Dreamforce 2026: Salesforce Unveils AIforce, Autonomous Marketing and Next-Generation AI Agents

Artificial intelligence continues to shape the agenda of the largest marketing and technology conferences in 2026. Dreamforce 2026, held September 15–17 in San Francisco, provided another example of this trend.

At its annual conference, Salesforce introduced updates to its CRM ecosystem, including AIforce, new Agentforce capabilities, and tools for automating marketing, sales, and customer service.

Speakers and participants included Salesforce executives such as Marc Benioff, along with major AI industry figures including Sam Altman, Dario Amodei, and Jensen Huang.

The central concept was the Agentic Enterprise: an organization in which AI agents can access corporate data, business processes, and tools to perform tasks autonomously under employee supervision.

For marketers, Salesforce outlined several development areas: autonomous campaign creation and optimization, brand visibility management in AI search, personalized customer experiences, and automated B2B lead generation.

Some of these products became available in September and October, while others were introduced in beta or pilot programs with commercial launches planned for later.

AIforce: Salesforce Opens Its CRM to ChatGPT, Claude, and Other AI Interfaces

One of Dreamforce's main announcements was the new AIforce architecture.

Traditionally, working with Salesforce requires employees to use CRM interfaces: open an application, select a section, analyze data, and perform an action.

AIforce offers a different approach. Salesforce's corporate data, business logic, user permissions, and automated workflows can now be made accessible to external AI interfaces.

For example, a manager can work with CRM information directly through Claude or Slack without switching to Salesforce for every action.

Core Components of AIforce

• Claudeforce. A Salesforce integration with Claude that allows users to access corporate data and perform permitted actions through an AI assistant.
• Slackforce. Tools for interacting with Salesforce data, agents, and business processes directly within Slack.
• Agentforce Coworker. An autonomous AI assistant within Salesforce that can analyze information, plan a sequence of actions, and execute tasks using connected tools.

Headless Architecture and Permissions

At the heart of AIforce is the concept of Headless Architecture. Instead of tying CRM functions to a particular user interface, Salesforce provides access to its data and business processes through APIs and AI agent tools.

Organizations can therefore use their existing corporate infrastructure in new workflows without completely rebuilding the CRM.

Salesforce places particular emphasis on preserving existing access controls. AI agents are expected to operate within established permissions and security policies rather than receive unrestricted access to corporate data.

Availability at announcement: Agentforce Coworker was generally available. Claudeforce was in beta, and Slackforce general availability was planned for October 2026.

Campaign Agent: Autonomous Campaign Creation and Optimization

Campaign Agent was one of Dreamforce's most significant announcements for marketers.

Salesforce presented it as an AI manager for advertising and marketing campaigns capable of independently handling a substantial share of tasks that traditionally require manual work.

Instead of building audiences, content, messaging journeys, and delivery rules step by step, a marketer defines the campaign objective.

For example, a company might aim to increase repeat purchases among customers who have not placed an order in the last 90 days. Campaign Agent uses customer information, interaction history, and constraints set by the marketer to build the campaign.

How the Campaign Agent Workflow Operates

1. Content creation. The AI agent uses marketing briefs, brand guidelines, and customer data to prepare messages and variations for different audiences and communication channels.
2. Audience creation. The system identifies suitable customer segments based on available data and campaign objectives.
3. Customer journey creation. Campaign Agent combines audiences, messages, and channels into sequences of customer interactions.
4. Delivery management. Rather than following only a predefined sequence, the agent can select the most appropriate message, channel, and communication time for an individual customer.
5. Post-launch optimization. The system analyzes available customer behavior and campaign performance signals, then adjusts subsequent marketing actions within established constraints.

For instance, when a customer has already purchased, the system can remove that customer from a promotional message that is no longer relevant. Another customer may receive alternative content or move to the next stage of a customer journey.

Salesforce is therefore proposing a shift from static marketing automation sequences to campaigns that continue adapting after launch.

