The Clean Break

Most campaigns start from a stock: every company that matches, whenever it was created. This one starts from a flow. The register publishes every new French registration, and a weekly pull filtered on the architecture and real-estate activity codes hands over everyone who incorporated in the last few days — a list nobody can buy, because a week ago it did not exist.

Freshness is the entire edge. An architect is required to carry décennale insurance before signing a first project; a real-estate agent cannot get a carte professionnelle without RC pro and a garantie financière. The need is not created by the campaign, it is created by the law, and it peaks in the first weeks. Call in month six and the contract is signed with whoever called in month one.

What survives the shell filter is still just paperwork: a legal form, an address, a declared activity. The question that matters is whether there is a person behind it building their actual livelihood — as opposed to a side structure, a holding for an existing practice, or a dormant registration made to reserve a name.

The tells are ordinary and readable: a sole director who is a physical person, a registered address that is a home or a small office rather than an accountant's letterbox, an activity declaration that describes doing the work rather than holding the assets. None of them is proof alone; together they sort the real starts from the administrative noise.

The register already gives the founder's legal name — there is no guessing who runs a two-week-old company. What it does not give is the ten years before it, and that history is the whole campaign. So each director gets matched to a LinkedIn profile on name, region and sector, with an AI judge scoring whether this profile genuinely is that person and not a namesake.

The bar for a match is deliberately high. A wrong profile does not just waste a call — it produces a pitch built on someone else's career, delivered to a person who notices immediately. Ambiguous matches are dropped, not coin-flipped; the register will mint replacements next Monday.

This is the step the recipe is named after. The profile's history answers one question: was this person, until very recently, an employee — an architect salaried at a firm, an agent salaried at an agency? If yes, the registration is not just a new company. It is the week their mutuelle, their prévoyance and their employer's retirement contributions all stopped at once.

That reading splits the list in two. The career-switcher has a personal protection gap opening at the exact moment of first contact — Profile A. The founder whose history is already a string of independent structures has no such gap, but has a brand-new entity with mandatory cover to arrange — Profile B. Same register line, opposite conversation.

Everything upstream was qualification; this is where the message gets decided. Profile A gets the protection conversation — what disappeared when the payslip did, and what a first year of independence looks like when it goes wrong. Profile B gets the structure conversation — décennale, RC pro, garantie financière, the premises, arranged once and correctly.

The split is what keeps a phone call from sounding like a script. Open with décennale to someone terrified about losing their salary coverage and the call is over in twenty seconds. Open with prévoyance to a serial founder on their third agency and it is over faster. The segment is the pitch.

An explicit rejection gate before any enrichment money is spent on a phone number. Out go the side hustles run next to a job that continues, the structures created for a single transaction, the registrations by people already retired, and anything where the confirmed story contradicts the register's.

As in every recipe, the discards are kept and read. If the same rejection reason keeps firing — a legal form that always turns out to be a shell, a code that always turns out to be holdings — that is a fix to the Monday pull, not a hundred more individual rejections.

A founder in their first month is a terrible email target and an excellent phone target. The inbox is a firehose of URSSAF, banks, insurers and software trials, and a cold email drowns in it. The phone is the opposite: no assistant screens it yet, and a call about mandatory insurance is a call about this week's actual to-do list.

So the last enrichment step is Kaspr, and the standard is strict: a direct mobile or nothing. A switchboard number for a one-person firm does not exist, and a generic line is not worth a dial. Contacts where Kaspr comes back empty fall back to email with the same profile-matched pitch — a weaker channel, but the research already paid for itself.