Integration Services
Business Consulting & Advisory
AI Solutions
AppExchange Development
Tailored Salesforce Engineering
Managed Services
Implementation
Salesforce Dreamforce 2026 runs September 15 to 17 at the Moscone Center in San Francisco[1], and it is easily the most interesting week of the year for anyone building on Salesforce. Three days, more than 1,600 sessions, and a keynote that tends to set the agenda for the next twelve months.
If you have been to a few of these, you know the shape of the week. Benioff opens. A big number lands. A customer tells a story that makes you rethink something you were sure about. By Thursday everyone is comparing notes on what it all means for next year’s roadmap.
We are writing this before any of it happens, so treat what follows as informed guesswork rather than reporting. We are curious to see how much of it we get right.
Table of Content
- Dreamforce 2026: What to Expect at Salesforce’s Biggest Conference of the Year
- Let’s Begin with The Theme: Becoming an Agentic Enterprise
- Expect the Conversation to Move from Capability to Control
- Watch What Happens to Pricing
- Listen for Proof, Not Promises
- Plan for the Boring Part, Because It Is Where Projects Die
- Stay Tuned
- FAQ
Let’s Begin with The Theme: Becoming an Agentic Enterprise
Salesforce has landed on “Becoming an Agentic Enterprise,” and the whole Dreamforce 2026 agenda sits underneath it[2]. Read that word “becoming” again. Two years ago, the pitch was that agents were arriving. Last year it was that they worked. This year the framing is a transition you are somewhere in the middle of, which is a quiet admission that most companies are not finished and plenty have barely started.
That one word tells you more about the tone of the week than any product name will. Our guess is that this year leans practical. Less spectacle, more plumbing. Fewer demos of one clever agent, and more sessions about running a lot of them safely.
Which is the more useful conference. It is just a less fun one to livetweet.
Expect the Conversation to Move from Capability to Control
If there is one shift we would bet on, it is this one. The question “can an agent do this” is no more in the limelight. The question now is whether you can predict what it will do, and prove what it did afterward.
You can see the direction in what Salesforce has already shipped this year. Deterministic logic sitting alongside model reasoning, so certain steps happen in a fixed order no matter how the conversation goes[3]. Controls over which external tools an agent is allowed to reach. Audit trails. Ways to govern the coding agents that now write a real share of the code.
None of that makes for a good demo. All of it is what your risk and compliance team will ask about the moment you propose putting an agent near a customer or a claim or a payment workflow.
So if you are attending with a buying decision in mind, consider skipping a keynote and checking the Dreamforce 2026 schedule for the security and governance sessions instead. That is where the answers you need tend to be, and they are easier to get into.
Here's Where Achieva's Work Sits Now: Building Agents That Behave Predictably
Watch What Happens to Pricing
This is the thread running under everything else, and it is the one nobody on stage will address quite as directly as you would like.
The old model was simple. More people using the software meant more seats, which meant more revenue. If agents do work that people used to do, that arithmetic gets confusing. The market spent the early part of this year treating it as an emergency, then spent the middle of the year deciding maybe not, and the honest answer is probably somewhere in between over a span of years rather than months.
What we expect at Dreamforce 2026 is a lot of energy behind consumption and credit-based pricing, framed as a natural evolution rather than a defensive move. Watch how the units get described. Watch whether anything genuinely outcome-based shows up, because that would be the real signal.
One practical note if you have a renewal coming up. When AI, data and automation all get bundled into one big deal, the vendor usually holds the stronger hand. Work out what you are actually using, line by line, before anyone shows you a package price.
Listen for Proof, Not Promises
The customer stories are the part of Dreamforce most worth your time this year. They are also the part most worth a little skepticism.
Expect large aggregate numbers. Tokens processed, work units delivered, cases resolved without a human. Those numbers are real. But they tell you how much work happened, not how much it was worth. A case resolved by an agent might be a case that should never have been opened.
Here is what we will be listening for instead. How long the deployment took from decision to production. How many people maintain it now. What broke in month three. And what the company stopped doing as a result. Saved time only counts if someone can tell you where it went.
If a session answers those four questions, it was worth the flight.
Plan for the Boring Part, Because It Is Where Projects Die
Every agentic AI conversation eventually arrives at the same place, and it is not the model. It is your data.
Salesforce has invested heavily here. Being able to reach data where it already sits, instead of copying it somewhere first, is a real step forward. But be clear-eyed about what that solves. Reaching your data faster does not clean it. If your customer records disagree across three systems, an agent will now surface that disagreement to a customer, confidently, at speed.
Our crystal ball view: the most quoted line of the week will not come from a keynote. It will come from a customer on a breakout stage saying they spent eight months on data before building anything, and that it was the right call.
Stay Tuned
That is our read going into Dreamforce 2026. We will be watching the announcements closely and thinking about what they mean for the people who have to turn them into a budget and a roadmap.
Two decades of Salesforce implementations have taught us one thing: the projects that go well start with an honest look at the org you already have.
More to follow after the keynote.
External Links:
