Most outbound teams don't have a targeting problem. They have a signal problem. They know who their Ideal Customer Profile (ICP) is, they just can't tell which accounts are ready to talk this quarter versus next year.

Signal-based prospecting fixes that. And Claude Code, used properly, turns a process that used to require an engineer into something a Go-to-Market (GTM) operator can build in an afternoon.

What signal-based prospecting actually means

Signal-based prospecting is the practice of triggering outreach based on observable buying behaviour, not just firmographic fit. A company hiring three Account Executives isn't the same as one that hasn't posted a sales role in 18 months. A product team that just shipped a billing feature is a different buyer than one that hasn't touched pricing in years.

The signals matter because they answer the question every Sales Development Representative (SDR) is really asking: why now? Without a "why now," your message is just another templated cold email. With one, your open and reply rates climb because the reader sees themselves in the first line.

The problem is that most teams know this and still don't do it. The reason is operational. Pulling signals from across the web, structuring them, matching them to accounts, and routing them into a sequencer is a real engineering job. Or it used to be.

Where Claude Code fits in the GTM stack

Claude Code is Anthropic's command-line coding agent. For GTM operators, that means you can describe a data pipeline in plain English and have working code at the end. It's not a magic outbound tool. It's a way to build the connective tissue between the tools you already pay for.

For Partner UP clients, the workflow usually looks like this:

The reason this matters is leverage. A GTM Engineer with Claude Code can ship a working signal pipeline in two days that would have taken a developer two weeks. We've used this approach to drive a 40% prospect-to-lead increase for a B2B marketplace client, mostly by removing the lag between signal and outreach.

The three signal types worth building for

Not every signal deserves a workflow. Build for the ones that correlate with revenue, not the ones that are easy to collect.

Hiring signals. A company posting roles in your buyer's function is one of the highest-intent signals available. A Revenue Operations (RevOps) hire at a Series B company tells you they're about to invest in systems. A second SDR hire at a 10-person startup tells you they're scaling outbound and need infrastructure.

Product and content signals. Changelogs, release notes, careers pages, and pricing page changes all leak strategy. If a SaaS company adds an enterprise tier, their buying committee just grew. If they ship a new integration, they probably just hired a partnerships lead worth talking to.

Funding and leadership signals. Funding rounds and executive hires are well-covered, which means they're also crowded. Use them, but expect lower reply rates than less obvious signals. The edge is in being faster than the rest of the market, not in finding rounds nobody else sees.

Building a hiring signal pipeline with Claude Code

Here's the actual pattern we use. Assume you want to track companies hiring SDRs or Account Executives within your ICP.

Step one: define the source. Most teams scrape LinkedIn job posts, company careers pages, or pull from a job board API. Claude Code can write a scraper in 20 minutes, but you need to be honest about terms of service and rate limits before you ship it.

Step two: structure the output. Each signal needs a company name, a domain, a role title, a posting date, and a source URL. Claude Code is good at writing parsing logic that normalises messy text into clean rows.

Step three: enrich. Push the company domain into Clay. Clay matches it to your ICP filters, runs an email waterfall on the right contact, and writes back a clean record with a person, an email, and a verified status. Clay's email waterfalls are available on all paid plans, which is the reason most mid-market outbound teams default to it.

Step four: route. Claude Code can write a small n8n workflow or a direct HubSpot API call that creates a contact, stamps the signal as a property, and adds them to a list. HubSpot then triggers the right sequence in HeyReach or Lemlist.

Step five: write the copy. This is where most teams fall down. The signal is only useful if the email mentions it. "Saw you're hiring two AEs in Berlin" beats "Hope this finds you well" every time.

What breaks if you skip the data layer

Signal-based prospecting falls apart without clean data underneath it. We see this constantly when companies bolt a signal tool on top of a messy HubSpot instance and wonder why reply rates are flat.

The order matters. Before you build signal workflows, you need:

• A defined ICP with firmographic filters that actually filter

• Account and contact records that aren't duplicated three times

• A reliable enrichment source, usually Clay for mid-market or ZoomInfo for enterprise targeting

• Lifecycle stages and lead statuses that mean something

If your HubSpot data is dirty, signals just route bad contacts into bad sequences faster. That's a worse outcome than no signals at all, because now you're burning domain reputation on the wrong people. The Clay and HubSpot native integration, available on Clay's Growth and Custom plans, is what we use to keep enrichment writing back cleanly. Lower Clay tiers need a Zapier or Make workaround, which works but adds a failure point.

For more on getting the data layer right, see how we think about [RevOps as the foundation under any outbound motion](/blog).

Where Claude Code helps and where it doesn't

Claude Code is not a replacement for GTM judgment. It's a replacement for the engineer you were going to ask to build the pipeline. The strategy, the ICP, the message, the offer, the routing logic, all of that is still on you.

The teams that get the most out of it are the ones that already know what they want to build. They use Claude Code to compress the build time from weeks to hours. The teams that struggle are the ones hoping the tool will tell them what to do.

What this looks like in practice for a Series A team

A typical Series A B2B SaaS company has one or two SDRs, a founder still in sales, and a HubSpot instance that's six months past its last cleanup. They don't need a 12-signal scoring model. They need two signals that work.

Start with hiring data for the roles your buyers manage. Add one product or content signal specific to your category. Route both through Clay into HubSpot, then into HeyReach with signal-aware copy. That's it. Three weeks of build, then iterate based on reply rates.

The mistake most teams make is trying to build the full Common Room or HockeyStack equivalent in-house on day one. Don't. Build the smallest pipeline that produces booked meetings, then add signals as you prove which ones convert.

FAQ

Do I need to know how to code to use Claude Code for outbound?

No, but you need to be comfortable in the terminal and reading code you didn't write. Claude Code generates working scripts, but you'll be the one running them, debugging when they break, and connecting them to Clay, HubSpot, and your sequencer. Most GTM Engineers we work with picked it up in a week.

How is this different from just using Clay?

Clay is the enrichment and orchestration layer. Claude Code is what you use to build the signal capture and custom logic that sits before Clay. For standard signals like funding rounds or job changes, Clay's built-in sources are enough. For custom signals specific to your buyer, you'll want Claude Code to scrape or parse the source yourself.

What's the minimum stack to run signal-based prospecting?

Clay for enrichment, HubSpot for the system of record, and either HeyReach or Lemlist for activation. Add Claude Code for custom signal pipelines and n8n if you need orchestration between tools. That's the lean stack we deploy for most Series A and Series B clients.

Partner UP works with GTM and RevOps teams on signal-based prospecting workflows and the data infrastructure underneath them. If you're trying to build signal pipelines without a full engineering team, reach out at hello@partneruphq.com or [book a call](/contact).