Takeaways from Sol Rashidi’s session at Broadcom’s 2026 Automation Virtual Summit
In her session at the 2026 Automation Virtual Summit, Sol Rashidi described a world where applications, data, models and agents converge into one connected mesh, with agents and automation as the connective tissue holding it together. Her advice for doing that well: know the problem you’re solving, prepare your people and your data, and choose carefully who you build with.
The session, “The Human Advantage: Trusted Orchestration for AI Amplification, Not Replacement,” was hosted by Automation by Broadcom and Google Cloud. Sol is known as the first Chief AI Officer and wrote the best-seller Your AI Survival Guide.
Most organizations are still finding their footing
She put the share of AI initiatives that stop, pause or get cancelled at proof of concept somewhere between 74 and 88 percent, and called it perpetual POC purgatory. It shows up most in enterprises with uneven data foundations and a lot of technical debt. Her point: if it feels like everyone else has this figured out, they don’t.
Start with the business problem
Her first lesson was to solve a business problem rather than do AI for its own sake. Using AI tools to speed up individual tasks has its place, but she sees the real return in redesigning the work: taking a function apart, looking at its tasks and how it runs today, improving the process, and then deciding where automation, agents or AI fit.
Build the foundation
She argued that project plans lean too heavily on models and infrastructure and too lightly on preparing the workforce, and that AI fluency means more than a short course on a tool. It means helping people see how their own roles evolve. On the data side, her advice was to classify it early, so you know what’s sensitive, where it flows and who has access, and to remember that governance looks different for applied, generative and agentic AI.
She also described what happens when teams rush: dozens of orchestration tools, scripts, monitoring systems and agents running independently, with inconsistent governance and security. The result is more risk, and more time spent triaging failures and working out who owns what.
Our view: this is the problem orchestration exists to solve. When one platform runs the work, you can see what ran, who owns it and what failed.
Choose a partner who shares the concern
Her advice on partners was direct: pick ones with governance and security built into how they work, because when something goes wrong you want people who feel the same weight of concern you do. She framed it as a choice about reliability, stability and governance more than features.
We’d also look for people who have run these environments themselves, who tell you plainly what they find, and who stick around after go-live.
Human led, AI supercharged
Sol closed with a line that’s close to our own tagline: be “human led and AI supercharged.” Let AI, automation and agents carry the execution, and keep the thinking, the judgment and the relationships with people.
Ours is Human-Led. AI-Enabled. People make the decisions and own the outcome. Automation handles the work around them.
AutomWorx specializes in Automic workload automation, orchestration, and AI-enabled infrastructure modernization consulting.