Dreamforce 2025 brought a lot to the table, from major product announcements to new developments around Agentforce, Data 360, and the idea of businesses becoming more agentic. It was pretty clear that AI agents were moving from something people were only discussing to something businesses were actually starting to use.
And now, Dreamforce 2026 comes at an interesting point. AI agents are no longer just about answering a customer question or writing a quick summary. They are starting to handle tasks, work across systems, use business data, and even take actions on behalf of teams.
So, what comes next? In this blog, we will look at how AI agents are already changing everyday business work, what could become more important at Salesforce Dreamforce 2026, and why the next phase may be less about simply having AI and more about having AI that actually gets things done.
The Shift From AI That Answers to AI That Acts: What AI Agents Are Already Doing
AI agents for business are already being used for work like lead qualification, follow-ups, customer research, record updates, and routine sales tasks. Salesforce is also expanding Agentforce across sales, service, commerce, marketing, and operations.
The interesting part is that agents are not just giving suggestions anymore. They can take actions across workflows, which means sales teams can spend less time jumping between small tasks and more time actually talking to customers.
A chatbot can answer a question. An agent can potentially understand the request, find the right information, take an action, and continue the workflow.
For example, an agent can help with:
- Finding customer information
- Handling routine service requests
- Updating records
- Escalating complex cases
- Following up after an interaction
That difference is pretty important because the customer does not really care whether the work was done by a chatbot or an agent. They just want the problem handled.
A lot of business work happens behind the scenes, and this is another area where agents are starting to help. Salesforce has already introduced Agentforce Operations for back-office workflows, showing that agents are moving beyond sales and customer conversations.
They can potentially help teams with tasks such as:
- Processing routine requests
- Finding and organizing information
- Triggering workflows
- Handling repetitive administrative work
- Supporting internal teams
And this is probably where businesses will discover a lot of their real value, because not every useful task is customer-facing.
Here is the very important part. Agents need good information to make good decisions.
Gartner predicts that by 2028, AI agents could outnumber sellers by 10 to 1, but fewer than 40% of sellers are expected to say agents improved productivity. The message is pretty clear: more agents alone will not fix broken processes.
Businesses will need to focus on:
- Clean and connected data
- Clear workflows
- Proper permissions
- Human oversight
- Measuring actual business results
Therefore, the future is not simply about adding agents everywhere. It is about putting them in the right places, with the right information, and letting them do work that actually matters.
The next big thing is agents working across different systems instead of staying trapped inside one application.
Salesforce is already showing this direction through its broader agentic platform and integrations, while Dreamforce 2026 itself is putting Agentforce, Slack, Data 360, and other technologies together in the Agentic Enterprise experience.
So instead of an agent saying, “sorry, I cannot access that system,” the future could look more like the agent finding the information, moving it through the required workflow, and coming back with the work already handled.
Shopify is also moving quickly toward agentic commerce. Its Agentic Storefronts are designed to make products discoverable through AI channels, while Sidekick can help merchants with store tasks, orders, products, content, and more.
This means Shopify AI agents are not only about helping shoppers find products. They can also help merchants manage parts of the business that happen before and after the purchase.
For ecommerce teams, the possibilities include:
- Product discovery
- Customer questions
- Order-related tasks
- Store management
- Product updates
- Connecting commerce activity with other business systems
Sync Made Easy also brings an AI Agent into the Shopify-to-Salesforce workflow, helping businesses connect their Shopify store with Salesforce while making the data flow between both systems easier to manage.
And honestly, this is where ecommerce could get very interesting. A customer could interact with an agent, find a product, complete a purchase, and have the relevant information flow into the business systems without someone manually moving everything around.
What We Can Expect From Dreamforce 2026
Salesforce Dreamforce 2026 takes place September 15–17, 2026, in San Francisco and on Salesforce+. Salesforce has already announced 1,600+ breakout sessions, 50+ keynotes, 150+ hands-on trainings and demos, and 240+ community roundtables.
So there is obviously going to be a lot happening, but the agentic theme is difficult to miss.
The current Dreamforce agenda includes sessions such as:
- 5 Steps to Build Your AI Roadmap With Confidence
- Accomplish More Every Day with Slack AI and Skills
- Accelerate Quotes with Agentforce Revenue Agents
- Accelerate Complex Data Management with AI Agents
- 5 Steps to Cut IT Costs and Boost Satisfaction with AI Agents
- 3 essentials to Scale Multi-agent AI Across the Enterprise
This suggests the focus will not only be on showing what agents are. There will be plenty of attention on building them, connecting them with data, securing them, and actually putting them to work.
Dreamforce 2026 is also highlighting sessions around bringing Slack and CRM together, using Slack actions with agents, and building multi-agent experiences.
That is important because businesses rarely operate from one system. Sales might be in Salesforce, conversations might happen in Slack, orders might sit somewhere else, and data can be spread across several places.
Dreamforce seems to be heading toward the bigger picture here, how all those pieces can work together instead of becoming another collection of disconnected tools.
One of the biggest things to watch will obviously be Agentforce and how Salesforce expects businesses to build and manage their agentic environments.
The main keynote is set to focus on the Agentic Enterprise, with Salesforce CEO Marc Benioff and guests exploring how humans, agents, and platforms can work together for customer success.
Salesforce Dreamforce 2026 is not only about listening to big announcements. Developers will get hands-on opportunities around Agentforce, Headless 360, Data 360, Apex, Flow, and other tools used to build agentic applications.
The Developer Keynote is specifically focused on building the Agentic Enterprise, while Trailblazer Forest will include technical demos and hands-on experiences.
Another important theme to watch will be how people actually work alongside agents.
Salesforce is positioning the Agentic Enterprise around humans and agents working together, rather than treating agents as something completely separate from employees.
And this could be where the conversation becomes much more practical. Which work should the agent handle? When should a person step in? What should require approval? These are the questions businesses will need to answer as adoption grows.
The Future Will Not Be Just About Having More Agents
The future is probably not about adding an AI agent to every single department just because we can.
It will be about choosing the right tasks, giving agents access to useful data, connecting them with existing systems, and making sure humans can step in when something needs a little extra thinking.
Gartner has already highlighted the growing issue of agent sprawl, predicting that by 2028, an average Fortune 500 enterprise could have more than 150,000 agents in use, compared with fewer than 15 in 2025.
That is a huge jump, and it also means businesses will need to think about data, automation, user experience, and actual productivity instead of simply adding more and more agents.
The better approach will be to:
- Give each agent a clear responsibility
- Connect agents to trusted data
- Keep humans involved where judgment matters
- Track whether the agent is actually helping
- Avoid building five agents for one simple job
Moreover, the future will probably not be about replacing every person with an agent. It will be about giving people a group of digital helpers that quietly handle the repetitive parts of their day.
Conclusion
Dreamforce 2026 is arriving at a pretty interesting point in the AI journey. Agents are already handling real work, ecommerce is becoming more agentic, and businesses are now figuring out how to connect all of this without creating complete chaos.
AI agents for business will continue moving into sales, service, operations, commerce, and internal workflows, but the companies that get the most from them will probably be the ones that focus on useful work instead of simply collecting more agents.
Salesforce Dreamforce 2026 will give businesses a closer look at that future through Agentforce, Data 360, Slack, multi-agent design, and real customer examples.
And maybe that is the biggest shift happening here. The future of business software may not be about opening an app and doing the work yourself. It may be about telling the system what needs to happen, and then watching the work quietly get done in the background.