SDR Outbound Engine

Inside this Project, Claude carries the material that makes sequences yours:

• ICP and persona definitions — who you target and what each persona actually cares about.
• Your messaging voice — sentence length, formality, what you never say.
• The patterns that have performed, once you feed results back in.
• Current target-list context — the segment you are working and what you have already sent into it, so touch four does not repeat touch one.

Run one Project per segment or motion. Voice and proven patterns differ enough between segments that mixing them dilutes both.

Reach for this Project when:

• You are building a sequence for a persona or segment.
• A trigger event has appeared and you want a message out while it is current.
• A prospect has gone quiet and you need a follow-up that adds something.
• You are re-engaging dormant or closed-lost accounts.
• You need to reach a second person inside an account without repeating yourself.

Use something else when:

• You need research on an account before writing — that is Prospect Research Engine, which produces the specific facts this Project needs as input.
• You are mapping a strategic account in depth — that is Account 360 Builder.

Per request, you provide:

• The target persona and the company type or named account.
• The trigger event or pain hypothesis — specific facts, ideally from a research brief.
• The sequence type: which of the five workflows you want.
• Any prior contact with this person or account, so the new message accounts for it.

The specific facts are what separate a usable sequence from a template. Without them this Project will tell you it cannot personalise rather than inventing something.

Keep the case studies to one paragraph each with the numbers in them. Long case studies get quoted at length into emails, which is exactly the wrong outcome.

Claude accepts files up to 30MB each, and there is no fixed limit on how many you add, but everything in a Project's knowledge has to fit the context window. Fewer, sharper files beat a full archive.

Prepare files this way:

• Strip cover pages, legal boilerplate, and navigation chrome — they consume context and carry no signal.
• Prefer text formats. A clean export beats a scanned PDF.
• Name each file for what it contains, not when you exported it. Claude reads filenames.
• One subject per file. A merged reference document is harder to cite precisely.
• Date anything that goes stale, inside the file itself, so Claude can tell you when it is working from old material.

1. Create the Project. Name it for the segment and motion — for example Outbound — Mid-Market Ops Personas.
2. Paste the Project Instructions.
3. Replace the four bracketed fields. The voice guide field matters most; it is what stops the output sounding like every other cold email.
4. Upload the knowledge files, voice guide and persona library first.
5. Run workflow 1 once and read the output aloud. If it does not sound like you, tighten the voice placeholder rather than editing the output — the fix belongs upstream.
6. Feed results back monthly. Tell the Project which messages got replies. That is the only way the pattern memory becomes real rather than notional.

The bracketed fields below are placeholders you replace with your own details before you save the instructions. Leaving them unfilled produces generic output — that is the single most common reason a Project underperforms in its first week.

This Project has four. The original mixed three different placeholder grammars, which made it unclear which were meant to be replaced; all are now in one form:

• [FILL IN: Company Name] — your employer.
• [FILL IN: Your Product One-Liner] — what you sell and the problem it removes, in one sentence.
• [FILL IN: Your Messaging Voice Rules] — three to five rules describing how you write: sentence length, formality, and specifically what you never say. Negative rules are the most useful.
• [FILL IN: Your Meeting Goal] — what a successful reply leads to, and the ask you are actually making. A sequence written without a defined ask drifts into brochure copy.

Role

You are an outbound messaging specialist for [FILL IN: Company Name], writing for an SDR working a defined segment. You write short, specific, human messages, and you refuse to write a good-sounding message you cannot support.

Operating rules

• Every message references something specific and verifiable about the prospect. No fact, no message — ask for one.
• Never fabricate a trigger event.
• Never claim a named customer result that is not in the knowledge files.
• Never imply a relationship or prior conversation that does not exist.
• No false urgency, invented deadlines, or manufactured scarcity.
• Email bodies under 150 words.

The five workflows

1. Cold sequence · 2. Trigger message · 3. Follow-up after no response · 4. Re-engagement · 5. Multithread into a second contact. Each has a defined output format.

The full instructions continue with all five workflow specifications and their output formats, the commands, and escalation.

Five workflows, each with a defined output format — the original specified none, which is why output shape drifted between requests:

• Workflow 1 — cold sequence. Five touches across channels.
• Workflow 2 — trigger message. A single message tied to a live event.
• Workflow 3 — follow-up. A next touch after no response.
• Workflow 4 — re-engagement. Dormant or closed-lost outreach.
• Workflow 5 — multithread. Reaching a second person in an account.

Workflow 1 output — a five-touch sequence for a fictional persona: Director of Operations at a 200–500 employee logistics company, following a first RevOps hire.

Touch 1 — Email, day 1

Subject: your first RevOps hire

Saw you're hiring a Head of Revenue Operations — usually the point where someone finally owns a number three teams have been reporting differently.

The part that tends to surprise people is how much of that first year goes into reconciliation rather than the analysis they were hired for.

Worth fifteen minutes on how other logistics operators handled the first ninety days?

Anchor: confirmed job posting. 61 words.

Touch 2 — LinkedIn connection, day 3

No pitch. Connection request only, referencing the same hiring signal in one line.

Anchor: same trigger. Channel change, not message repetition.

Touch 3 — Email, day 6

Subject: re: your first RevOps hire

One concrete thing rather than a follow-up: a regional carrier we work with cut month-end reconciliation from nine days to four in their first quarter — mostly by fixing where the data landed, not by adding headcount.

Happy to send the two-page version if it's useful. No meeting needed.

Anchor: case study from knowledge files, numbers verified. Gives before asking. 48 words.

Touch 4 — Phone, day 9

Voicemail referencing the emails, naming the reconciliation problem specifically, with a single question to reply to.

Touch 5 — Email, day 14

Subject: closing the loop

I'll stop here — either the timing's off or this isn't the problem you're solving this quarter. Both are fine.

If it becomes relevant after your RevOps lead starts, I'm easy to find.

Anchor: trigger timeline. Closes without guilt or false scarcity. 38 words.

Sequence notes: every touch anchors on one confirmed fact — the job posting — approached from a different angle rather than restated. No fabricated urgency, no invented trigger, no unverified customer claim. Case study numbers trace to the knowledge files.