Cold Outbound Automation Researched, personalised, sequenced — inside the limits that protect your domain
Each prospect gets a first line written from real research — and the email is refused when there is nothing specific to say.
- Installed for you · one-time$249
- Run for you$297/month
- Custom build · this process$4,000–7,000 one-time
- Starts whenSchedule Trigger (daily batch) + reply event
Covers up to 2,000 new prospects a month with one sequencer; data and sequencer subscriptions are yours. LinkedIn steps, multi-language, or a custom research tool set are a custom build. Custom builds start at $2,000 site-wide; this band is for this process and its integrations.
Cold Outbound Automation is an automation Agentmelt installs in your own tools. The free template is a file you import yourself; the kit ($49) adds the setup guide and prompts. Installed for you costs $249 one-time and is live within two working days of access, with a recorded walkthrough and 30 days of fixes. Run for you costs $297/month, with setup and maintenance included. A custom build of this process is $4,000–7,000 one-time; custom builds start at $2,000 site-wide. No call is needed for any option. Written by Max Zeshut; last revised 2026-09-21.
Reference stack: Claude + n8n. Node names are on each step; the file is under the template.
Founder at Agentmelt
Reviews every build before it goes live; the workflow escalates to a person, never past one.
What is cold outbound?
Cold outbound automation does the research and the writing a good SDR would do for each prospect, at a volume a person cannot sustain, inside limits that protect your domain. The workflow builds the batch from your ideal customer profile, researches each account — site, news, hiring, technology — writes a first line and a short email grounded in that research, passes a quality gate that rejects the generic, loads the sequence into your sender with reply and bounce detection, classifies replies, and hands interested ones to a rep with the thread. Daily send caps, warm-up and suppression lists are part of the design, not an afterthought.
- Reps spend the morning researching accounts and the afternoon typing the same email.
- The “AI SDR” we trialled sent 2,000 emails in a week and burned the domain.
- Every first line is “I noticed you’re the VP of Sales at…”.
- Replies land in a shared inbox and interested prospects wait a day.
- Nobody knows which angle or segment actually gets meetings.
If two of these are yours, this is the process to look at first. Free audit — or read on.
What it looks like once it runs.
An outbound sales agent does the research and writing a good SDR would do for each prospect, at a volume a person cannot sustain, inside limits that protect your domain. The workflow builds the list from your ICP, researches each account (site, news, hiring, tech), writes a first line and a short email grounded in that research, sequences with reply and bounce detection, and routes replies to a rep.
- 6:00 Today's batch of prospects is pulled: matches your profile, not in the CRM, not contacted recently.
- 6:30 Each company is researched; a note with three to five facts is written.
- 7:00 A first line and a sub-80-word email are drafted from the note; anything generic fails the quality gate and is not sent.
- Through the day Replies are classified; interested ones reach a rep with the thread and the research note.
Outbound is either generic at volume (ignored, and damaging to your domain reputation) or personal at a volume of ten a day. SDRs spend most of their time on research and copy, not conversations.
Every email in the sequence references something true and specific about the prospect. Volume stays within deliverability limits, replies are detected and handed to a person within minutes, and the team measures reply and meeting rate per segment and per angle instead of guessing which messaging works.
Trigger, then 8 steps.
Seven of the eight steps run without anyone. Step 04 waits for a person.
Schedule Trigger (daily batch) + reply event
Build the batch
Pulls prospects matching the ICP from Apollo or Clay that are not in the CRM, not suppressed, and not contacted in 90 days. Daily batch size is capped per sending mailbox.
Research each account
Reads the company site, recent news, job postings and tech stack (via enrichment and a web-fetch tool) and writes a five-line research note: what they do, what changed recently, a plausible trigger for your offer.
Write the first line and email
From the research note and your messaging playbook, writes the opening line and a sub-80-word email for the chosen angle. Structured output enforces length, no adjectives about your product, and one clear ask.
Quality gate
Emails are checked for specificity (must reference the research note), banned phrases, and length. Failures are rewritten once, then dropped from the batch rather than sent generic.
Load into the sequencer
Pushes the prospect and the personalised steps into Instantly, Smartlead, Lemlist or Outreach, which handle sending, warm-up and inbox rotation.
Detect replies and classify
Every reply is classified: interested, question, not now, wrong person (with a referral), unsubscribe, out of office. Sequences are stopped on any human reply.
Hand to a rep
Interested and question replies go to the rep with the thread, the research note and a suggested response; 'not now' is scheduled for a follow-up; referrals create a new prospect.
Report by angle and segment
Weekly: sent, replies, positive replies, meetings by segment, angle and mailbox; deliverability metrics from the sequencer; the losing angle is retired.
Every email in the sequence references something true and specific about the prospect
Every email in the sequence references something true and specific about the prospect.
One decision, at step 04. Seven steps ran without you.
Each pattern page explains it in plain words and lists every automation that uses it.
The template is built on n8n (how the n8n AI Agent node works); the same steps run on Make, Zapier or plain code. The tool is chosen to fit your accounts, you own the result, and the template comes in all three formats.