Under Armour Example and Business Implications

At the conference, Salesforce highlighted Under Armour, which works with an audience of more than 50 million subscribers across channels.

At this scale, manually preparing individual messaging sequences for every customer is practically impossible. Campaign Agent is designed to automate a substantial portion of communication decisions.

Autonomous management, however, does not eliminate the need to define business objectives, maintain customer data quality, and establish guardrails for AI.

Availability at announcement: Salesforce identified October 2026 as the planned general availability date for Campaign Agent.

Palmata and Contentful: Salesforce Enters Answer Engine Optimization

Another major Dreamforce topic was managing how brands appear in answers generated by AI search platforms.

Salesforce views ChatGPT, Gemini, and other AI services as additional channels through which customers discover brands.

Unlike conventional search engines, AI assistants do more than provide a list of links. They can compare products, recommend solutions, and shape users' perceptions of a company before they visit its website.

To address this opportunity, Salesforce presented Palmata, an Answer Engine Optimization platform developed by Contentful. Salesforce completed its acquisition of Contentful on September 1, 2026, incorporating its technology into the development of its marketing ecosystem.

What Palmata Measures

Palmata is designed to analyze brand visibility in AI-generated answers. Its capabilities include:

• Tracking brand mentions in generative search responses.
• Analyzing how AI systems describe the company and its products.
• Comparing brand visibility with competitors.
• Identifying information sources that influence AI-generated answers.
• Recommending content changes to improve visibility.
• Modeling the potential impact of proposed changes.

Connecting AI Visibility with Content Management

Salesforce's approach is distinctive because it connects AI visibility analysis with content management.

Contentful provides infrastructure for storing and publishing structured content through APIs. This allows organizations to use a common information base for websites, mobile apps, marketing communications, and AI interfaces.

For example, Palmata might discover that AI search services are describing a product's features inaccurately. The system can identify potential sources of the problem and suggest changes.

Through integration with Contentful and MCP tools, the corresponding content can be updated within an existing workflow.

However, such an integration does not guarantee any particular brand placement in ChatGPT responses. Generative search results depend on independent AI platforms' algorithms and information sources.

Availability at announcement: Contentful had introduced Palmata in June 2026, and it was already offered to users at Dreamforce. Its inclusion in Salesforce was part of a broader AI Search and marketing content strategy.

Headless Marketing: Managing Marketing Cloud Through AI Assistants

The development of AIforce also extended to Marketing Cloud. At Dreamforce, Salesforce introduced Headless Marketing, an architecture for interacting with marketing tools through external AI interfaces.

The solution is built on the Model Context Protocol (MCP), which allows compatible AI agents to connect to Marketing Cloud capabilities and carry out authorized operations.

For example, a marketer could ask an AI assistant:

“Find customers who made their first purchase in the last 30 days but have not placed a second order. Prepare an email communication journey for them.”

A connected agent could then use the relevant Marketing Cloud tools to work with that audience and automation.

Depending on permissions and available functionality, the agent can interact with customer data, content, marketing journeys, and communication settings.

The essential change is that employees no longer have to carry out every operation directly through the Marketing Cloud interface.

However, Headless Marketing is not a standalone replacement for Salesforce. Agents still require appropriate data, a configured infrastructure, and permission to perform actions.

Availability at announcement: Salesforce announced October 2026 as the planned general availability date for Headless Marketing.

B2B Lead Generation: Account Discovery, Piper, and Hunter

Dreamforce also focused on automation for B2B marketing and sales.

Salesforce introduced tools designed to work not only with individual leads, but also with companies, potential buying groups, and sales opportunities.

Three products stood at the center of this strategy: Account Discovery, Piper, and Hunter.

Account Discovery: Identifying Potential Buyers

Account Discovery is designed to identify companies showing potential interest in an advertiser's products. It analyzes available engagement signals and helps identify promising accounts and buying groups.

