Dell Converts Company Practices Into Instructions for AI Agents

Dell Technologies is converting its internal procedures into instructions for AI agents, putting control of company knowledge at the center of a strategy meant to work across models and platforms.

At a New York symposium Oct. 20, Dell Global Chief Technology Officer and Chief AI Officer John Roese described how subject matter experts turn their working methods into machine-readable skill files. Agents use those files for tasks involving coding, pricing and customer service.

The files give agents instructions drawn from Dell’s own practices. Roese argued that a company’s lasting advantage lies in the knowledge behind its processes rather than the tools that run them. Keeping control of that knowledge would allow a company to move a process to another agent platform.

Dell already uses agents in production to clean up customer relationship management data, write code and handle special pricing. Roese said in August that Dell had built its first agents about two years earlier and had put agents into production over the preceding year. The two time frames refer to when Dell began building agents and when it began using them for operational work.

Roese distinguishes agents that carry out tasks from chatbots that respond to requests. He also described changes to how companies organize people and processes as work shifts between employees and agents. Putting an agent into a workflow, in his account, involves more than choosing a model.

In Sept. 2025, Roese described about a dozen agent proofs of concept and roughly 20 digitized enterprise processes. Those were estimates of Dell’s work at that time. Dell has also identified process redesign, data preparation and governance as parts of its broader enterprise AI strategy.

Dell plans to add Knowledge Agents to its AI Data Platform in the first half of 2027. The planned customer feature would offer guidance on specific topics under rules customers set for data, quality and spending.