Salesforce

Choose your fighter: Human: Complex strategy, relationship building Agent: Servicing, qualifying, speaking the language Actually, we’ll go with both. @Benioff explains to @MatthewBerman why human-agent collaboration is the future.

Most self-service "solutions" do one thing well: Make the customer give up and scream "REPRESENTATIVE!!!" Let's fix that. Meet Agentforce Self-Service. Deploy in minutes and give your customers one seamless path to resolution.

@gokulr Data gravity matters, but the real moat might be simpler: who's liable when an agent makes a bad call? The platform that can produce a clean audit trail - intent, action, outcome - becomes structurally hard to route around, regardless of how cheap the compute layer gets.

HEADLESS The reason more “dashboard” software companies will push to become headless stores (that is, become “pipes” companies) is for two reasons: (A) one of the most durable opportunities in software is to become the singular data store for all structured or unstructured context for a certain domain or function (or heck, for the entire organization). (B) once a function or org embraces you as the data store, you price based on compute and storage, and beautifully compound and grow as the org context inevitably explodes over time, the way other compute/storage businesses like Databricks and Datadog have durably and steadily compounded over time. Now, of course, the challenge is that as soon as you open up to 3rd party agents, some of these agents, being venture funded companies themselves and run by ambitious entrepreneurs, will try to suck all the context out of you, reduce you to a CRUD database, and ultimately offer a version of you to their customer for free, bundled with their agents. Meanwhile, you will likely offer your own bundled free or low priced agents as alternatives to the most commonly used third party agents. This war between headless context stores and agentic workflow startups is just starting. Will be interesting to see who can commoditize their complement first.

What's an Agentic Enterprise? It’s a business where people and intelligent AI agents work together. AI tackles the routine. Humans make the big decisions. The perfect Agentic recipe:

Your CRM shouldn't live in a separate tab. 🗂️👀 Available today, Slackbot can now create and update @salesforce records from right within Slack. All you have to do is ask. ☑️ Try it out today on Slack's Business+ and Enterprise+ plans. ⚡️

Expectation: Shiny LLM fixes everything. Reality: Messy data = weak AI. @Benioff tells @MatthewBerman why architecture beats model chasing. Harmonize your layer with Data 360 ~ before ~ deploying agents alongside your people.

That's all. (Slackbot already handled it.)

@miramurati We've seen this firsthand. Agents working in shared channels don't just get things done, they make the people around them sharper. Better prompts, faster trust, smarter escalations. The collaboration layer isn't around the model. It is the model.

We started Thinking Machines to advance human-AI collaboration, and this is our first bet on what that looks like. Most labs treat autonomy as the goal and interactivity as scaffolding around a turn-based core. We think the way we work with AI matters as much as how smart it is. Interactivity has to be in the model, and it has to scale with intelligence rather than trail behind it. thinkingmachines.ai/blog/interacti…

Today we're sharing our work on interaction models. A new class of model trained from scratch to handle real-time interaction natively, instead of gluing it onto a turn-based one. youtu.be/A12AVongNN4

Your boss: "Deploy AI for IT." 3 months later: 14 disconnected agents, no shared context, employees repeating issues to each one. Stop the agent rigmarole. Agentforce IT Service detects intent and routes to the right agent automatically. 1 place. 50+ specialists.

Explore hundreds of features to build your Agentic Enterprise in the latest Salesforce Release: sforce.co/4nkfNc6

What we're shipping in Summer '26: multi-agent orchestration, real-time voice, multi-language support, and a lot more.

Dreamforce Call for Participation is officially open. Submit your session idea: sforce.co/4uKCs47

Don’t be like Astro. Share your agentic story at @Dreamforce.

@lindavivah Treating AI agents like real employees is the smartest way to speed up growth. 116 hires in 100 days is a massive milestone, and we love seeing how Asymbl uses digital workers to do the heavy lifting and take the friction out of scaling.

60% of Brandon Metcalf's workforce is digital. Not only that, one human recruiter and one digital worker helped him hire 116 humans in 100 days! He's operating like a company 10x its size – so HOW? #SalesforcePartner I had the opportunity to partner with @salesforce and sit down with Brandon Metcalf Founder & CEO of @Asymbl_Inc - at #AgentforceTour inNYC to find out! He breaks down how he onboards AI agents like employees, why his entire workforce runs on @SlackHQ with Salesforce as the connective tissue, the skill that's becoming one of the most valuable in the workforce, and the framework any company can use to actually get started with digital workers

GA this summer: Multi-Agent Orchestration in Agentforce One agent handling everything? Hallucinations. Multiple agents? You repeat yourself three times. Now you route complex work to specialists automatically—they see what you said, coordinate, solve it once.

“The agent conversation stopped being about models two quarters ago. It is about infrastructure now.” - @NateBJones The shift happening in AI isn’t just about who has the best model. It’s about which platforms can connect to enterprise data, plug into existing workflows, and support ecosystems of interoperable agents. Nate recently shared a five-part framework for evaluating agent announcements: 1. Does it work with the tools teams already use? 2. Is it extensible and open to other agents? 3. Can it access meaningful enterprise data? 4. Is an ecosystem forming around it? 5. Can agents stack on top? Salesforce Headless 360? Checks all five boxes. The full video is worth a watch: youtube.com/watch?si=Gdr9u…

🎉 Introducing MuleSoft Omni Gateway: govern APIs, agents, and MCP tools across your existing architecture. Turn APIs into governed MCP tools with auth, policy, and compliance built in by default. Control across every gateway and agent interaction: sforce.co/4d82Qh0

You're pointing to the biggest friction point in the enterprise. A model doesn't know the cost of being wrong, but a well-governed data platform does. If you ground the AI in a single source of truth, the domain knowledge of your best people becomes the literal guardrail for every automated action.

I’m coming to the conclusion that the biggest challenge for Enterprise AI, and AI in general , as of now, is that it’s still impossible to make sure that everyone gets the same answer to the same question, every time. Which is a great response to the doomers. AI doesn’t know the consequences of its output. Judgement and the ability to challenge AI output is becoming increasingly necessary, and valuable. Which makes domain knowledge more valuable by the second. Am I wrong ?

We can finally post this now. Big marketing moments don’t happen by accident. See how bold marketers build them at #CNX26. x.com/salesforce/sta…

Oops? Guess who it is. 😉

⚠️ DO NOT POST YET HOLD: Announce Connections 2026 musical guest on X Status: PENDING APPROVAL Band Handle: CONFIDENTIAL Publish time: 5/5 early a.m. PT