Salesforce has unveiled a broad collection of artificial intelligence products, cloud partnerships, customer deployments and workforce initiatives around Dreamforce 2026, led by the introduction of AIforce, its new interface layer for bringing Salesforce data and business logic into external AI environments, and Koa, its first CRM reasoning model developed with NVIDIA.
The announcements collectively illustrate Salesforce’s strategy for what it calls the Agentic Enterprise, where humans and AI agents work across company data, applications and business processes without requiring employees to remain inside traditional Salesforce interfaces.
Rather than treating AI as an additional feature within CRM applications, Salesforce is attempting to make its platform available wherever employees and agents already work, including Claude, Slack, Google Gemini Enterprise, Amazon Quick, and other environments.
AIforce Brings Salesforce To Any AI Interface
AIforce sits at the center of that strategy.
The new live interface layer exposes Salesforce data, workflows, business logic, semantics, permissions, security, and governance to AI interfaces outside the traditional Salesforce user experience.
Employees can ask questions, update records, and trigger workflows through other interfaces while remaining grounded in the information and permissions maintained inside Salesforce.
AI agents can simultaneously examine information across hundreds of records and connected systems, reason over that information, and perform actions using Salesforce business logic.
Salesforce is positioning AIforce as another layer in an architecture that includes Data 360, Customer 360 and Agentforce.
Data 360 provides harmonized and federated data, metadata and memory.
Customer 360 provides business logic, processes, permissions and actions across sales, service, marketing, commerce and other applications.
Agentforce supplies the digital workforce, while AIforce is intended to make that combined foundation accessible through virtually any AI interface.
AIforce maintains existing Salesforce permissions and business rules, meaning an agent should only be able to access information that the requesting user is authorized to see.
Salesforce is also applying Zero Data Retention, under which business information used to answer a request is not retained by the external model provider.
The architecture is designed to operate without introducing a separate permissions framework or requiring companies to migrate their existing Salesforce environments.
Claudeforce, Slackforce And Agentforce Coworker Extend The Platform
Salesforce is launching AIforce through several initial experiences, including Claudeforce, Slackforce, and Agentforce Coworker.
Claudeforce extends Salesforce’s relationship with Anthropic by placing Salesforce capabilities directly inside Claude through a prebuilt Model Context Protocol server.
Salesforce in Claude initially includes 37 prebuilt sales skills covering tasks ranging from prospecting to pipeline management.
A Salesforce Development plug-in for Claude Code includes more than 40 skills, access to Salesforce’s larger skills library, and specialized plug-ins capable of handling development tasks.
Deloitte, GitLab and Legora have already piloted Salesforce in Claude, which is now available to customers in beta.
Slackforce follows a similar philosophy by bringing Salesforce information and business actions directly into Slack conversations.
Slackforce Surfaces can assemble Salesforce, Slack and other connected information into interactive interfaces that teams can filter, explore and act on collaboratively.
Slackbot can reason over both Slack conversations and governed Salesforce information, potentially identifying accounts that are becoming inactive, examining related support cases and internal discussions, reassigning ownership, creating follow-up tasks and drafting customer communications.
Slack CRM also allows users to create accounts, log notes and update Salesforce records using prompts without opening a separate CRM interface.
Salesforce is further positioning Slack Code as a collaborative environment where developers and AI agents can produce code, prototypes and documents together inside Slack.
Agentforce Coworker brings the model directly into Salesforce’s Lightning interface.
The AI teammate can reason across accounts, activity, and historical information, surface insights, and perform actions while operating within existing Salesforce permissions and business rules.
Salesforce said 100,000 users activated Coworker during its first 35 days.
Fulton Bank moved from zero to more than 20 production use cases within weeks, supporting approximately 3,000 users across the organization.
Salesforce Opens Its AI Architecture To A Larger Partner Ecosystem
The architecture underneath AIforce is designed to remain open.
Salesforce’s Headless Toolkit exposes platform components through MCP, APIs, plug-ins, skills and development tools.
Its AgentExchange ecosystem allows partners to create and distribute interfaces, agents, applications, integrations, workflows and actions using that architecture.
Participants include Anthropic, Amazon Web Services and Google, along with AI builders such as Lovable and Vercel and software companies including Docusign, Gamma, Jasper, Ramp and Rippling.
Salesforce And NVIDIA Launch Koa CRM Reasoning Model
Alongside AIforce, Salesforce and NVIDIA introduced Koa, Salesforce’s first CRM reasoning model for Agentforce.
Koa is built on NVIDIA Nemotron and is intended specifically for complex, multistep enterprise workflows.
