In partnership with FullEnrich
Every B2B team is sitting on email addresses nobody has identified: webinar registrants, event signups, gated downloads, trial requests and CRM records with an email and nothing else. FullEnrich’s reverse email lookup turns one email, personal or work, into the person and company behind it, including name, title, seniority, company, headcount and LinkedIn, for 1 credit per match and nothing on a miss. If you want to build your own waterfall with custom logic, Clay is the workshop for that, and FullEnrich runs inside it as a provider. If you just want the answer, FullEnrich runs the waterfall across 20+ providers for you, with paid plans from $29 a month. Try it on your last webinar list with 50 free credits.
The 10-intro ask that built $10M in pipeline
The Swarm has about a dozen people. Over the last 12 months they built more than $10M in pipeline, and a big chunk of it came from people who don’t work there: their investors, their advisors, and their team’s old colleagues.
David Connors, their co-founder and CEO, came on the show and walked through the whole motion live. I came away with a play I think most B2B teams are leaving on the table, and a strategic question I can’t stop thinking about.
The problem everyone has
Every company I’ve worked with has the same moment. A rep needs into an account, so someone drops a message in Slack: “Does anyone know someone at Walmart?” Then nothing. Or three people check LinkedIn, one remembers a guy from 2019, and the trail goes cold.
Teams try to fix this two ways. They buy a tool that requires everyone to connect their LinkedIn and email, and adoption stalls at a handful of people. Or they lean on advisors who promise intros and never make them, because they’re busy and nobody asks them in a way that’s easy to say yes to.
Meanwhile cold outbound keeps getting harder. AI made it cheap to send, so every buyer’s inbox is full. David put it plainly: trust and a human relationship are what cut through now.
What David actually does
The Swarm’s internal motion has four parts.
1. Map the network passively. You don’t need everyone to sign up. The Swarm maps former colleagues, school overlaps, investor networks, and LinkedIn engagement from public data, then lets a few key people add their private email and calendar data. David’s rule: get the top 20% to contribute, and you usually get 80 to 90% of the best relationships. Usually that’s the leadership team and a few investors.
2. Score which relationships are real. A LinkedIn connection alone is “Potential.” Add emails or a meeting and it becomes “Likely.” Working together at a small startup, same office, same function, is “Confirmed.” Same big company, different country, doesn’t count for much.
3. Carve the warm accounts out of your list. Start with your target accounts (they build theirs in Clay). Layer intent signals to narrow 10,000 accounts to roughly 1,000. Then split: accounts with a warm path go to the intro motion, everything else stays in cold outbound and ads.
4. Pick the play and batch the ask. There are three plays. Direct: your connector knows the buyer and forwards your note. Indirect: your connector knows someone else at the account who passes it along. Mention: you reach out yourself and reference a shared customer. Then, once a month, each connector gets one email with the top 10 intros you think they can help with. They reply with the numbers they know well. David’s typical result: four out of ten.
The stack for them is simple. Clay for the account list, The Swarm for relationship data, Claude through the Swarm MCP for ad hoc hit lists, Attio as the CRM, and Google Workspace to send.
Where I’d push this further
Midway through the conversation, I asked David: what if this was the strategy? What if, instead of spending the next dollar on ads, a company spent it acquiring networks?
He agreed, and I think it deserves more attention than it gets. Most companies pick advisors for credibility on a slide. Very few pick advisors for who they can reach. If you treated advisor recruiting like channel strategy, you’d look at your top 50 accounts, find the people with real relationships into them, and offer advisory shares or revenue-tied comp for a monthly intro commitment. David said most good advisors will do it for advisory shares if they believe in the product.
Two more things stood out.
First, I guessed on air that the mention and one-hop plays would drive most of the volume, and direct intros would drive most of the pipeline. David confirmed it. So design for both: use mentions to cover ground, and protect direct intros for your highest-value accounts.
Second, the Sequoia “top 1%” exercise. They gave partners and founders a list of 200 to 300 people from their own networks and asked them to mark the ten they’d vouch for. Five minutes each. That’s proprietary data no competitor can buy, and it works for recruiting as much as for pipeline. When I’ve hired sales reps, the best signal always came from people who’d worked with the candidate twice. Almost nobody writes that down.
