← all playbooks

Work your hundred most wanted accounts every week, ten at a time

Each week, the ten people on your Dream 100 most worth a message, with anyone contacted recently held back, fresh research on each one, and a connection note or warm message drafted in your voice. Staged for you to send by hand.

use when
you keep a list of the accounts and people you most want to work with, and it only gets touched when the desk is quiet
starts from
A list

The prompt

Paste it into Claude Code or the Claude desktop app with Hyreflow connected. The first line tells your agent to use Hyreflow, so it reads the play, asks before it spends anything, and hands the work back to you.

paste this into Claude
Use hyreflow and run this week's sweep of my Dream 100.

The list is here: <PATH_OR_TABLE>. My notes on which segments matter most
and how I write are here: <PATH_TO_MY_NOTES>.

1. First, hold back anyone I have contacted in the last <SUPPRESSION_DAYS>
   days. Do this before anything else.
2. Refresh the rest for a change of title or company. If the name that
   comes back is not the person on my list, do not trust the row.
3. Rank them on recent change, how long since I last wrote, and the
   priority of their segment. Take the top 10.
4. Research those 10 for something current worth mentioning.
5. Draft a connection note for each, under 300 characters, in my voice.
   If I am already connected to someone, draft a warm message instead.

I send at most <DAILY_SEND_CAP> a day. Do not hand me more than I have
left today.

Refresh five rows first and tell me what the full run costs before you do
the rest. Stage it, do not send anything. I send every message myself.

Replace every <PLACEHOLDER> with your own detail. Everything else can stay as written.

What you need first

  • Your Dream 100 in a table you own, such as a spreadsheet or a database, with a LinkedIn URL and a last-contacted date on each row
  • Written notes on which segments matter most and how you write, so the ranking and the drafts are yours
  • A Hyreflow workspace with credits
  • Someone who will send the messages by hand each week and mark them as sent

Tools it can reach for

The agent picks per step from what your workspace has. Nothing here is required by name.

What happens when you run it

Free steps are marked free. Anything that spends credits is marked, and the agent asks before the first paid run of any size.

  1. 1

    Hold back anyone contacted recently

    free

    The first step of every sweep, before anything is ranked. Rows with a last-contacted date inside your window are set aside. The default window is ninety days, and people return to the pool by themselves once it has passed.

  2. 2

    Refresh the rest

    credits

    The remaining rows go through the enrichment waterfall on their profile URL to catch a change of title or company. The name that comes back is compared with your list, because a dead or reassigned profile address can return a stranger with no error.

  3. 3

    Score, rank and take ten

    free

    Recent change, time since you last wrote and the priority of the segment are combined into one score. It is plain arithmetic with no AI call, so it is free. The top ten go forward.

  4. 4

    Research the ten

    credits

    Each of the ten is researched for recent news, funding or another company signal that a message can refer to. This is charged per request, and a broader question costs more.

  5. 5

    Draft the note in your voice

    credits

    A connection note per person, held to the 300-character limit, built from the research and your notes on how you write. Someone you are already connected to gets a warm message draft instead, for whoever in your firm knows them.

  6. 6

    Stage within your daily cap

    free

    Drafts are written back to your table and marked as staged. The number handed over today is your daily cap less what you have already sent today, so the list never invites you to overshoot.

  7. 7

    You send, and the row is dated

    free

    A person sends every message by hand. When you mark a row as sent, its last-contacted date is set to today. That date is what the first step of next week's sweep reads.

What a run costs

Credits are spent per contact the play actually works, and a lookup that finds nothing usually costs nothing. The two figures are the run where the first provider answers and the run where every lookup walks its full chain.

contactsif the first provider answersif every lookup walks the chain
25under $13.3 credits$1.919 credits
100under $17 credits$6.868 credits
500$2.727 credits$33330 credits
1,000$5.252 credits$65650 credits

Free before anything is charged

  • Hold back anyone contacted recently
  • Score, rank and take ten
  • Stage within your daily cap
  • You send, and the row is dated

What moves the number

  • Coverage on work history. The chain stops at the first provider that answers, and only that provider bills.
  • How many contacts survive the free filters. Everything dropped before the paid steps costs nothing.
  • The scoring and drafting steps run on the metered agent, charged on what they read and write rather than per contact, so they sit outside this table.
  • Providers you connect with your own key. Those calls bill your account, not your credits.

