Here is what we built, and how it works.
Cyber Insurance Academy trains brokers, underwriters, and claims professionals, and runs the largest global community in the field. The growth question is the same one every focused team has. How do you reach the right carrier, the right broker, the right person, at the moment they are actually ready to hear from you, without burning your name on spray.
This page walks through the system built to answer that. The companion to it is your live dashboard, where the qualified plays show up, each one traced to a real source. Owned by you. Built around your ICP, not a template.
Everything here is built from what we know and researched so far.
We stood this up from your ICP and public signal, on our own research. It is real, and it works today. With you it gets sharper. Your read on each account, the context only your team has, the edge cases, all of that tightens every play. We build this together.
Your world, mirrored back.
The system starts from what you already know about who buys. Your authority is real, and it is the wedge: a gold standard certification and a community the market trusts.
The targeting spec Guy wrote
Pulled straight from the dream client list you sent. The system carries this as its brief and verifies every account against it.
Dream clients
- Global carriers with in house cyber teams (CNA Hardy, Travelers, Sompo, The Hartford)
- Top specialty brokers (Lockton, Alliant, AON)
Indirect reach
- Insurance networks and associations, US and UK
- Attendees of cyber risk conferences (Cyber Risk London, NetDiligence Toronto)
The people
- Head of Cyber Underwriting, Cyber Product Manager
- L&D and Training Managers in insurance
- Chief Underwriting Officers, broker cyber practice leads
Four agents. One job each.
Not one model trying to do everything. A small team, each one good at a single step, passing work down a line. The same pattern runs in production for our other clients today.
Hunts the web and Apollo for carriers and brokers showing a real, dated cyber signal: a new Head of Cyber, a cyber product launch, an underwriter job posting, a conference sponsorship. It works two tracks, your named dream accounts and look-alikes it finds from the signal itself. It only writes down a lead if there is a real source page behind it.
Takes Ranger's finds and goes deep. Re-opens every source link, confirms the quote is really on the page, checks the company and the date, then finds the specific person to approach, one of your five personas, and their LinkedIn. If it cannot verify the signal or find a real person, it kills the lead.
Takes the verified lead and writes the outreach. Picks the one CIA program that fits the situation, an enterprise program, CCIS, or the right course, and drafts the message so it reads like Guy or Asaf sent it themselves. Value before the ask, nothing template-y. The more you use it, the more it sounds like you.
Reviews everything adversarially before it reaches you. Re-checks the source, hunts for any fabricated fact, confirms the person is a real ICP buyer, and that a CIA founder would actually send the message. Three verdicts: approved, sent back for revision, or dropped. Nothing reaches your dashboard until it clears this gate.
What holds it together
Each agent has its own memory in plain files: your ICP, your programs, what it has learned. They share state through a database with four tables (accounts, triggers, people, plays), and a watcher script moves work between them automatically. The finished plays land in a small Base Camp style dashboard, your dashboard, not another platform your team has to learn.
The agents actually think.
This is not fill-in-the-blanks. At the start, Guy and Asaf brief the system deeply: your ICP, your programs, how you sell, who you are. From there the agents reason. The deep research is where it shows. They go after the non-obvious hook a sharp human would find with hours to spare on a single account.
Research that goes deep instead of wide, paired with agents that genuinely reason about each account. The more you tell them up front, the better they think. That is the real edge.
How one play gets built.
Every play moves through the same line. Nothing reaches your dashboard until it has cleared the last gate.
Ranger
Finds the account and a fresh, dated signal worth acting on.
→Analyst
Researches the account and the person, sources every fact.
→Penman
Drafts the message and picks the right program to lead with.
→Gatekeeper
Approves, sends back for a fix, or drops it. Then it reaches you.
The gate that protects your name
Gatekeeper gives one of three verdicts: approved, needs revision, or dropped. A draft that needs work goes back to Penman for up to three rounds, then stops. You only ever see plays that survived an adversarial check. That is the difference between intelligence and noise.
What lands in your dashboard.
Every approved play arrives as five fields. Enough to act in the hour you have, nothing to dig for.
"Saw your team is adding cyber underwriters. The hard part is rarely the hire, it is getting them fluent in cyber risk fast enough to underwrite with confidence. We certify cyber insurance specialists for a living, and run whitelabel programs for carriers scaling exactly this way. Worth a short conversation before your next intake?"
Depth over volume.
The edge is not how many accounts the system touches in a day. It is that every account is real and every signal can be checked.
The targeting spec is the one Guy wrote, not a generic list. The system hunts your dream clients and the personas you named.
Nothing sends itself. Plays land in your dashboard for your team to approve and send. The system does the research, you keep the judgment.
If a claim cannot be traced to a real, live source, it gets cut. No invented companies, no made up triggers. Gatekeeper enforces it.
Intelligence shows up where your team already works, a clean dashboard, not one more tool to log into and forget.
The system gets sharper every week.
Every play your team approves, edits, or passes on teaches it. The agents keep a memory of what landed and what missed, so the next batch is tuned to what Guy and Asaf actually respond to. The longer it runs, the better it gets. That feedback loop is the whole point.
Built for Cyber Insurance Academy.
Around your ICP. Owned by you.
The dashboard link comes next, with the first plays already loaded. Open it, read a few, and tell me what feels right and what to sharpen. We tune it from there.