For example, the system might detect that multiple employees from one company are interacting with an enterprise software provider's content. This could suggest that demand is developing at the organizational level.

Such signals, however, do not guarantee an approved budget or readiness to purchase.

Piper: Inbound AI SDR

Piper is an AI SDR agent designed to handle inbound inquiries and interact with prospective buyers.

It can answer website visitors' questions, conduct initial qualification, provide product information, and schedule meetings.

Piper supports text, voice, and video communications. Interaction data is recorded in CRM, allowing the commercial team to continue working with the prospect.

Hunter: Outbound Sales and Opportunity Development

Hunter focuses on outbound sales and developing existing commercial opportunities.

It helps analyze CRM information, find promising engagement opportunities, and prepare personalized communications.

Salesforce views the products as parts of a connected process:

1. Account Discovery identifies promising companies.
2. Piper handles emerging inbound demand.
3. Hunter supports further opportunity development.

The primary objective is to reduce information loss when prospects move between marketing and sales.

However, automating a B2B pipeline still depends on the quality of CRM data and the accurate identification of target accounts.

Availability at announcement: Account Discovery was scheduled for general availability in October 2026. Piper was already available, while Hunter's expanded offering within the Agentforce portfolio was planned for November.

Marketing Intelligence: AI-Powered Budget Management and Data Quality

Salesforce also introduced Marketing Intelligence updates focused on automating analytics and operational tasks.

Budget Optimization Agent

One new tool was the Budget Optimization Agent, designed to manage the allocation of marketing budgets across channels according to predefined goals and planning periods.

Marketers can define budget constraints and desired outcomes. The agent analyzes available performance indicators and recommends or executes budget reallocations within configured parameters.

For example, a company could use this approach to adjust spending across several marketing channels.

Optimization still depends on the underlying performance data and the selected measurement methodology. Automation cannot independently fix an analytics system that incorrectly attributes channel contributions.

Data Guardian

Another product, Data Guardian, uses AI to identify data quality problems and compliance issues involving established information governance rules.

Its tasks include detecting suspicious changes, monitoring compliance with requirements, and supporting processes for correcting identified problems.

This matters particularly for marketing analytics because customer data quality directly affects segmentation, personalization, and AI agent performance.

Availability at announcement: Salesforce identified October 2026 as the planned general availability date for Budget Optimization and Data Guardian.

Personalizing Customer Journeys: Adaptive Web, Personalized Paths, and RCS

In addition to autonomous campaign management, Salesforce expanded its customer experience personalization capabilities.

Dreamforce included updates for websites, mobile apps, and marketing communications.

Website and Journey Personalization

• Adaptive Web: Quick Setup simplifies the launch of personalized web interfaces and is intended to make dynamic experiences based on customer data easier to deploy.
• Personalized Paths in Flow enables adaptive customer journeys in which each customer's next step depends on current signals and objectives. Unlike rigid sequences, the system can account for recent customer behavior and AI recommendations.

RCS Messaging and Mobile Flash Sending

• RCS Messaging expands mobile communications beyond traditional SMS with interactive buttons, images, carousels, two-way replies, and other elements, enabling richer messaging experiences.
• Mobile Flash Sending is designed to deliver urgent mobile notifications at scale. Salesforce positions it for communicating with millions of users when delivery speed is particularly important.

Formula 1 Example

Salesforce highlighted Formula 1 as an example of its infrastructure in use.

The company consolidated customer information from more than 100 systems, creating infrastructure for personalized interactions with fans.

According to Salesforce, this architecture can send millions of notifications in 90 seconds. These figures describe a specific implementation, not a universal performance benchmark for every Marketing Cloud customer.

Availability at announcement: RCS Messaging was already generally available. Salesforce targeted October 2026 for the general availability of Adaptive Web: Quick Setup, Personalized Paths in Flow, and Mobile Flash Sending.