An estimate, not a quote, priced at the volume credit rate. Your agent sizes the run against your own workspace and tells you what it will cost before it spends anything.

A short list worked weekly beats a long list worked when it is quiet

Every recruiter has a version of this list: the hundred accounts and people who would change the desk if they became clients. Nearly everyone works it the same way, in bursts, when a deal falls through and the pipeline looks thin. Business development done only in the gaps produces exactly the feast and famine it was meant to cure.

The fix is not more effort. It is a fixed list, a fixed day and a small number. Ten a week, every week, gets round a hundred names several times a year without anybody hearing from you twice in a quarter.

What makes that sustainable is that the slow parts are done before you sit down: checking who has moved, finding something current worth saying, writing the first draft. What is left for you is judgement and the send.

The list itself can come from anywhere. One good source is your own network: turn a LinkedIn connections export into a scored prospect database builds and screens the pool. This play assumes the hundred are chosen and is only about the weekly rhythm.

What you get back

Every week, ten rows from your list. Each carries what changed since the last refresh, a research note with its source, and either a connection note inside the character limit or a warm message for someone already in your network. All of it is marked as staged in your own table. You also get a line on who was held back and why, so you can see the suppression working.

Variations worth knowing

Build the hundred first. If you have the criteria but not the names, the list can be built with one people search per segment, de-duplicated on profile URL, and checked against your existing connections so that nobody you already know receives a cold invitation.

Segments with different weight. The score uses the priority you give each segment, so say in your notes which ones matter most.

A one-off list. For a campaign list, not a standing one, see write one sourced opening line per person on a list you have already built.

Where this goes wrong

Ranking before suppressing. The order is the point. Rank first and the most recently contacted people float to the top.

Not marking rows as sent. Suppression only knows what the table tells it. Send a note without dating the row and the same person is back next week.

A profile address that belongs to someone else. Handles get abandoned and reassigned. The refresh can return a stranger's details without any error, which is why the returned name is checked against yours.

A hundred that is really four hundred. A loose list turns the sweep into a feed of whoever changed job this week.

Automating the send. It is the one hard rule here. Everything before the message is automation. The message is yours.

Questions

Why only ten a week?

Because ten is a number that gets researched properly and sent by a person, week after week. The default cap on manual invitations is twenty to twenty-five a day, and the sweep stays well inside it on purpose. A hundred generic notes in one afternoon is a campaign. Ten specific ones every week is a relationship with a market.

Why does the suppression step run first?

Because a sweep that ranks first will put someone you wrote to three weeks ago back at the top of the list, and a second message that soon costs more goodwill than silence would have. So the hold-back runs before scoring, before research and before anything that costs credits. Nobody has to remember to release people afterwards. The window simply expires.

Why will it not send the invitations for me?

Because automated sending on LinkedIn puts your account at risk, and the account is the asset this whole play depends on. The research and the drafting are the slow part, so those are what the agent does. The send is a hard stop by design, including when the sweep runs on a schedule.

What about people on the list I am already connected to?

They never get a connection note, because there is no connection left to request. The same research feeds a warm direct message instead, routed to whoever in your business is connected to them, and staged and sent by hand in the same way.

Can it run by itself every Monday?

Yes. The weekly sweep is the part built for that: once you trust a hand run, it is saved as a workflow on a weekly schedule and the staged drafts are waiting for you when you start the week. It still stops at the draft and never sends. Building and screening the list stay as runs you start yourself. Prove the sweep by hand first, because a scoring rule that is slightly wrong picks the wrong ten every week.