In September, Salesforce announced AIforce, its strategy for bringing Salesforce to any AI model, agent, or interface where your teams already work. To see where it fits, picture the Salesforce platform as four layers. Data 360 brings your data together. Customer 360 holds your workflows, data models, and business logic. Agentforce is where agents get built and run. AIforce sits on top as the live interface, carrying everything underneath out to wherever people are working.
AIforce also covers a few named experiences. Claudeforce is the partnership with Anthropic, and its product is Salesforce in Claude. Slackforce brings Salesforce into Slack. Agentforce Coworker lives in Lightning. The Headless Toolkit gives builders direct access through MCP servers, APIs, and developer tools.
Powering it all is the Headless Toolkit, formerly known as Headless 360, which brings specific technologies to make AIforce work. The Headless Toolkit makes the work you already do more powerful and more reusable. Your flows, your data models, and your security rules are now accessible to systems and agents outside of Salesforce.
To be specific: the goal of AIforce is not to replace the Salesforce interface you work in every day. Your Lightning App Builder, Reports and Dashboards, and the UI your users rely on are not going away. AIforce expands what the platform can do.
Instead, AIforce lets people work with Salesforce from the tools and interfaces they already use, rather than switching to the Salesforce UI. Have Slackbot bring records to you. Want to use Claude to talk to your Agent Script-powered Agent in Salesforce? Expose it via Model Context Protocol (MCP). The agentic tools you can use to read, analyze, and help maintain your data and automations are now as broad as any client that supports MCP – which is most of them these days.
In some ways, this shouldn’t be new to admins – this is similar to any API related automation or integration you’ve helped deploy. The big distinction is that AIforce and MCP bring a universal language for agentic clients to operate.
Let’s break down some of the components that power the AIforce strategy.
Headless is not a new concept. Developers have used the term for years to describe systems where the UI is separated from the backend logic. Salesforce has now embraced that same principle with the Headless Toolkit. Your flows, your automation, and your data models live in Salesforce as they always have, but they can now be experienced from anywhere: other apps, external systems, AI agents. You configure it once. It works wherever it’s needed.
The toolkit has five tools in it. Let’s look at each one:
Under the hood, the Headless Toolkit isn’t a new product you need to buy or a separate platform you need to learn. It’s an umbrella term for a suite of existing and new capabilities that are already built into Salesforce: 60+ MCP tools, 4,000+ existing APIs, and 220+ CLI commands, all accessible to authenticated users governed by the same security model you already manage.
AIforce includes tools for any other agent that can make API or MCP calls. Once a Salesforce Agent is written in Agent Script, an admin can create a custom MCP server, add the agent to the Agentforce Asset Library, and configure the access. It’s even easier with Salesforce in Claude: 37 prebuilt sales skills for Claude across 8 categories, delivered through the Claudeforce partnership.. Meanwhile, check out our growing family of agents for specific tasks like supply chain management with Marshall or customer service with Fin.
With AIforce and the Headless Toolkit, your automation goes further: The flows and logic you build can be triggered from outside Salesforce. Developers build the integration, but they rely on the data models, automation, and guardrails you’ve already created. Instead of recreating logic across tools, you centralize it in Salesforce and extend it outward, design once, reuse everywhere.
Developers building external applications who want to use your Salesforce configuration have two options. They can use the Headless Experience Layer (HXL), a developer tool that handles rendering your workflows natively in any UI, Slack, WhatsApp, voice, or mobile without building custom rendering code. Or they can call your Salesforce APIs and MCP tools directly and build their own UI.
The first option is faster and more consistent. The second gives developers more control over how the experience looks and behaves. Either way, what they build on is your configuration. Understanding these options means you can have an informed conversation when developers ask what’s possible, and increasingly, they will. With the Salesforce Multi-Framework, developers have more options than ever for how they build those frontends. But none of it works without the data models, automation, and guardrails you’ve already put in place.
And when the next platform or device comes along, your Salesforce configuration is ready. You future-proof the org without writing a line of code.
As AIforce is a strategy for enabling customers with AI, Salesforce Guardian is the group of products that keep it safe.
A common question with any new access layer is: What about security? AIforce leverages your existing security model: Every sharing rule, field-level security setting, and permission set is respected automatically, regardless of how the data is being accessed.
When data runs through Salesforce-provided MCP servers, it passes through the Trust Layer, ensuring your enterprise data remains secure, private, and compliant by default.
Four things travel with every API call, every MCP tool invocation, and every CLI command.
One thing to be mindful of: When an agent acts on behalf of a user, it can do anything that user is permitted to do. That means your permission model matters more than ever. As you open up your org to external systems and agents, it’s worth reviewing what your profiles and permission sets actually allow. The security model protects you, but only if you configure it thoughtfully.
That is about to change, though. In the near future, Salesforce plans to roll out Agentic Identity. With Agentic Identity, agents will have their own permissions and credentials. This will allow admins more precision in controlling the agents, including revoking them specifically.
The agentic era does not require you to start over. Everything you’ve built: from your data models and your flows to your security rules is exactly what makes AI agents useful in your organization. The better your configuration, the smarter your agents.
Your primary responsibility has always been to understand your business and encode that understanding into Salesforce. Every field you create, every flow you build, and every rule you configure generates and shapes metadata. That makes you the metadata expert in your organization. In the agentic era, that expertise matters more than ever. Your metadata is what gives agents meaning. It’s the semantic layer that tells them how your org works, how your business operates, and what your data actually represents.
Whenever you build out your flows, custom data models, or security protocols, you’re doing more than configuring Salesforce. You’re defining the logic, structure, and policies that agents will rely on to function correctly. Your skills in building flows, defining objects, and managing data are exactly what make headless solutions work accurately and at scale.
This also means that the impact of your role and decisions is expanding. The work you do inside Salesforce now has reached far beyond it. Every decision you make about how to structure data, automate a process, or govern access shapes what agents can do on behalf of your users, wherever they are.
Admins are managing more requests than ever, and let’s be honest, it’s getting harder to keep pace. Turning business ideas into production-ready apps and agents takes time. You’re constantly stitching together features, navigating complex data models, and toggling between Setup and different builders. At the end of the day, it means that you can’t keep […]
AIforce expands possibilities for admins Salesforce Admins are adaptable. Over the years, we’ve transitioned through many technology shifts. Look at automation– we went from workflow rules to Process Builder to becoming full-fledged-flowantics. When AI entered our scope, we quickly got up to speed on Prompt Builder and Agentforce, bringing AI into the flow of work, […]
I’m going to start this post with one of the boldest statements I’ve made in a long time: The way Salesforce Admins build apps has changed forever with the introduction of Builder Central. I’ll let that sink in for a moment. Okay, the moment’s over. Let’s get moving! I spent a few afternoons getting hands-on […]