Agentforce: Long-Running Tasks, Multi-Agent Collaboration, and the Koa AI Model

Alongside its specialized marketing products, Salesforce introduced new Agentforce capabilities.

Long-Horizon Agents

One major development was Long-Horizon Agents: agents designed to perform tasks over days, weeks, or months.

Unlike AI tools that finish their work after a single request, these agents can preserve task state and continue through a series of actions.

For example, an agent might be tasked with re-engaging inactive prospects. This would require analyzing CRM data, preparing communications, monitoring replies, and adjusting subsequent actions.

The architecture includes mechanisms for preserving context, resuming work after errors, and obtaining user approval for certain operations.

At Dreamforce, Long-Horizon Agents were in pilot, with general availability planned for November 2026.

Multi-Agent Orchestration and Agent Optimizer

Multi-Agent Orchestration allows several specialized agents to work together on one task. One might analyze a customer, another prepare an offer, and a third interact with CRM.

A primary agent coordinates the sequence of operations.

Salesforce also introduced Agent Optimizer, a tool designed to analyze deployed AI agents, identify recurring errors, and suggest improvements.

According to Salesforce, general availability for Agent Optimizer was planned for October 2026.

Koa: Salesforce's CRM-Focused AI Model

A separate technology announcement was Koa, Salesforce's specialized AI model built on NVIDIA Nemotron.

Unlike general-purpose language models, Koa focuses on CRM business processes. It is intended for tasks involving customer context analysis, corporate tool use, and sequential decision-making.

Salesforce emphasizes that additional model training used synthetic scenarios rather than actual Salesforce customer data.

This allows common sales, marketing, and service situations to be simulated without including confidential customer information in training datasets.

Internal Salesforce tests showed improvements in several metrics compared with the company's selected baseline models. However, these results are not an independent assessment of Koa's effectiveness in real enterprise deployments.

Availability at announcement: Koa was in pilot, with general availability expected in winter 2026.

Which Salesforce Capabilities Were Available and Which Were Still in Development?

Dreamforce announcements included both existing products and tools with planned launch dates. Based on the September 2026 announcements, Salesforce outlined the following availability stages:

Importantly, general availability does not mean automatic access in every Salesforce edition. Individual features may depend on licenses, region, system configuration, and additional activation requirements.

Announced launch schedules also should not be treated as confirmation of actual availability without checking the relevant Salesforce documentation.

Conclusion: Dreamforce 2026 Reinforced the Shift from Marketing Automation to Agentic Marketing

Dreamforce 2026 continued a trend visible at the year's biggest technology conferences: AI is becoming the foundation of enterprise process transformation rather than an additional tool.

Salesforce, however, presented a more specific direction for AI development. The company aims to bring customer data, marketing communications, CRM, business logic, and AI agents together within a single architecture.

For marketing, this represents a shift from automating isolated actions toward automating sequences of decisions.

• Campaign Agent can independently build and optimize communications.
• Palmata and Contentful extend capabilities for AI search.
• Headless Marketing makes marketing tools accessible through external AI assistants.
• Agentforce automates lead qualification and sales opportunity management.

At the same time, AI agent autonomy does not eliminate the fundamental challenges of marketing analytics. These systems still depend on the quality of first-party data, accurate audience identification, performance measurement, and permission management.

The main trend illustrated by Dreamforce 2026 is the gradual shift from preprogrammed marketing workflows to systems that independently select actions to achieve established business objectives.

Salesforce sees this transition as the foundation of the next generation of enterprise marketing and customer relationship management.

Sources

• Salesforce — 21 Things We Announced at Dreamforce 2026
• Salesforce — Dreamforce 2026: 5 Ways Marketing Cloud Is Building a New Model for Growth
• Salesforce — Marketing Cloud Innovations
• Salesforce — Salesforce Unveils AIforce
• Salesforce — Five Big Ideas Coming Out of Dreamforce 2026