Post by Helen Bevan

Peer influence as a driver of change adoption: part 2. A summary of responses/comments to my previous post, reflecting on new Microsoft research on AI adoption (& applicable to other forms of change). 4 themes emerged from commenters: 1) Peer influence as the engine of adoption Mark Green says: tools, training & leadership enthusiasm are insufficient if "Dave from Finance" isn't quietly saying "this saved me 2 hours." Ian Kendall prefers pointing to someone "who walks in your shoes" over his own advocacy. John-Paul Crofton-Biwer adds that change sticks when people choose for themselves after being inspired by others. Lisa Evans (GACN), Sharon Mickan, Reza Hosseini Ghomi, MD, MSE & Thomas Kempin describe a similar pattern: informal relational networks & lived practice shift behaviour more than formal policy. 2) Networks & who we activate Marcella Bremer draws on Dr Leandro Herrero's concept of "viral change", finding visible go‑to colleagues & engaging them so new behaviours become "the way we do things around here." Patrick Whalen actively seeks first movers & early champions as deliberate allies. Jeppe Vilstrup Hansgaard points to Innovisor's work on mapping who might be activated for success, reinforcing that informal influence matters more than formal hierarchy. 3) Psychological safety, capacity & sequencing Andrew Jacobs (FCIPD) says that psychological safety must precede visible peer exchange, not emerge from it - as making learning visible in a fearful culture can increase anxiety rather than normalise practice. Anthony Lawton argues that many public sector programmes run "mandate‑first, social conditions somewhere after that" — a sequence that pushes learning underground. 4) Strategy, governance & context Dr Catriona Bradley warns that peer influence can either support or distract from strategy depending on how it is framed. Dr. Robert Marotta gives a frontline view: tools purchased & mandated without workflow fit or peer learning networks become "expensive shelfware." Miguel Guevara reframes peer influence as structural, not just social: trusted colleagues demonstrating AI in real work remove the path of least resistance back to the old way. Evelien Verschroeven describes how semi-structured conversations that create mutual vulnerability & complementary ownership make private experiments visible & collective learning possible. Lynsae Tulloch connects this to communities of practice, arguing that social learning backed by leadership advocacy can optimise tech adoption. Lillian Chiwera raises the balance between patient safety benefits & risks of cognitive decline, asking how resource redistribution can support culture change alongside technical adoption. The overall message (with a few dissenters): If we want AI (&/or other forms of change) adoption to be meaningful & sustainable, we have to invest at least as much in relationships, networks & social conditions as we do in technology, mandates & messaging. Thanks to all commenters.