Data it touches
- Prospect data (Apollo, Clay, LinkedIn Sales Navigator exports)
- Company sites, news, job posts (web fetch)
- CRM for suppression and logging
- Sequencer (Instantly, Smartlead, Lemlist, Outreach)
- Messaging playbook and angles
Guardrails
- Daily volume per mailbox is capped and warm-up is respected; the workflow never sends outside the sequencer's limits.
- Emails that fail the specificity check are not sent; generic filler is worse than silence.
- Any human reply stops the sequence immediately; unsubscribes are honoured across all lists.
- Suppression lists (customers, open deals, competitors, opt-outs) are applied before research spends any credits.
Personalisation that survives the quality gate
'I saw you are the VP of Sales at Acme' is not personalisation, it is a mail merge. The research step produces a note with something that changed — a new office, a job post for the role your product replaces, a product launch, a quote from the founder — and the copy step must reference it. The quality gate checks that it did. Emails that cannot be made specific are dropped. Reply rates on batches that pass this gate are typically several times those of template outbound, and the domain reputation stays intact because the volume of low-quality mail is zero.
Deliverability is the constraint, not the model
The model can write a thousand emails an hour; a mailbox can send 30–50 a day without harm. The workflow treats the sequencer's limits as hard constraints, rotates across mailboxes it is given, and reports bounce and spam-complaint rates weekly. If those move, volume drops automatically. Outbound automation that ignores this burns the domain in weeks.
Angles as experiments
Each batch is tagged with the angle — the problem you led with — and the segment. The weekly report shows reply and positive-reply rate per angle per segment. After a few weeks the workflow retires losing angles automatically and the team writes new ones. This is the part that compounds: the messaging gets better because the data says what worked, not because someone had a feeling.
What it connects to.
| Tool | Role in this workflow |
|---|---|
| n8n | List building, research loop, quality gate, reporting |
| Claude | Research notes and copy, with structured output |
| Apollo / Clay | Prospect data and enrichment |
| Instantly / Smartlead / Lemlist | Sending, warm-up, deliverability |
| HubSpot / Salesforce | Suppression and logging |
Want this running without building it?
$297/month
We set it up in your tools, run it on our servershost it on our n8n cloud (or deploy it to your own n8n, Make or Zapier), watch it, and fix it when something changes. Setup included, cancel monthly; the automation stays yours if you leave.
$4,000–7,000 one-time
Your systems, your rules, your edge cases. A one-off build delivered into your accounts to keepon Claude and n8n, delivered into your accounts to keep, with documentation and a recorded walkthrough.
Covers up to 2,000 new prospects a month with one sequencer; data and sequencer subscriptions are yours. LinkedIn steps, multi-language, or a custom research tool set are a custom build.
Also sold one-time, installed in your own accounts, as part of the sales reps package and the sales teams package — do it yourself $49 · installed for you $249 · done for you $490.
When this becomes a custom build.
The subscription covers the standard version — $297/month, setup and maintenance included. Three things move it to a custom build ($4,000–7,000 one-time), delivered into your accounts to keep. If one of them is you, say so on the audit and the plan prices both.
LinkedIn steps in the sequence, with the limits and tooling that requires.
Outreach in several languages or markets.
A custom research tool set — your own data, scraped sources — or above 2,000 new prospects a month.
Running the standard version first is the usual path: the exceptions it logs in the first month are the scope of the custom build, with your own numbers. How custom builds work.
Frequently asked questions.
Is this compliant with anti-spam law?
B2B cold email is lawful in most jurisdictions with a clear identity, a working unsubscribe and honoured opt-outs, which the workflow enforces. Regions with stricter rules (some EU countries, Canada under CASL) are configured with consent or opt-in requirements per segment.
Which sequencer do you recommend?
Instantly or Smartlead for volume and inbox rotation; Lemlist for smaller, more personal programmes; Outreach or Salesloft if your team already lives there. The workflow feeds any of them.
Can it do LinkedIn steps?
LinkedIn automation carries account risk and is offered only as a custom build with conservative limits, or via a sequencer that supports it natively.
Is cold outbound automation compliant with anti-spam law?
B2B email to business addresses is lawful in most jurisdictions when the message identifies the sender, is relevant to the recipient’s role, and honours opt-outs immediately; the EU and Canada are stricter and the workflow’s suppression lists and per-country rules reflect that. Send caps, warm-up and bounce handling protect deliverability; a person owns the list and the rules.
Which sequencer does it work with?
Lemlist, Instantly, Smartlead, Apollo, Outreach or Salesloft — the workflow writes the personalised steps into the sequencer you already run and reads its reply and bounce events. LinkedIn steps, with the limits and tooling those require, are a custom build.
How is this different from an AI SDR product?
The AI SDR products own the sending and the volume; this workflow owns the research and the writing and hands the sending to your sequencer under your limits, with a quality gate before anything loads. The difference shows in the reply rate and in whether your domain is still delivering in month three.
What this blueprint is built on.
Vendor documentation the steps rely on, and the rules the guardrails are written against.
- Google — Email sender guidelines support.google.com/a/answer/81126
- FTC — CAN-SPAM Act compliance guide www.ftc.gov/business-guidance/resources/can-spam-act-compliance-guide-business
- Apollo.io — API documentation docs.apollo.io/