Salesforce developed the model by post-training NVIDIA Nemotron 3 Super using a proprietary synthetic dataset modeled on nearly three decades of CRM knowledge.
The company said Koa matches or exceeds leading model performance on its CRM benchmark while producing three times fewer errors on tasks such as updating opportunities, routing customer cases and scheduling follow-ups.
No customer data was used to train Koa.
Salesforce instead generated synthetic scenarios representing the reasoning and tool-use requirements of CRM work across more than 14 industries, including manufacturing, financial services, healthcare and travel.
The model was post-trained using supervised fine-tuning and reinforcement learning through NVIDIA technologies including NeMo RL, NeMo Gym and NeMo AutoModel.
Salesforce controls Koa’s model weights and operates inference within its own infrastructure so customer information does not leave its trust boundary.
Koa is already being used internally at Salesforce and is moving into customer pilots with 1-800Accountant, Baxter Credit Union, Engine, Formula 1, UChicago Medicine and Xero.
Koa is currently available to select Agentforce pilot customers and is expected to reach general availability in U.S. regions during winter 2026.
NVIDIA Technology Expands Into Missionforce
The Salesforce-NVIDIA partnership also extends into Missionforce, Salesforce’s platform for government and highly regulated organizations.
NVIDIA models and accelerated computing are being integrated into Missionforce so organizations can train, tune and operate mission-specific AI in environments such as private clouds, classified networks and air-gapped systems.
Post-trained NVIDIA models will initially support Missionforce Operations, which automates government processes including procurement, supplier management and logistics.
Missionforce Operations is already generally available in U.S. regions, with post-trained NVIDIA models expected to become available to select customers in October 2026.
Google Cloud Partnership Connects Salesforce With Gemini Enterprise
Salesforce is simultaneously expanding its strategic partnership with Google Cloud.
Hyperforce on Google Cloud is already handling live production customer traffic, and Salesforce plans to begin moving selected U.S. customers onto the infrastructure during the fourth quarter of 2026.
North American general availability is scheduled for November 2026, with additional geographic expansion during 2027, including Germany.
Salesforce’s headless architecture is also being connected with Gemini Enterprise.
Companies can surface Salesforce information and capabilities directly inside Gemini without building custom integrations.
The architecture exposes thousands of Salesforce capabilities to models and agents through the MCP standard, enabling workflows such as checking pipeline health, summarizing accounts and triaging service cases across sales, service, finance and advertising.
Tableau analytics can also be accessed through Gemini Enterprise while retaining governance, row-level security and semantic models.
Google Gemini models remain available inside Agentforce for autonomous AI agents and Salesforce’s Reasoning Engine.
Salesforce And Google Expand AI Commerce And Data Sharing
The Google relationship also extends into commerce.
Beginning in fall 2026, Salesforce Commerce Cloud merchants will be able to surface products within Google Search, including AI Mode, and the Gemini app.
Google’s Universal Commerce Protocol will allow shoppers to complete purchases through Google surfaces while merchants continue using Commerce Cloud for payments, compliance and order management.
The companies are also expanding Zero Copy integration between Salesforce Data 360 and Google BigQuery across additional regions and adopting Iceberg REST Catalog standards and Private Connect.
Salesforce’s catalog feed integration with Google is generally available, Universal Commerce Protocol integration is expected to become generally available in October, and expanded Data 360 Zero Copy functionality is scheduled for fall 2026.
AWS Partnership Brings Salesforce Data And Agents Into More Workflows
Salesforce announced another major expansion with AWS.
Salesforce business context can now appear directly inside Amazon Quick through the company’s headless architecture, allowing teams to retrieve pipeline information, account summaries, service cases and operational analytics without building custom integrations.
AWS frontier agents are also moving into Slack, beginning with AWS DevOps Agent and expanding to Security Agent, FinOps Agent and Partner Central agents.
That means engineering and operations teams can interact with AWS agents directly within existing Slack conversations rather than switching into separate infrastructure consoles.
Amazon Bedrock is also becoming a broader source of model choice for Agentforce.
Salesforce customers can access models from Anthropic and NVIDIA through Bedrock, with OpenAI models expected to come to Agentforce as well.
Bedrock applies Zero Data Retention, while Salesforce’s Trust Layer adds capabilities such as dynamic grounding and toxicity detection.
Salesforce and AWS are additionally extending Data 360 Zero Copy access to AWS Glue-managed Apache Iceberg tables, S3-backed Apache Iceberg tables, Amazon Aurora, Amazon RDS and SageMaker Lakehouse, building on support for Amazon Redshift and Amazon S3.
Informatica MCP servers are also being integrated with Amazon Bedrock AgentCore and Amazon Quick to make governed enterprise data accessible to AI agents.
