Elena Verna is one of the sharpest voices in B2B growth. She's Head of Growth at Lovable, spent years running growth at Miro, SurveyMonkey and Dropbox before that, and writes a widely-read newsletter on product and go-to-market strategy. Last week she published a piece aimed at software companies that every school communications professional should read sideways.
Her argument: your product has a new user, and it isn't human. As personal AI agents start acting on people's behalf — screening inboxes, summarising documents, deciding what's worth surfacing — the assumption that a human will ever actually see your work is quietly collapsing.
She was writing about SaaS. The logic travels.
If anyone is going to deploy a personal agent to manage the daily load, it's a working parent with three children at two schools and a job. The school newsletter is exactly the kind of thing agents are built to handle. And once that happens, most of what school marketing and communications teams currently measure — opens, clicks, read time — stops measuring what we think it measures.
This isn't a forecast about 2030. A parent with a Raspberry Pi and a free afternoon can already do this. The question isn't whether this reaches schools. It's what communications directors should do about it now.
The easy version of this argument is that open rates are becoming unreliable. When a parent's agent opens the newsletter, archives it, and surfaces only the RSVP link the parent actually needs, the dashboard still registers the open. Reach looks fine. The human never read a word.
That's true. But if we stop there, we end up buying sentiment analysis tools and calling it a day.
The deeper point is Elena's: when an agent sits between your output and your audience, everything you built for human attention becomes invisible. The tone of voice. The photo choice. The workshopped subject line. The design. The agent extracts what is happening, when, where, and whether action is required. Everything else is stripped out.
Which raises the question school comms teams have so far avoided:
If the craft of your communication is being stripped out before it reaches a human, what exactly is your function?
Elena frames this as a moat problem. What happens to your product when agents can reproduce the useful parts of it without ever touching your UI? For a lot of SaaS, the honest answer is: not much is left. The moat turns out to have been friction.
School communications has a version of the same problem. For years, the defensibility of a school's comms function has rested on things that only matter to human readers: the voice of the school, design polish, editorial judgement, the rhythm of the weekly update. None of this survives agent mediation intact. An agent doesn't register brand. It doesn't feel tone. It reads for facts and actions and discards the wrapper.
This is uncomfortable, because most school comms roles are substantially about the wrapper. And the wrapper is going to stop being seen.
Three things survive agent mediation, and they're where the function is heading.
Information that can't be replicated elsewhere. If the agent can get the same facts from the school website or the MIS parent portal, your message has no reason to be read at all. What's left is genuinely new information — decisions, deadlines, changes.
Connections that require a human on the other end. An agent can tell a parent the Year 4 trip needs volunteers. It can't volunteer. It can tell a parent another family is looking for a carpool. It can't be the carpool. The parts of school life that require a human to show up, in person, for another human — these don't abstract away.
The felt sense of belonging. The one thing a school community uniquely produces is the sense among parents that this is their school, that other families here are like them, that they are known. This isn't a message. It's a state. It's generated by interaction, not broadcast. And it's completely invisible in open-rate dashboards, because it was never measurable there in the first place.
Map these three against a traditional school comms function and a pattern emerges. The function isn't dying. It's being pulled upstream — away from packaging messages and toward producing the conditions that generate the things agents can't replicate.
If the packaging layer is being stripped out and the information layer is being abstracted into something agent-readable, the question is what Marketing and Comms Directors are actually for.
The answer, I think, is this: the role moves from broadcaster to community architect.
Everything an agent can do well — summarise, route, remind, surface, filter — reduces the school's burden of broadcast. That's useful. Comms teams should welcome it. But it also means the remaining value of the function concentrates in the things agents can't touch. The interest group that forms because two parents met at a coffee morning. The Reception-parents network that becomes the actual support system for a family going through a hard year. The PTA volunteer who recruits the other five because she knows them personally.
None of this is legible to an agent. All of it is legible to a child who feels their family is part of the school. And all of it depends on in-real-life meetups — the events, the mornings, the get-togethers that turn a parent list into a community.
This is also where the word-of-mouth flywheel lives. Parents don't recommend a school to other families because the newsletter opens cleanly. They recommend it because their child is happy, because they've made friends among the other parents, because the school feels like somewhere they belong. Reputation, brand, admissions pipeline — all of it sits downstream of whether parents actually know each other.
A communications function built around broadcast can't produce that. A community architect can.
The metric that matters in this world isn't open rate. It's whether parents, when asked, say they feel connected to the school and to each other. It's whether new families form real ties in their first term. It's whether parents on the edges — SEN families, out-of-catchment families, families whose first language isn't English — are inside the community or watching from the fringe.
These things can be measured, but not by the tools most schools currently use. They require a space where parents actually interact with each other — a safe, school-sanctioned space, not an ad-hoc one the school can't see or moderate. They require a way to run events, coffee mornings, interest groups, volunteer call-outs, meetups — all the small gatherings that produce real connection. And they require someone in the building whose job is to pay attention to the texture of that space.
That person is the community architect. They think less about subject lines and more about whether the new Reception cohort has actually met each other. They plan fewer newsletters and more in-person touchpoints. They measure success by the strength of the parent network, not the health of the mailing list.
Elena's phrase for this, applied to software, is worth stealing: some products will become invisible infrastructure, and some will become genuinely loved. School communications has to do both at once. The information layer has to become clean, structured, and agent-ready — invisible infrastructure that simply works. The community layer has to become something parents choose to participate in, because no agent will do that for them.
The parents aren't going anywhere. But increasingly, they're sending their agents to the mailbox. The question worth asking — before the shift fully lands — is what we want them to find when they get there, and what we're building for the humans behind them.