AI is quickly moving beyond chatbots and productivity assistants toward systems capable of reasoning, taking action, and working directly within business applications.
At Salesforce Dreamforce, NVIDIA founder and CEO Jensen Huang described artificial intelligence as an emerging infrastructure layer and predicted that enterprises will increasingly become both AI companies and “agentic” organizations. His appearance coincided with the announcement of Koa, Salesforce’s new CRM reasoning model built on NVIDIA Nemotron 3 Super.
For MSPs, Huang’s comments offer a glimpse at how enterprise AI could reshape customer technology environments—and the services required to support them.
1. The Agentic Enterprise Is Taking Shape
Huang predicted that enterprises will increasingly become agentic, with AI systems working across business processes rather than simply responding to prompts.
For MSPs, that could mean supporting customers where AI agents become part of everyday workflows across sales, service, operations, and other functions.
MSP Action: Start identifying customer workflows where agentic AI could create measurable efficiency or service improvements.
2. AI Is Becoming Part of Business Infrastructure
Huang compared AI’s emergence to previous infrastructure shifts involving electricity and the internet, describing AI as an infrastructure layer capable of expanding what organizations can know and do.
That suggests MSPs may increasingly need to treat AI as part of customers’ core technology strategies rather than as another standalone application.
MSP Action: Incorporate AI readiness into technology roadmaps and strategic conversations with customers.
3. Custom AI Could Become More Important
Salesforce created Koa by post-training NVIDIA Nemotron using synthetic enterprise CRM scenarios. The model runs within Salesforce’s infrastructure, and Salesforce says customer data was not used during training or inference.
The example illustrates how businesses and software providers can build AI around specific workflows and requirements.
MSP Action: Evaluate where specialized AI solutions may deliver greater customer value than general-purpose tools.
4. AI Safety Can’t Be an Afterthought
Huang called safety “job one” and argued that organizations can pursue innovation and speed while still building secure, reliable AI systems.
For MSPs, that reinforces the importance of security, governance, testing, permissions, and oversight as customers expand AI adoption.
MSP Action: Build AI security and governance requirements into customer adoption plans from the beginning.
5. MSPs Have an Opportunity to Guide AI Adoption
Huang encouraged organizations to engage with AI rather than risk being left behind. NVIDIA itself is using Salesforce across multiple business functions and piloting Agentforce for customer-support workflows.
As more organizations explore similar technologies, many will need help separating practical opportunities from experimentation.
MSP Action: Position AI conversations around business outcomes, identifying where automation and agents can solve specific customer problems.
The MSP Opportunity
The shift toward agentic AI could significantly expand the MSP’s role. Customers may need help evaluating AI platforms, preparing infrastructure and data, securing agent access, establishing governance, and integrating AI into existing workflows.
For MSPs, the opportunity isn’t simply to follow the latest AI technology. It’s to help customers determine where AI delivers real value, how to deploy it responsibly, and how to manage it as part of the modern technology environment.