Agentforce Voice will interoperate with Amazon Connect Customer using Agent2Agent support for multi-agent orchestration in fall 2026.
Salesforce Clouds are now available through AWS Marketplace in 33 countries, allowing eligible customers to apply AWS spending commitments toward Salesforce products including Agentforce.
Amazon Quick with Salesforce’s headless architecture, Agentforce model choice through Bedrock, expanded Data 360 Zero Copy capabilities and Informatica headless integrations are available now.
Siemens Deploys Agentforce Across Industrial Sales And Service
Salesforce also provided a concrete example of how its agentic strategy can operate inside a large industrial company through an expanded partnership with Siemens.
Salesforce Agentforce is being connected with Siemens Teamcenter Service Lifecycle Management so engineering information can move directly into customer-facing sales and service workflows.
A technician could identify the correct spare part for a specific machine before arriving at a customer site, while a sales representative could determine whether a proposed equipment upgrade is technically valid and manufacturable without waiting for engineering teams to manually respond.
Salesforce and Siemens said the combination could compress some industrial processes from weeks to hours.
The companies also pointed to aftermarket services as a particularly attractive opportunity because industry analysis suggests aftermarket revenue can grow approximately six times faster and carry roughly four times the margin of new-equipment sales.
Siemens has already deployed Agentforce for inbound sales development.
The company had been receiving more than 2,500 unqualified leads per month without an efficient way to determine which opportunities deserved seller attention.
Two AI agents now work together inside Sales Cloud to engage incoming prospects, qualify them and route stronger opportunities to Siemens’ approximately 18,000 sellers.
Siemens is now engaging 100% of those incoming leads across 132 countries.
Salesforce and Siemens are also planning to automate partner onboarding.
An AI sales development agent called Piper is expected to create personalized pitches and book qualified prospective distributors directly with partner account managers.
After a partner signs, Slackbot can coordinate certification, financial and risk reviews, while Agentforce Operations powered by Regrello can execute final onboarding steps in SAP.
The companies expect that model to reduce supplier onboarding from weeks to days.
Trailblazer Community Reaches 20-Year Milestone
The Dreamforce announcements also highlighted the human ecosystem Salesforce expects to build, govern and deploy these systems.
Salesforce’s Trailblazer Community is celebrating its 20th anniversary in 2026.
The community spans millions of administrators, developers, architects, engineers and technology professionals who build and customize Salesforce environments.
Salesforce said Trailblazers have built approximately 12 million applications, support 9.6 billion API calls per day and have produced approximately 156 million lines of AI-generated code.
More than 151 million Trailhead badges and 1.48 million certifications have been earned, while Salesforce’s community now includes more than 700 groups worldwide.
AIforce will provide those builders with more than 60 MCP tools and custom MCP servers, more than 30 coding skills, more than 4,000 APIs and more than 220 command-line interface commands.
Salesforce Commits $27 Million To Education
Salesforce is pairing the technology push with new investments in education.
The company announced $27 million in education grants, including $18 million for public school districts in San Francisco, Oakland, New York City, Chicago and Indianapolis.
Salesforce has now contributed more than $180 million to public schools nationwide, including $156 million to Bay Area public schools.
The company is also relaunching its Adopt-a-School program across more than 25 public schools, including 15 middle schools in the Bay Area, and approximately 100 students are expected to participate in mentoring and workshops during Dreamforce.
The remaining $9 million will support organizations focused on educators and career opportunities for students.
Recipients include New Teacher Center, Leading Educators, TNTP, aiEDU and Partners in School Innovation for teacher development, instructional coaching and leadership support.
CodePath, Genesys Works, Eat. Learn. Play. Foundation and Jobs for the Future will support hands-on learning and career-connected programs.
Salesforce cited measurable results from earlier education investments.
Oakland math teacher coaches supported by the company helped produce an eight-point improvement in standardized math scores, while middle-school math proficiency in San Francisco has increased nearly 8% since 2022.
Salesforce-backed computer science programs have reached more than 130,000 Bay Area students.
Among educators surveyed through grantee aiEDU, 91% reported considerable or significant improvement in AI literacy.
Salesforce’s broader philanthropy has provided approximately $970 million in grants since the company’s founding, while employees have contributed nearly 11 million volunteer hours.
More than 65,300 nonprofits and educational institutions use Salesforce technology for free or at discounted rates.
Salesforce and Marc and Lynne Benioff have collectively contributed more than $1 billion to the Bay Area over more than 27 years.
Missionforce Military Fellowship Targets AI And Technology Careers
Another workforce initiative is focused specifically on military personnel transitioning into civilian careers.
Salesforce launched the Missionforce Military Fellowship, a 12-week Department of War SkillBridge program that places transitioning service members alongside Missionforce Customer Success teams.
Participants receive Salesforce certification support, AI and consulting training and direct experience with technologies used across government and national security.
Those who complete the fellowship can pursue available full-time Missionforce positions.
The program builds on more than a decade of Salesforce Military programs offering veterans, active-duty service members and military spouses technology training, certifications and career resources.
Salesforce sees that experience as increasingly relevant as government agencies deploy AI and seek professionals who combine technology skills with an understanding of mission operations, public service and complex decision-making environments.
The company’s Missionforce platform is already being used in government-related deployments, including support for 9.2 million soldiers, veterans and military families through the U.S. Army Human Resources Command, modernization efforts associated with a $13.5 billion U.S. Air Force vehicle fleet and a $1.6 billion Department of Veterans Affairs contract.
Salesforce Pushes CRM Beyond The Traditional Application
Taken together, the announcements show Salesforce attempting to expand the definition of CRM well beyond a traditional application accessed through a browser.
AIforce turns Salesforce into an underlying context, governance and action layer that can appear through other AI interfaces.
Koa gives the company its own specialized reasoning model for enterprise workflows.
NVIDIA provides model and accelerated computing technology, Google Cloud and AWS extend Salesforce into additional infrastructure and AI environments, and Siemens demonstrates how agents can connect customer-facing work directly with industrial engineering information.
At the same time, Salesforce is investing in the developers, administrators, students, educators and transitioning service members that it expects will build and operate those systems.
The common strategy is to make Salesforce less dependent on users navigating a fixed CRM interface and more capable of serving as an underlying business platform that provides trusted data, permissions, workflows and actions to AI agents wherever enterprise work takes place.
KEY QUOTES:
“AI is creating an interface revolution. We are combining model intelligence with all the context that customers have built into Salesforce to create an intelligent, dynamic, composable system that is securely governed, built with Zero Data Retention, and designed to work with the core systems that already run your business.”
“The most valuable thing Salesforce has built isn’t our platform — it’s the accumulated knowledge of how enterprise business actually works. With Koa, the knowledge is put inside the model itself. We trained a reasoning engine that understands the structure of a deal, the lifecycle of a service case, and the workflows that vary across industries. That’s a different kind of intelligence, and it runs entirely inside your trust boundary.”
Marc Benioff, Chair and CEO of Salesforce
“AI is creating a much larger opportunity for software. Every company needs useful AI, tailored to its knowledge, expertise, and work. NVIDIA Nemotron open models give Salesforce the foundation to turn decades of enterprise expertise into specialized AI with Koa, creating a CRM model that can reason and securely take action.”
Jensen Huang, Founder and CEO of NVIDIA
“Enterprise AI only pays off when agents can understand a company’s business context and act on it. Gemini Enterprise brings Google’s most advanced AI together with a company’s data, applications and everyday workflows in one platform. By bringing Salesforce — where critical customer and business processes already live — into Gemini Enterprise, customers can build trusted agents that reason across their business context and securely take action.”
Kevin Ichhpurani, President of Global Partner Ecosystem at Google Cloud
“The companies moving fastest and delivering differentiation with AI are the ones building on their own data, with the applications and cloud infrastructure they already trust. That’s why removing the friction between environments matters. These new integrations with Salesforce give customers data access without migration, model choice at the right economics, and AI agents where their people already work.”
Rahul Pathak, VP of Data and AI GTM at AWS
“For decades, the expert knowledge of the engineer has been separate from the person servicing the machine in the field. We are closing an important gap. By embedding our digital twin into the commercial workflow, we are putting a virtual engineer in the hands of service technicians and salespersons. Siemens brings industrial AI, Salesforce enterprise AI. Together, they create a critical building block for an industrial AI operating system that connects engineering, operations and business.”
Roland Busch, President and CEO of Siemens AG
“AI is changing how national security organizations operate, but technology alone doesn’t create readiness. In the agentic era, the expertise of our service members — their understanding of the mission, judgment under pressure, and ability to lead through complexity — has never been more important. This fellowship helps transitioning service members pair that experience with new AI and technology skills so they can continue contributing to national security missions in their civilian careers.”
Kendall Collins, CEO of Missionforce at Salesforce
“Trailblazers have guided businesses through every wave of innovation for 20 years, and now they’re leading the way into the AI era. We’re grateful for this community, for how they show up, go first, and support an entire ecosystem of Trailblazers skilling up for what’s next. They’re not just part of the Salesforce story, they’re the heart of it.”
Meredith Brown, SVP of Trailhead & Community at Salesforce
