The AI era is here. Local newsrooms aren't in it yet.

The problem

AI is moving fast and it doesn't know where you live

Every major AI model — Claude, ChatGPT, Gemini — is being embedded into search, into the tools people use daily to understand the world around them. That's not slowing down. It's accelerating.

And every single one of those models has the same critical blind spot: no reliable, verified, current knowledge of what's actually happening in your community. They know what your city is. They don't know who's on your planning commission, what was decided at last Tuesday's water board meeting, or what's driving the housing crisis on your specific street.

When they answer local questions, they're guessing from general patterns. The result is plausible-sounding responses that could apply to any comparable community anywhere. The specificity, the nuance, the institutional memory that local journalism exists to create — none of it makes it into the answer.

Local knowledge gets flattened. Every community starts to sound the same. That's not a hypothetical. It's already happening — and the window to change it is closing as platforms decide which sources to trust and which knowledge bases to retrieve from.

The opportunity

Local newsrooms are sitting on the most valuable untapped data asset in the AI era

Every story a local newsroom has published, every meeting it has covered, every document it has obtained — that is a verified, community-specific knowledge base that no AI company can replicate at scale. Six years of coverage of a local water rights dispute. A decade of school board meetings. A reporter's notes from three years on the housing beat.

The problem is that this asset is invisible because it's unstructured. It lives in a CMS, in reporters' notes, in PDFs, in transcripts. It exists but it can't be queried. It can't be handed to an AI at the moment it needs to answer a local question. And it walks out the door whenever a reporter leaves.

Civant turns that pile of knowledge into something permanent, searchable, and machine-readable — without requiring a newsroom to do anything differently.

How it works

The newsroom keeps reporting. Civant captures everything.

Every published story, government document, meeting transcript, and reporter note that flows into Civant gets converted into a form an AI can search and retrieve from. Not by training a new model — that's expensive, slow, and becomes stale the moment new articles are published.

Instead, Civant maintains a live, continuously updated local knowledge base. When a reporter asks a research question or an AI platform queries for local context, Civant finds the most relevant passages from across years of coverage and hands them to the AI along with a simple instruction: answer this using these sources, cite every claim, and don't make anything up.

Source protection

Built for newsrooms that take their sources seriously

Civant has four record statuses: on record, background, deep background, and off record. These aren't labels. They're structural controls enforced at the database level.

Off-record material never enters the corpus, never reaches an AI, and cannot be overridden by a UI bug or a reporter mistake. The exclusion happens server-side before any prompt is assembled. A newsroom can adopt Civant knowing that the tool cannot accidentally expose protected sources — because the system architecture makes it technically impossible, not just policy.

The community intelligence layer

Local newsrooms as the source of record in the AI era

Beyond the newsroom, Civant exposes a verified local knowledge endpoint that AI platforms can query directly. When someone asks Claude or ChatGPT about a zoning dispute, a water rights fight, or a school board decision — and a Civant-powered newsroom has covered it — the answer should come from that reporting, not from a model's best guess.

That's what the community intelligence layer means. The newsroom's work doesn't just inform the community through published stories. It becomes the authoritative local data source that shapes how AI answers questions about that community across every platform.

The patterns defining how AI interacts with local information are being set right now. Newsrooms that have a structured, machine-readable intelligence layer will be part of that ecosystem. Those that don't will be bypassed — not maliciously, but by default, because there's nothing to retrieve.

Where Civant comes from

Built by a local newsroom, for local newsrooms