Find brands, agencies and talent before they sign up, research them, build their account and invite them to claim it
Most people who would be good on Zooly have never heard of it. The prospect pipeline is how the team finds them, learns enough to be useful, builds their account ahead of time, and invites them to take it over.
A prospect is one of those people before they claim anything. The pipeline keeps one row per prospect and moves it from "we just found this name" to "they own their account now". It is a team tool. Prospects never see this list.
The team looks for four kinds of people:
Brands and agencies are found because they might spend money. Talent is found because a brand will want them before they ever sign up. A talent agency can bring its whole roster in one go, and each person on it becomes a prospect too.
On the Prospects page, the team pastes what it has. One prospect per line, and a line can be any of these:
@handle or a link to a social profileA whole list can go in at once, one per line, or as a table copied from a spreadsheet with columns for name, website, handle, kind, email, notes and owner. Rows with nothing usable are skipped and shown with the reason.
The team picks the kind once for the paste. A spreadsheet row with its own kind column keeps that kind. Once the account is built, the kind cannot change. An owner and a note can be added now or later; the owner is the person adding unless they type another name. An email is not required to start looking someone up.
For an agency, a Roster research run reads the agency's public roster page and adds what it finds as prospects of their own, marked with where they came from: talent for a talent agency, brands for a brand agency. One run adds at most 200. It only finds the list when the roster sits on a page the run can reach, which is not always the case; see what it does not do.
| Stage | What it means | What moves it on |
|---|---|---|
| New | On the list, nothing looked up yet | A research run that finds something |
| Researched | Research found who they are and enough to be useful | Building the account |
| Account built | An account exists, filled from research | Preparing the invite |
| Invite ready | The claim link and both drafts are ready | Sending the email, or marking it sent by hand |
| Invited | Someone sent the email or the LinkedIn note | The person claiming the account |
| Claimed | The account has an owner | Nothing, it is the end of the line |
| Declined | The team decided not to approach them | Reopening it puts it back at new |
| Bounced | The team marked that the email came back | Fixing the address and sending again, or marking it sent by hand |
Claimed is never set by hand. It follows the account actually getting an owner. An agency that claims its account marks the talent on its roster as claimed by agency.
Each row shows the name and the one line research wrote, the kind, the stage, the owner, the best contact found (an email, a contact form or LinkedIn), what research has cost for that person, and where they came from.
Above the list are the totals: how many prospects there are, what research has spent in total, what it has spent on the rows in view, and whether email sending is on or off. The list can be filtered by kind and by stage, and the stage filter shows how many sit in each one.
Selecting rows opens the two bulk actions. Research runs the chosen depth on every selected prospect. Build account and invite creates the private account for each one and prepares the invite. Agencies are built before their roster, so the roster joins the right account.
Research reads public sources only. Depending on the depth chosen, it collects:
The four depths are Light (who they are and what they are doing now), Deep (adds social activity and past deals), Contact (finds a way to reach them), and Roster (reads an agency's list). Each run is saved onto the prospect, and a later run adds to what is already there rather than starting over. Claimed and declined prospects are skipped.
Research costs money every time it runs, and the page is built to make that visible before anyone presses the button.
Each single run has a small cap, five cents by default. Each press of Research has its own cap for the whole press, two dollars by default, and that number can be changed on the page. When the press cap is reached, the remaining prospects are skipped rather than charged, and the result says so. Every run records its cost on the prospect, and the list totals it.
A pre-built account is a real account the team made before the person arrived. Nobody owns it yet. It is not an empty shell. It is filled with what research found:
The point is that the invite lands on something real. When the person opens the claim link, Z can already say "here is what I know about you" instead of starting from a blank page. The account has no owner until someone claims it.
Unclaimed accounts can take deals. Their public page stays hidden. An account Zooly built from research can be matched, receive offers, be part of a blast, host a deal room and be seeded, the same as any account. What stays hidden is the public page, because research may be wrong and nobody has confirmed it yet: the public talent page shows "not found", and a shared link shows the plain Zooly card, without their name or photo.
The hidden page covers talent the pipeline built, while unclaimed and without an owner, and only while no agency that represents them has joined Zooly:
Nothing is sent to an unclaimed account on its own. When an offer, blast, seed or deal reaches an account nobody has claimed, no email or SMS goes out. It waits on the Outreach page under Offers to unclaimed accounts, marked "Offer to unclaimed account, waiting for admin approval". An admin sees who it goes to, the address research found and where it came from, and the offer in a line or two, then presses Approve or Skip. Approve sends one email to that address (or an SMS only if the person agreed to SMS on that number), with the offer and a link: "Claim your account to reply". Who approved and when is kept, and the same offer is never sent twice. Skip sends nothing. Updates, reminders and the weekly note only ever go to real people who own or manage the account.
Preparing the invite creates a claim link and two drafts: an email and a short LinkedIn note. Both are written from what research found, so they can name the person's work instead of sending a generic message.
The email is never sent on its own. A person on the team reads it, can change the address, subject and body, and then approves it. Email sending is off for the product unless it is switched on, and the page header says which. While it is off, approving saves the email, and the team copies it, sends it themselves, and presses Mark sent. The claim link always has to stay in the email. If two people approve the same invite at once, it is sent once.
LinkedIn is always by hand. The note is kept under 300 characters, with the claim link at the end. It is there to copy, and the page opens the person's profile when research found one. Nobody can send it from Zooly, and no automation touches it. When it has been sent, the team marks it sent on LinkedIn.
A claim link lasts 30 days. New link makes another one and rewrites both drafts. The old link stops working if nobody has used it yet, which is useful when an address was wrong or a link was forwarded.
A talent whose page is hidden can still be offered a deal through the agency that represents them, when that agency has not claimed yet. The offer text is written for the agency and shows on the talent's prospect, marked agency offer to send.
When the agency later claims its account, these offers are linked to it.
The claim link opens a welcome page that greets the person by name and says their agent is ready. One button, Meet your agent, opens a chat with Z.
Z's first message is the card: what it already knows. It shows their name and kind, their social handles, past brands and work, past deals, anything they told us they are looking for, and, for an agency, the roster it has already set up. Under it are two choices: That's right, or Let me fix something.
Fixing happens by typing to Z, not by editing the card. Whatever they correct is saved onto the account and the card comes back updated. Then Z asks only the questions still missing, one at a time, with quick replies: a phone number, the kinds of deals they want, their smallest offer, a photo. A brand is asked who it is looking for and its budget. An agency is asked how to get its roster. Creating a login finishes the claim.
A link that has expired, has already been used, or belongs to an account that is already owned shows a short message instead of the card. Someone who is already signed in is asked to log out first, because claiming makes a new account.
Before anything is added, the list checks the new names against accounts that already exist and against prospects already on the list.
If a new prospect looks like an existing account, it is flagged and left out, with the account's Zooly address and a plain reason: the same website, the same handle, or the same name. The team can add it anyway by ticking Add even if it looks like an existing account before pressing Add. When it is added anyway, the row keeps a possible duplicate mark.
A name that is already on the prospect list is not added a second time. It appears in the result as already known, with the reason. Repeats inside one paste are reported the same way.