Shared ownership: dev teams own their service's SLOs + on-call (SRE enables, not owns)

Make reliability visible + rewarded: SLO dashboards, postmortem learnings shared widely

Provide paved roads/platform so the reliable path is the easy path

Leadership buy-in: error budget policy signed by product; blameless culture from the top

Teach via game days, postmortem reviews, embedded SREs

A reliability culture can't live in a single SRE silo — if SRE is the only team that cares, every other team optimizes purely for features and SRE drowns. The lead's job is to spread ownership: dev teams own their SLOs and on-call, with SRE providing the platform, tooling, and expertise (paved roads make the reliable path the path of least resistance). It needs genuine leadership buy-in — product signing the error budget policy, blamelessness modeled from the top — and continuous teaching through shared postmortems, game days, and embedded SREs. Culture is the hard 80%; tooling is the easy 20%.

SRE enables; teams own: The end state isn't "SRE keeps everyone reliable" — it's every team owning its reliability with SRE providing leverage. That scales; a central team being everyone's safety net does not.