Dreamforce 2026 – 05 Takeaways
This year, Daeu AIon GDC Vietnam had the opportunity to be part of Dreamforce ’26 – learning, connecting, and exchanging ideas with technology leaders, innovators, partners, and businesses from around the world.
Beyond the sessions and product announcements, Dreamforce offered a broader view of where enterprise technology is heading.
One message stood out clearly: Enterprise AI is moving beyond copilots and chatbots. The next phase is about reasoning, agents, trusted data, orchestration, and action.
Here are our 5 key takeaways:
01. CRM is becoming the foundation for enterprise AI
AI can generate insights, reason about possibilities, and suggest what could happen next.
But AI alone does not know the reality of a business.
That reality comes from customer data, business processes, relationships, and context.
This is where CRM becomes critical.
AI tells you what is possible.
CRM tells you what is true.
The value of AI comes from connecting intelligence to the business reality captured in CRM.
02. Salesforce is evolving into an Agentic Enterprise platform
Dreamforce ’26 demonstrated how Salesforce is bringing multiple layers of enterprise technology together into an agentic architecture:
AIforce → Interface
Bringing Salesforce capabilities into the places where people work, including Claude and Slack.
Agentforce → Digital Workforce
Agents that can reason, collaborate, and execute business processes.
Customer 360 → Business Applications & Semantics
Connecting customer relationships, applications, and business context.
Data 360 → Trusted Context
Providing connected and trusted data for AI.
As AI becomes increasingly embedded in business processes, trust and data protection are becoming fundamental to enterprise AI adoption. One important capability within Salesforce’s Einstein Trust Layer is its Zero Data Retention (ZDR) policy. With Einstein Trust Layer’s Zero Data Retention (ZDR), when a prompt is sent to a third-party LLM (like OpenAI or Anthropic), the provider is required to discard the prompt and response after processing and not store or train on it, with Salesforce also applying controls like data masking and permissioning. This creates an important layer of protection for enterprises using AI with sensitive business and customer data.
03. AI is becoming more specialized and context-aware
One of the examples was Koa, Salesforce and NVIDIA’s CRM reasoning model designed for complex CRM and Agentforce tasks.
The bigger takeaway is not simply another AI model.
It is the combination of:
AI Model + Enterprise Data + Domain Knowledge + Business Context
For enterprise use cases, intelligence needs to be relevant to the domain and grounded in the context of the business.
The question is increasingly not:
“How big is the model?”
but:
“How effectively can the intelligence solve a specific business problem?”
04. Copilots are evolving into a digital workforce
AI is moving beyond simply assisting people—it is increasingly working alongside them, connecting across systems, and taking on real business tasks.
At Dreamforce ’26, this evolution was reflected across the Salesforce ecosystem:
- Claudeforce brings Salesforce capabilities directly into Claude, with 37 prebuilt Salesforce skills, allowing users to work with Salesforce data and workflows through natural language.
- Slackforce brings AI agents and Salesforce intelligence into the flow of work, enabling teams to collaborate with agents directly within Slack.
- Agentforce Coworker takes this concept further by positioning AI agents as digital coworkers that can take on defined roles, collaborate with human teams, and execute tasks across business processes.
- Agentforce Multi-Agent Orchestration enables agents to reason, coordinate, use tools, and execute workflows, rather than simply answer questions.
The operating model is shifting from:
Human + Software
to:
Human + AI Agents + Software
And increasingly toward:
Human + Digital Workforce + Software
This is more than automation. It represents a new way of organizing, executing, and scaling enterprise work.
Real-world examples make this shift tangible:
- Piper: an inbound AI SDR (Sales Development Representative) hat engages website visitors 24/7, answers questions, qualifies prospects, and books meetings.
- Hunter: an outbound prospecting agent that uses CRM, intent, and third-party data to build and prioritize target accounts and initiate outreach.
- Marshall (Operations Agent): a back-office/process agent that helps orchestrate and execute repetitive operational work (ex: extracting data from docs, updating spreadsheets, running comparisons/calculations, spotting delays/compliance gaps) with humans in the loop and an audit trail.
AI agents are becoming part of the workforce – not simply tools used by the workforce.
They can take on repetitive and operational tasks, while humans remain focused on strategy, relationships, creativity, and decisions that require human judgment.
05. SaaS is transforming - not disappearing
The discussion around the “SaaSpocalypse” reflects a fundamental change in enterprise software.
The shift is from:
“Software that humans” operate to “Software that AI agents operate on behalf of humans.”
But this does not make applications irrelevant.
Instead, applications increasingly become the layer that provides the data, business logic, workflows, and governance behind the AI experience.
In other words, SaaS is not necessarily disappearing.
Its role is evolving.
Our Perspectives
For us at Daeu AIon GDC Vietnam, Dreamforce ’26 highlighted a broader transformation:
Data → Intelligence → Agents → Orchestration → Action
And this transformation is not about AI alone.
It requires the right combination of:
Trusted Data + Business Context + Applications + Integration + Governance + AI
The future of enterprise technology will increasingly depend on how these layers work together — enabling people and AI agents to work with the same business data, applications, and processes.
The future is not simply AI replacing software. It is AI transforming how people, data, applications, and business processes work together.
Let’s rewatch Dreamforce 2026 Keynote:
