Most public agencies now operate under a standing directive to modernize. Boards tell communications staff to adopt AI, streamline workflows, and do more with a budget that isn’t growing. The instruction usually stops there. What lands on a comms director’s desk is “use more automation,” not a plan for what that means inside an agency that also has to think about public records requests, procurement rules, and a public that didn’t ask to be marketed to.
The response is often to shop for a platform. A district evaluates engagement software. A transit agency compares CRMs. The evaluation itself becomes the project, and it produces a decision about software rather than a decision about what the communications function should be doing differently. Somewhere in that sequence, the tool quietly became the strategy.
Tooling is a real part of the problem, worth saying plainly. Most agencies are running some combination of a decade-old website, a handful of platforms that overlap in ways nobody has audited, and at least one system nobody remembers why they bought. Knowing what’s useful, what’s redundant, and what the agency can run without outside help is a legitimate skill, and a necessary one. On its own, though, it isn’t leadership. A tool audit tells you what you own. Whether the communications function is doing what it needs to, who it’s reaching, and what changes if it starts working better are separate questions, ones a spreadsheet of licenses can’t answer.
What the tools do, as of September 2026
The platforms agencies evaluate fall into a few categories, and the category matters more than the brand, because each one answers a different question about who the agency is talking to and why.
The first is subscription communication: the email and text updates residents sign up for. Granicus govDelivery is the long-standing platform here; it has held FedRAMP Moderate authorization since 2016, with authorizations from 46 federal agencies. In July 2025 Granicus added GXA, an AI agent that answers resident questions around the clock, with text, voice and WhatsApp versions in preview as of January 2026. For cities and counties, CivicPlus covers much of the same ground through its website platform, SeeClickFix for service requests and Notify Me for alerts. Five CivicPlus products were rated GovRAMP Ready in August 2025 (SeeClickFix and Notify Me were not among them), and in December 2025 the company bought Streamline, which builds websites for special districts.
The second is public engagement: consultation and comment analysis. Granicus EngagementHQ (formerly Bang the Table), Open Point (the name Social Pinpoint took in March 2026), Zencity and Polco sit here. Much of the recent AI work has landed here, mostly as summarization that sorts thousands of public comments into themes; EngagementHQ’s comment theming and sentiment analysis were in preview as of January 2026. That is useful and deserves more scrutiny than it gets. A theme model decides which concerns reach a decision-maker in readable form, and in Washington the summary is as much a public record as the comments it summarizes.
The third is alerting. Everbridge and Motorola’s Rave Alert run emergency notification, and this is the category where a targeting error is most visible. In January 2025 Los Angeles County reported that a wildfire evacuation alert “was erroneously sent out to nearly 10 million residents across the County” (county statement).
The fourth is general-purpose marketing automation, the same software private companies run. Salesforce extended FedRAMP High authorization to Marketing Cloud in June 2025. Microsoft is a different case: its own documentation says Dynamics 365 Customer Insights – Journeys “isn’t yet available for US GCC, DoD, GCC High or other sovereign clouds” (Microsoft Learn), which is worth confirming before assuming an existing Microsoft agreement covers it. At the small end, Mailchimp’s free plan now caps out at 250 contacts, which matters for a small district that started there.
The rules that shape the choice
Public agencies carry constraints that private marketing teams rarely think about, and those constraints tend to settle the platform question before any demo does. Washington’s Public Records Act defines a “writing” to include “every other means of recording any form of communication,” so an automated email sequence and a chatbot transcript are both records subject to retention and disclosure. The Department of Justice’s accessibility rule under ADA Title II holds agency websites and apps to WCAG 2.1 AA; an interim rule in April 2026 moved the compliance dates to April 26, 2027 for governments serving 50,000 people or more and to April 26, 2028 for smaller governments and special districts. Text messaging has its own exposure: the FCC’s 2016 ruling exempted the federal government from the Telephone Consumer Protection Act, but not states or cities, so consent still governs an agency’s SMS program.
Granicus’s 2026 State of Digital Government report, a survey of more than 550 government communicators, puts the gap in numbers: “71% of respondents report using artificial intelligence in some capacity, and 44% have formal AI policies in place.” The distance between those two figures is where most of the risk sits, and no platform closes it on its own.
Where judgment matters more than the platform
I sit on the board of The Urbanist, a Seattle publication that covers how the region plans and governs itself, so I spend a good deal of time on the receiving end of agency communication. Before Cedar Collaborative I ran marketing at Funnelback, an enterprise search company whose client list was heavy with government. Neither is the same as running an agency’s communications from the inside. Both make it hard to miss how rarely the underlying systems line up. When Seattle’s finance director, Dwight Dively, spoke with The Urbanist in September 2026, he put it plainly: “Some of the very basic things that any organization, whether it’s a government, a nonprofit, a business, needs to do well, we don’t do well anymore.” Mayor Katie Wilson’s Executive Order 2026-08 now asks departments to modernize technology, grants tracking and customer service together, with recommendations due March 30, 2027. Communications sits inside that same picture.
The instinct I would push back on most is treating a data dump as transparency. Publishing everything feels accountable, but raw data invites cherry-picking, in the same way a narrowly drawn records request can produce criticism or celebration that has little to do with the whole record. The stronger practice is to communicate events, activities and data as one coherent account of how the government serves people, with the agency’s own work legible inside it. An agency that has made that decision can evaluate any of the platforms above in an afternoon. Without it, the evaluation becomes the project.
How this gets hired
Washington doesn’t require most local governments to run a full competitive process for this kind of work. Cities, counties, school districts, and special purpose districts other than ports and public facilities districts set their own rules for what the state calls a personal services contract, a category built for bringing in outside expertise to solve a defined problem or complete a temporary project, not to run an ongoing department. Most agencies also keep a consultant roster with a dollar threshold below which they can skip competitive bidding entirely. For a district or a mid-size agency trying to close a specific gap in how it communicates, that’s a considerably lighter process than “government contracting” tends to sound.
Larger agencies are a different case. An organization the size of Sound Transit runs competitive procurement as standard practice, with board approval required on anything substantial. It also runs a Technical Assistance Program built specifically to help smaller firms understand and work through that process, which functions as a real door in rather than a formality. Getting hired at that scale looks less like a single pitch and more like building a relationship with procurement staff well before there’s a specific opening to fill.
Leadership, not a longer contract
Agencies sometimes treat this kind of engagement as temporary staffing, someone to hold the seat until there’s budget or justification for a full-time hire. That framing undersells what a senior person can do inside a fixed window. What they contribute is judgment applied to a specific mandate: what the agency’s existing tools are for, what a three-person comms team can realistically own, and what changes once someone senior is accountable for the answer. Structured that way, the engagement tends to leave behind something that outlasts the contract, not just a report nobody reopens.
In practice, that means holding two things at once. One is a working knowledge of the marketplace: what a given automation platform genuinely does well, where it duplicates something the agency’s IT department already licenses, and where the honest answer is that the agency doesn’t need new software at all, it needs someone using what it already has correctly. The other is an understanding of what changes when the client is public money: claims that would be routine in a private-sector case study become public record, and vendor relationships carry conflict-of-interest rules private companies rarely think about. Working inside those constraints is a different skill from working around them, and it’s usually the one an outside vendor hasn’t had to develop.
None of this makes the tooling question irrelevant. It just changes where it sits. An agency that knows what its communications function is supposed to accomplish can evaluate a platform in an afternoon. Skip that step, and the platform decision gets made again in eighteen months, by whoever’s in the seat next, with the same uncertainty about what problem it was ever supposed to solve.