Dell sees company knowledge, not AI models, as its competitive edge

Dell Technologies has spent close to two years running fully autonomous agents inside its own business, and it has come away with a blunt lesson: AI only matters if the company changes with it. John Roese, Dell’s global chief technology officer and chief AI officer, said firms that bolt agents onto old workflows will mostly end up doing the same work, just with machines mixed in.

That view puts the operating model, not the tool stack, at the center of Dell’s AI pitch. Roese made the case during an interview with theCUBE Research’s Dave Vellante at the Dell Technologies New York Symposium. The basic point was simple. If people, processes and responsibilities stay frozen, the AI project becomes a cosmetic layer instead of a real shift in how work gets done.

Dell’s own approach started with knowledge capture. Subject matter experts helped turn company practices into machine-readable skill files that guide how agents act. Roese described them as a Markdown file, but a very special one: text that encodes the rules, bias and parameters a company uses to distinguish itself in work like writing code, pricing a deal or servicing a customer.

And that’s the real prize, in Dell’s telling. The company says its agents get better as they work, forming relationships in knowledge graphs and developing new digital skills. That is also why Roese warns against handing proprietary knowledge to third parties. Frontier models and open models can still be used, he said, but they are not the moat. The moat is the special knowledge a company has captured and put to work inside its own system.