and wait for an automated email, Salesforce see the future with Piper holding a live dialogue, surfacing documentation and guiding visitors toward conversion on the spot.
- Hunter (Outbound): Scheduled for general availability in November 2026, Hunter promises to take a proactive approach. Similar to how marketers launch trigger-based campaigns today and assuming all of the data needed is in Salesforce, Hunter will scan lifecycle signals like product usage telemetry and content engagement to find the best moment and channel for outreach, rather than sending emails on a fixed schedule
“When you’re working in Salesforce apps and when you’re working on our platforms, your data is your data. It does not go in the models… Your data is your data. You’re not training any other model when you’re using our products. We’ve audited it. We’ve tested it. We’ve tried it.“
Marc Benioff
CEO Salesforce.com
Dreamforce 2026 Keynote Address
Customer lifecycle management isn’t a straight line; it hasn’t been for many years. The internet kicked off the change in buyer behaviour in the 1990s. Again, assume all data on every purchase, meeting, or site visit lives in Salesforce; this will feed context into Piper and Hunter for the next interaction.
To make the most of these agents, you need clean CRM data and clear data governance. Because these tools make autonomous choices, running thorough evaluations and maintaining human oversight until models meet your standards is vital.
The promise is pretty clear and leaves essentially no room for anything other than what he’s articulated. We see a busy time ahead for IT, lawyers and marketers.
Putting that aside, let’s look further at the introduction of Claudeforce, Slackforce
AIforce, Koa & the headless* interface revolution
The biggest structural update at Dreamforce 2026 was AIforce, a live interface layer that changes how you interact with Salesforce applications.
Instead of requiring you to log into Marketing Cloud dashboards to run tasks, AIforce brings Salesforce data, workflows, and logic straight into the platforms you already use every day.
Powered by Koa, a CRM reasoning model built with NVIDIA and integrated through tools like Claudeforce and Slackforce, you can execute complex actions using plain language. You can ask Slack or Claude for campaign performance updates, adjust an audience parameter, or review new creative assets without opening a browser tab.
What this means for your team:
- Headless 360 architecture: Connects Marketing Cloud capabilities through APIs, Model Context Protocol (MCP) servers, and conversational plug-ins
- Existing governance: Keeps your current sharing rules and user profiles, running under Salesforce’s Zero Data Retention policy so customer data stays secure
- Permission audits are essential: Because anyone can ask an AI agent for information in plain English, broken permissions or messy sharing rules could expose sensitive campaign strategies or customer data
The future interface is AI-built around what you want to achieve. That means your primary focus must shift from UI customisation to strict data cleanliness and rock-solid permission control.
Which Salesforce have said ad nauseam.
* Headless architecture is a software design pattern where the frontend (the “head” or user interface) is completely decoupled and managed independently from the backend (the “body” or business logic/database).
Before we let Piper & Hunter loose, we have to discuss your data quality
Recent industry research shows a significant gap between AI ambitions and actual data readiness across marketing teams:
- Validity State of CRM Data Management Report: Only 21% of marketers believe their CRM data is adequately prepared to support AI agents, even though two-thirds of organisations have already started delegating decisions to automated tools. On top of that, 78% of C-suite leaders acknowledge making decisions based on AI suggestions they later questioned because of underlying data flaws
- Integrate and Demand Metric Data Governance Study: Just 24% of high-growth B2B companies consider most of their marketing data to be AI-ready (clean, governed, and accessible). For lower-growth organisations, that figure drops to just 10%, proving that upstream data governance is the primary blocker to automated execution
- Supermetrics AI Readiness Research: Only 6% of surveyed organisations strongly agree that their marketing data is high quality, accurate, and accessible across systems. Fragmented data remains the single biggest structural barrier preventing teams from turning agentic AI into dependable outcomes
These findings confirm that while platforms like Salesforce are releasing autonomous agents, siloed systems, poor CRM hygiene, and missing governance frameworks hold back most marketing departments. Until organisations establish clean data pipelines and automated upstream validation, deploying autonomous agents risks accelerating mistakes rather than driving growth.
Press sources
[1]: The Agentic Marketer. “Dreamforce 2026: the marketing funnel just got simpler.” Published 19 September 2026. Detail focuses on the shift toward goal-oriented Campaign Agents and the introduction of Piper and Hunter. SOURCE
[2]: Codiot. “Dreamforce 2026 Recap: What Was Announced and What It Means.” Published 18 September 2026. Detail focuses on the systemic rollout of AIforce, Koa CRM reasoning models, and headless architecture implementations across the Salesforce stack. SOURCE
[3]: ITPro. “Dreamforce 2026: All the big talking points from the annual conference.” Published 18 September 2026. Context surrounding Marc Benioff’s keynote on AI safety, model autonomy, and the launch of the AIforce UI replacement. SOURCE
Other sources
Validity State of CRM Data Report: Read the MarTech breakdown showing only 21% of marketers feel their CRM data is prepared for AI agents
Integrate and Demand Metric Data Governance Study: View the press release detailing why 24% of high-growth B2B teams have AI-ready data
Supermetrics AI Readiness Research: Access the Supermetrics research finding that only 6% of organisations strongly agree their marketing data is accurate and accessible
Frequently Asked Questions
Q: What major shift occurred at Salesforce Dreamforce 2026?
A: Dreamforce 2026 marked a transition from static software management towards an autonomous, agent-driven model. Marketing teams can now move away from manual journey mapping to goal-focused execution, whether they manage campaigns in Marketing Cloud Next.
Q: How does Campaign Agent change standard journey building?
A: Instead of requiring you to build complex logic, branching rules, and static list pulls by hand, Campaign Agent uses goal-driven AI to generate multi-channel assets and dynamically update target audiences in real time.
Q: What are the primary roles of Piper and Hunter?
A: Piper and Hunter are job-ready AI agents designed for instant action. Piper handles inbound web engagement by having live conversations with prospective customers, whilst Hunter scans lifecycle signals to find the ideal moment for proactive outbound outreach.
Q: What is AIforce and how does Koa support it?
A: AIforce is a live interface layer that lets you execute tasks using natural language inside daily tools like Slack or Claude. Powered by NVIDIA’s Koa CRM reasoning model, it removes the need to log into traditional dashboards.
Q: How do these AI features keep customer data secure?
A: Salesforce runs under a Zero Data Retention policy, meaning your customer data is never used to train external models. However, because agents accept plain-English commands, maintaining strict permission controls and sharing rules remains crucial.
Q: Why is data governance essential before deploying autonomous AI agents?
A: Research shows most marketing data isn’t AI-ready. Because agents make autonomous choices, flawed data can automate errors at scale. Establishing clean CRM hygiene and maintaining human oversight ensures your automated actions deliver trustworthy, growth-focused results.
