The Amplemarket Duo Playbook: How to Auto-Push Buying Signals Into Personalized Sequences

I spent three weeks building something that replaced an entire SDR's signal research workflow. Not a tool. Not a platform. A pipeline.

One that discovers buying signals across the internet, qualifies them against an ICP, and pushes them directly into Amplemarket Duo so it can generate personalized multichannel sequences automatically.

No spreadsheets. No copy-pasting. No "let me check if this person is a fit." The signal shows up, the qualification happens, and Duo starts writing. Here's exactly how it works.

Why Signal Data Dies in Dashboards

Most B2B outbound teams run the same playbook they ran in 2019. Build a list. Write a template. Hit send. Maybe add a few personalization tokens so it doesn't look completely robotic.

The results? Industry average reply rates sitting between 1-3%. Sales cycles stretching past 60 days. Reps burning through lists faster than marketing can build them.

The issue isn't the sending. Tools like Amplemarket handle execution beautifully. The issue is what happens before the send. Specifically, the signal layer.

When an outbound rep reaches out to someone who just talked about their exact pain point on a podcast two weeks ago, the reply rate jumps to 8-12%. When they reach out to someone from a cold list with no context? That's the 1-3% world.

The gap between those two numbers is the signal.

The Signal-to-Sequence Architecture

Duo is Amplemarket's AI copilot for outbound. Feed it context about a prospect and it generates personalized multichannel sequences — emails, LinkedIn messages, the whole thing — tuned to that specific person's situation.

But Duo isn't magic. It's only as good as the context it receives. Give it a name and a job title, you get generic outreach. Give it a name, a job title, the podcast they appeared on last week, the specific pain they described, and the growth signal that explains why they need to solve it now — Duo writes something that actually sounds like you did your homework.

This playbook is about automating that context delivery. The system runs in five phases.

Phase 1: Signal Discovery

An AI agent scans the internet for fresh buying signals. These aren't generic intent signals like "visited a pricing page." These are public, specific, verifiable moments where someone in your ICP demonstrated pain.

Each of these is a reason to reach out. Not a cold reason. A warm one.

Phase 2: ICP Qualification

Not every signal is worth pursuing. A podcast interview might surface someone interesting who sells to a completely different market. The pipeline runs every discovered signal through an ICP qualification layer.

This checks company size, industry fit, buyer role match, and what I call "pain alignment." Does the signal they demonstrated actually match the problem you solve? A founder complaining about slow lead response on a podcast is a perfect signal for a speed-to-lead solution. That same founder complaining about their product roadmap is not.

Phase 3: Enrichment

Qualified signals get enriched through Amplemarket's data layer. Email, phone, LinkedIn profile, company details, tech stack, growth metrics. This is where the prospect goes from "interesting signal" to "complete profile ready for outreach."

I use Amplemarket's enrichment API here. Create a lead list, add the prospects, let enrichment run. This usually takes 30-60 seconds per batch.

Phase 4: Sequence Generation via Custom Signals

Here's where Duo enters the picture.

Amplemarket's Custom Signal API lets you create webhook endpoints for different signal types. Each signal type gets its own webhook URL and its own set of messaging instructions.

For example, I have one Custom Signal called "Podcast Pain Signal" with instructions that tell Duo:

"This prospect appeared on a podcast and discussed a pain point related to our solution. Reference the specific podcast, the specific pain they described, and connect it to how we solve that exact problem. Tone should be casual, founder-to-founder. No corporate speak."

Another Custom Signal called "Funding Round Signal" has different instructions:

"This company just raised capital and is likely investing in go-to-market infrastructure. Lead with the growth angle, not the pain angle. They're excited about scaling. Match that energy."

When the pipeline discovers a podcast signal, it fires a POST request to the podcast webhook. When it discovers a funding signal, it fires to the funding webhook. Duo receives the context and generates the sequence accordingly.

The JSON payload looks like this:

POST /custom_signals/{webhook_token}/entries

{
  "first_name": "Sarah",
  "last_name": "Chen",
  "company_name": "Acme Corp",
  "email": "sarah@acme.com",
  "linkedin_url": "linkedin.com/in/sarahchen",
  "extra_context": "Appeared on SaaS Scaling podcast
    March 2026. Discussed challenge of converting
    website visitors to pipeline. Said their sales
    team responds to inbound leads in 4+ hours.
    Mentioned evaluating de-anonymization tools.
    Company recently expanded SDR team from 2 to 5."
}

That extra_context field is everything. It's the signal, the pain, the timing, the evidence. Duo reads it and generates outreach that references the podcast, the specific challenge, and connects it to your solution.

Phase 5: Review and Send

Duo generates the sequences. You review them. Adjust if needed. Hit send.

The review step matters. I don't recommend going fully hands-off on outreach copy. But the difference between reviewing AI-generated sequences with rich context versus writing every email from scratch is the difference between 20 minutes and 4 hours per batch.

Signal Types That Convert Best

After running this pipeline for several months, certain signal types consistently outperform others.

Podcast appearances convert at 2-3x the rate of any other signal. When someone sits on a mic for 45 minutes and talks about their challenges, they've given you a roadmap for exactly what to say. The outreach feels relevant because it is relevant.

Funding rounds convert well but need the right angle. Don't lead with "congrats on the raise." Everyone does that. Lead with the operational challenge that comes next: "Series A usually means 3x pipeline pressure in the next 12 months. Most teams aren't built for that yet."

Hiring signals are underrated. When a company posts for a role they've never had before, it means they've identified a gap. That gap is your opening.

Competitor complaints are gold but rare. Someone publicly saying they're frustrated with a competing solution is about as warm as outbound gets.

Getting Started

You don't need to replicate this exactly. The core principle is simple: automate the connection between signal discovery and sequence generation.

The whole thing can run as a Claude skill, an n8n workflow, or a simple Python script. The architecture matters more than the implementation.

Frequently Asked Questions