Seven emails before the news cycle wakes up - a spoofed journalist, an embargo to hold, half-year results to handle. Counsel screens out the fakes and puts the real media requests in front of the right person, scored against what's live.
The fraud verdicts on this page are the actual output of the CounselAgent screening engine run on this sample inbox - not a mock-up. Re-screen with node rescreen-demo.js pr.
Each client's current situation is on file - the results due, the embargo running, the story circulating. That context is what makes triage sharp: the same media request matters more or less depending on what's live right now.
Seven messages. To a person they all look plausible - a journalist after a steer, a results query, a renewal invoice, a webinar. Counsel screens every one at the door before anybody opens them.
Every inbound is screened - deterministic rules plus an AI verdict - before anyone replies. A spoofed journalist fishing for the results early, and a "changed bank details" renewal, both stopped. A guard that runs quietly so the team can trust their inbox.
This is where the day is won. Only the genuine mail reaches triage. Each request is scored against its client's live situation, given a recommended posture - on the record, hold the line, escalate - and routed to the right person. An embargo mid-results-week ranks differently to a cold request on a quiet account.
The high-impact items become private signals woven into each client's morning brief - and the whole lot is written back into the knowledge vault, so tomorrow the brain is current. The loop closes.