Fletcher Previn, senior vice president and chief information officer at Cisco Systems since 2021, is rethinking how work gets done as generative Artificial Intelligence tools become more widely used across the company. He frames his approach as collaborative, balancing bottom-up demand from employees for new tools with top-down decisions on how roles should evolve. Previn oversees an IT organization that supports about 10,000 workers and is considering changes that would affect Cisco Systems’ wider workforce of roughly 90,400 employees.
Previn is pushing developers to adopt Artificial Intelligence coding tools such as Cursor, Windsurf, and GitHub Copilot and is tracking both usage and acceptance of generated code. About 70% of Cisco Systems’ roughly 20,000 developers log into Artificial Intelligence coding tools at least once per month, and the acceptance rate for AI-generated code has risen to about 24% from 4% nearly a year ago. Previn says Artificial Intelligence is improving and supporting more languages, and he hopes that over time up to 70% of the company’s code could be AI generated. He is also reassessing job responsibilities, titles, and training needs to account for new toolsets.
Beyond development, Cisco Systems is using Artificial Intelligence to boost productivity for nontechnical staff, including more intelligent onboarding that identifies the precise tools a new hire needs and AI-driven hardware planning. By analyzing device memory, application performance, and network telemetry, the company aims to distinguish transient performance issues from genuine hardware failures, potentially avoiding unnecessary laptop replacements. Previn emphasizes small, cross-functional IT teams—six to ten people focused on systems like Workday, SAP, or Oracle—to reduce the productivity loss of constantly forming project teams.
Previn is cautious about adopting costly AI add-ons from vendors, warning they can raise costs and confuse employee experience. He expects an agent-to-agent future in which digital assistants can act autonomously and coordinate tasks on behalf of workers. Cisco Systems has built an internal digital AI teammate that infers intent and selects appropriate models and data sources so employees do not need to pick between language models or data classifications. Previn warns that if developers use outdated models, the software they write will be correspondingly out of date.
