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 question, five platforms: stance, your message, and the sources behind every answer.
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.
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.
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
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 Framing shifts, new voices in the answers, outdated claims still circulating.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.
How it works
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.
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.
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
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.
The playbook
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.
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
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.
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.
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.
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.
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.
A thirty-minute walkthrough with your questions, not ours.
Request a Demo Request Pricing