Find in-house teams with a capacity gap, not a performance gap
Your ICP for this playbook is marketing leaders at companies with roughly 50 to 500 employees who run marketing in-house but are stretched thin on a specific channel or skill. Target titles: Marketing Director, VP Marketing, Head of Growth, Demand Gen Manager, CMO.
Look for companies where a specialist role sits open, where the team lost someone in the last two quarters, or where a hiring freeze landed on marketing while growth targets stayed the same. These are capacity signals, not performance complaints, and that distinction matters. Opening with "your ads are underperforming" puts a stranger on the defensive. Opening with "your team looks stretched" starts a different conversation, because it names a problem they already feel and did not create.
Write down two or three specific gaps you want to target before you build the list: a missing paid social hire, a lifecycle email program nobody owns, an SEO backlog with no one assigned. A retainer pitch aimed at a named gap reads as relevant. A generic "we help companies grow" pitch reads as a form letter.
It also helps to separate companies that are permanently understaffed from companies mid-hire for the role you'd fill. Both are fits, but the email you send each one differs: one is a longer-term retainer conversation, the other is a bridge conversation while they finish hiring.

Source contacts with AI Leads and filter for strain signals
Describe your target in AI Leads: company size range, industry, and the titles from step one. AI Leads finds and enriches matching prospects, and the quality funnel filters out weak matches before they cost you credits.
If you already have a warmer list, for example companies you found through LinkedIn or a conference, import that list as a CSV instead. Either path gets you into the same campaign; the difference is just where the contacts start.
Keep the list narrow rather than wide. A tight list of 150 to 300 companies that genuinely match your capacity-gap thesis will outperform a loose list of 2,000 that only match on industry. If your first batch comes back too broad, describe the target again with a narrower size range or a more specific industry vertical, and let AI Leads run a second pass rather than sending to a mismatched list.
Let contact research surface the actual gap before you write
Once contacts are in the campaign, automatic prospect research runs before drafting. The transparency view shows you what it found on each contact: company facts, the person's role, and any signals it picked up, plus what it actually used in the email.
For this playbook, check that research is surfacing the kind of detail that supports a capacity-gap opener: team size, function coverage, or recent changes at the company. If a contact's research comes back thin, that prospect is a weaker fit for a personalized retainer pitch and worth reviewing before you send.
Spend ten minutes reading through research on a sample of twenty contacts before the campaign goes live. If most of them surface something usable, the list and the offer are aligned. If most come back generic, the fix is usually the ICP description in AI Leads, not the email itself.

Set the campaign goal and describe your offer
Create a new campaign and set the goal to book a meeting, since that is the natural next step for a retainer conversation. Campaign guidance walks you through Describe, Teach, and Practice.
In Describe, write the actual retainer offer: which channel or function you cover, roughly what scope looks like, and who it is for. Be concrete. "We run paid social as an extension of your team" gives the AI more to work with than "we help companies grow faster."
Use Practice to review sample drafts against a few contacts before turning the campaign on. This is where you catch a generic opener before it goes out to three hundred people. Read at least five drafts end to end, not just the subject lines, since a strong opener followed by a generic closer still reads as a form email.

Teach the AI to lead with capacity, not results
Paste a few of your own past emails, or write two or three samples in the tone you want, and the AI reuses that tone across the campaign. Keep the capacity framing explicit in your samples: name the gap, name what filling it looks like, ask a small question.
Subject: extra hands on paid social
Hi Priya,
Saw the Growth Marketing Manager role has been open at Kindled for a couple months. Filling a specialist seat like that takes time, and paid social usually doesn't wait.
We run paid social as an extension of in-house teams, so someone still owns strategy on your side and we handle execution day to day. A few teams we work with brought us in during exactly this kind of gap.
Worth a 20-minute call to see if it's a fit while you're hiring?
DanNotice the opener names a specific, observable fact rather than a guess about performance. That is what keeps a capacity pitch from sounding like every other cold email in the inbox. Avoid the temptation to add a results claim into the same email; save that for the call, once there's a real conversation to have it in.
Build a short, staged follow-up sequence
Sequences send multi-step follow-ups on a delay, and every step stops automatically the moment someone replies. For a retainer pitch, three touches spaced three to four days apart tends to work better than a long drip.
Vary the angle across the three touches instead of repeating the same pitch. Touch one names the capacity gap. Touch two can reference a specific example of the kind of work you'd take off their plate. Touch three can lower the ask entirely, something like offering to send a one-page scope instead of asking for a call.
Resist adding a fourth or fifth touch to this particular sequence. Capacity-gap prospects tend to reply within the first three emails or not at all, and pushing past that point mostly adds noise without adding replies.

Handle the two objections that show up most
AI reply handling reads the actual content of a reply and answers it, so when "we already have an agency" or "we handle this in-house" comes back, the reply won't ignore what was said. Review a sample of these replies yourself early in the campaign to confirm the tone matches how you'd actually respond.
Both objections are openings, not closes. "We have an agency" often means capacity is still tight even with outside help. "We do this in-house" often means one person is covering more than a full-time role. The reply worth sending acknowledges the objection directly, then asks one specific question rather than repeating the pitch.
Once a real conversation opens up in the inbox, use that thread to ask about scope and timeline rather than immediately pushing for a call. Some of these prospects will book straight from the reply thread once the capacity gap is confirmed.
Track saturated-market metrics and tune the angle
Agency-to-in-house outreach sits in a saturated market: expect roughly 1 to 3 percent reply rates and 0.3 to 1 percent meeting rates. Check opens, replies, and meetings per campaign, and per-step performance to see which touch in the sequence is doing the work.
If reply rate sits below that range after a few hundred sends, the fix is usually the list or the opener, not more volume. Narrow the ICP further, or rewrite the sample emails you taught the AI with a sharper, more specific capacity gap.
Review results by touch, not just by campaign total. A sequence where touch one is carrying nearly all the replies suggests the follow-ups aren't adding anything and could be trimmed or rewritten with a genuinely different angle.

Pro tips
Time the send to a real trigger
A contact whose research shows a recent hiring gap or a stalled job posting is a better send target than a random name on the list. Prioritize contacts with a visible signal over ones without.
Keep the first ask small
Ask for a reply on touch one, not a calendar booking. A low-commitment ask gets more responses from people who are curious but not ready to commit time.
Let one sample email carry the tone
You only need two or three well-written samples in Teach for the AI to hold the capacity-gap framing consistently. More samples with mixed tone will muddy it.