AI answer intelligence for public affairs

Know what AI tells the public about your issues.

Every day, people put your policy debates to ChatGPT, Claude, Google and Perplexity, and take the answers as fact. Breakwater watches those answers, scores them against your positions, and tells you when the story shifts, who is shifting it, and what to do about it.

One policy question scored across ChatGPT, Claude, Perplexity, Grok and Google: stance, message presence, and sources for each platform

One question, five platforms: stance, your message, and the sources behind every answer.

Asked of AI, answered by AI

Track the questions people actually put to AI assistants, each backed by search-demand evidence, asked in several phrasings, and sampled week after week so you see the distribution, not a coin flip.

Your message, from any source

What counts is your arguments appearing, whoever carries them. Every answer is judged against positions you write and own, on a spectrum from favorable to hostile.

Know what to do next

Every question gets a strategy grounded in who actually supplies its answers: which reporter to brief, what authoritative piece is missing, where accurate voices are thin, each play with a stated purpose.

The week, read for you

Open the dashboard, know where you stand.

Every collection ends with a plain-language snapshot: the single most important truth right now, then the items that need eyes, each one linked to its evidence. Framing shifts, new voices in the answers, outdated claims still circulating. No dashboard archaeology; the reading is done for you.

The account snapshot: a plain-language headline on message presence plus tone-coded items linking to evidence

How it works

Three steps to your first answers.

1

Pick what to monitor

Type an issue or a bill. Breakwater proposes the questions real people ask AI about it, with the search volume to prove each one earns its place.

2

State your positions

Write where you stand, in your words. We map what the opposition argues and surface the key events found in the answers. Every answer is scored against positions you own.

3

Read the week, act on it

A digest that starts with what matters: framing shifts with evidence, new sources entering the answers, demand rising, and an action plan grounded in who supplies the answers today.

The sources

See exactly who is shaping the answer.

Every link an AI answer cites is classified supportive, neutral or adverse, answer by answer, not by outlet reputation. Roll them up across 30 days and you know which outlets, journalists, and pages actually supply the framing, which articles do the damage, and where supportive ground is thin. Action plans start from that list, not from guesswork.

Every cited source classified supportive, neutral or adverse
An action plan drafted from live answer data: media outreach, authoritative content, and official-record plays, each measured before and after

The playbook

From who's shaping it, to what you do about it.

Every question gets a plan written from its own answer data: the outlet to brief, the explainer to publish, the record to file, each play with a stated purpose. Mark a play enacted and Breakwater measures the answers before and after, so you know what moved. Grow supportive voices, never suppress, never astroturf: comms work you'd be glad to explain.

Citations matter. The message landing matters more.

AI assistants compose answers from sources they trust. Breakwater measures what actually matters: whether your side of the story makes it into the answer, from your site or anyone else's, and shows you exactly where to grow the supportive ground when it doesn't.

Questions, answered

The things people ask us first.

Which AI platforms do you monitor?

ChatGPT, Claude, Perplexity, Grok, and Google's AI answers on search. Each tracked question is asked in several phrasings and sampled on a schedule, daily for high-priority questions, so you see the distribution of answers rather than a single roll of the dice.

What does "scored against your positions" mean?

You write where your organization stands on each topic, in your own words. Every collected answer is judged against that position on a spectrum from favorable to hostile, and we track whether your arguments appear in the answer, whoever is cited. An adverse answer that still repeats your argument counts; that nuance is the point.

Do you manipulate the AI platforms?

No. Breakwater reads publicly available answers and helps you do legitimate communications work: brief the journalists AI already trusts, publish the authoritative explainer that is missing, put your evidence on the official record. Grow supportive voices, never suppress, never astroturf.

How much work is it for my team?

Setup is a working session: pick your topics, approve the questions we propose from real search demand, and make the drafted positions yours. After that the rhythm is weekly: a digest of what changed, alerts only when framing genuinely shifts, and a plan when a question needs one.

What does it cost?

Pricing follows the size of your monitoring: how many topics and questions you track. No seat math, no feature gates, no tier sheet. Tell us what you'd watch and we come back with a quote shaped to it.

See what AI is telling the public about your issues.

A thirty-minute walkthrough with your questions, not ours.

Request a Demo Request Pricing