Salesforce used the opening day of Dreamforce 2026 to introduce AIforce, a new interface layer designed to bring Salesforce data, business logic, workflows, permissions, and actions into the AI tools where people already work.
That distinction matters.
For years, getting value from Salesforce largely meant logging into Salesforce, navigating an application, opening the right records, and working through a defined interface. AIforce begins to separate the power of the Salesforce platform from the traditional Salesforce user experience. Instead of requiring every employee to come to the CRM, Salesforce can now come to them through Slack, Claude, Lightning, Gemini, and other AI-powered experiences. This extends the direction Salesforce established with Headless 360, which exposes Salesforce capabilities through APIs, MCP tools, and commands that people and agents can use on any surface.
In practical terms, a user could ask an AI assistant to summarize an account, assess pipeline health, update an opportunity, surface open service issues, or initiate an approved workflow. The response and action would still be grounded in Salesforce context, permissions, governance, and business rules.
That makes AIforce more than a new feature. It represents a shift in how Salesforce expects people and other AI agents to interact with the platform.
The most important idea behind AIforce is that the Salesforce interface may become optional, while the Salesforce platform becomes even more essential.
Companies have spent years building valuable institutional knowledge inside Salesforce: customer histories, revenue processes, service procedures, automation, security models, calculated insights, and connections to other enterprise systems. AIforce is intended to make that context available to AI interfaces without rebuilding those capabilities separately for every assistant or application.
This is also why Salesforce's expanded partnerships with Google Cloud and AWS matter. Salesforce capabilities can increasingly be surfaced inside Gemini Enterprise, Amazon Quick, Slack, and other environments using open interoperability standards such as Model Context Protocol. At the same time, organizations retain the governance and access controls already established in Salesforce.
The strategy is clear: Salesforce does not need to own every screen if it remains the trusted system that gives AI the context and authority to act.
Salesforce also announced Koa, its first CRM reasoning model, developed with NVIDIA and built on Nemotron. Koa is designed specifically for complex, multistep business tasks such as qualifying opportunities, routing service cases, and coordinating follow-up actions.
While AIforce determines how Salesforce context and capabilities can reach different interfaces, Koa strengthens the reasoning available within Agentforce. Together, the announcements show Salesforce moving beyond AI that simply retrieves information or drafts content. The goal is AI that can understand business context, select the appropriate tools, and complete governed work across a process.
Salesforce says Koa was trained using synthetic enterprise scenarios rather than customer data and will operate within Salesforce's trust boundary. It is currently moving through customer pilots, with broader U.S. availability expected in winter 2026.
For revenue leaders, AIforce becomes especially interesting when paired with Agentforce Revenue Management. Revenue processes rarely stop at the opportunity. They extend through product configuration, pricing, quoting, approvals, contracts, orders, invoicing, and recurring or usage-based billing. Salesforce is bringing all of these areas together on a common platform and data model.
That creates the potential for an entirely different revenue experience. A seller working in Slack or another AI interface could ask for a renewal summary, identify an at-risk account, assemble the right product configuration, generate a compliant quote, or determine which approval is holding up a deal. Finance and revenue operations teams could investigate billing issues, contract changes, or consumption trends without first navigating a series of objects and screens.
This is not just theoretical. Agentforce for Revenue already allows sellers to generate quotes using natural-language instructions while applying the appropriate products, pricing, and terms. Salesforce reported a 75% reduction in quoting time and an 87% reduction in clicks from its own internal use. The Winter ’27 release expands that direction with self-service quoting for buyers, integrated partner quoting, and automated renewal package preparation for account executives.
AIforce broadens the significance of those capabilities. Revenue actions no longer have to begin inside Revenue Cloud. They can begin wherever the seller, partner, customer, or operations team is already working, while the catalog, pricing engine, approval rules, contracts, and billing processes remain governed within Salesforce.
This also makes the quality of the revenue foundation critical. An agent cannot reliably create the right quote when product rules are inconsistent, pricing logic is undocumented, entitlement data is incomplete, or approvals vary outside the system. AI may simplify the interaction, but it does not eliminate the need to design the underlying quote-to-cash process correctly.
The arrival of AIforce does not reduce the importance of a well-designed Salesforce environment. It increases it.
An AI assistant can only act reliably when it has clear data, consistent definitions, usable metadata, appropriate permissions, and dependable workflows. If account hierarchies are incomplete, customer identities are fragmented, automations conflict, or business rules live only in employees' heads, AI will inherit those weaknesses.
Organizations preparing for this next generation of Salesforce should focus on a few fundamentals:
The question is no longer simply, “Where can we add an agent?” It is, “Is our business architecture ready for people and agents to work through any interface?”
AIforce was the strategic story of Day 1 because it expands the reach of everything organizations have already built in Salesforce. Koa may have been the splashier product reveal, but AIforce points toward a larger change: Salesforce becoming the governed business layer beneath an ecosystem of AI experiences.
For customers, the path forward is not to chase every new AI announcement. It is to build the foundation that makes those capabilities useful: connected data, clear processes, secure access, and an architecture designed for action. For revenue teams, that means treating AI readiness and quote-to-cash modernization as the same conversation, not two separate initiatives.
Clean Salesforce foundations just became more valuable, not less.
Clean Salesforce foundations just became more valuable, not less.
Every Salesforce admin has a list. You know the one: the messy permission model you inherited, the SPIFF spreadsheet finance emails out once a month, the Workflow Rules that still need to move to Flow before Salesforce sunsets them for good.
WRITE ABOUT AI, but DO NOT WRITE WITH AI.
WRITE ABOUT AI, but DO NOT WRITE WITH AI.