Run this Monday
List every investor, advisor, and former-colleague-rich leader who’d plausibly help. Pick the top 20%.
Ask each of them for their LinkedIn Connections export (Settings, Data privacy, Get a copy of your data).
Match their connections against your top 1,000 accounts. Score each path Potential, Likely, or Confirmed.
Build each connector’s top 10 for this month. Send one email per connector. Ask for numbers back.
Write a forwardable note for every yes. Under 120 words, one proof point, one ask.
Tag the lead source “Network intro” in your CRM so you can show the pipeline in 90 days.
Free: The Warm Intro Engine
I turned David’s play into a Claude skill you can install. Give it your target accounts and your connectors’ LinkedIn exports. It scores every path, picks the right play, builds each connector’s monthly top 10, drafts the forwardable notes, and sets up a tracking sheet. If you have The Swarm MCP connected, it uses that instead.
To install: download the file, go to Claude Settings > Capabilities > Skills, and upload it. Or paste it into a Project’s instructions.
[Get the Warm Intro Engine at GTMAIpodcast.com]
Watch or listen
Full episode: GTMAIpodcast.com
The Swarm: theswarm.com
David on LinkedIn: linkedin.com/in/connorsdavid
Warm Intro Engine Skill.md
---
name: warm-intro-engine
description: Turn your team's, investors', and advisors' networks into a monthly warm-intro motion. Scores relationship strength, picks the right intro play, batches asks, drafts notes, tracks pipeline.
---
# Warm Intro Engine
Inspired by David Connors, Co-Founder and CEO of The Swarm, on the GTM AI Podcast. The Swarm built over $10M in pipeline in 12 months using its own cap table, investors, advisors, and team for warm intros. This skill runs that play with exports you already have. GTMAIpodcast.com
## What this does
1. Maps who in your extended network knows people at your target accounts.
2. Scores each path Potential, Likely, or Confirmed, then picks the right intro play.
3. Builds one batched monthly ask per connector (max 10 intros), drafts forwardable notes, and sets up tracking so you can prove the pipeline.
## What you need
- **Target accounts** (CSV): company name, domain, fit tier (A/B/C), optional intent signals column. Export from Clay, your CRM, or ZoomInfo.
- **Target contacts** (optional CSV): name, title, company, LinkedIn URL for the buyers you want to reach.
- **Connector roster**: everyone who might introduce you. Name, relationship to you (team, investor, advisor, customer, partner), and their LinkedIn URL.
- **Connector networks** (any you can get, with their permission):
- LinkedIn: Settings > Data privacy > Get a copy of your data > Connections. Produces `Connections.csv` (First Name, Last Name, URL, Email Address, Company, Position, Connected On).
- Work history: each connector's past employers, dates, location, function (copy from their LinkedIn profile).
- Optional: email or calendar contact exports.
- **Suppression list**: people or companies to never route through or contact.
- **Optional upgrade**: if The Swarm MCP is connected in Claude, use it for path mapping instead of steps 2 and 3 below. Prompt pattern: "Use Swarm MCP to find intro paths to [domain]. List connectors, connection strength, and how they know each other."
Start small. Getting the top 20% of your people to contribute (leadership, a few investors, your best advisors) usually covers most of the strongest relationships.
## Process
### Step 1. Narrow the account list
Start from fit, then intent, then relationships. Example: 10,000 target accounts, filtered by intent signals to the top 1,000, then checked for warm paths. Everything without a warm path stays in your normal cold and ads motion.
### Step 2. Find paths
For each target account and each connector, check:
- **Connection match**: a row in the connector's `Connections.csv` where Company matches the target (normalize names: drop Inc, LLC, punctuation; match on domain where possible).
- **Work overlap**: connector and target contact worked at the same company with overlapping dates.
- **Education overlap**: same school, overlapping years.
- **Interaction**: emails, calendar meetings, or LinkedIn engagement (comments, reactions) between them.
- **Investor or customer tie**: connector invested in, advises, or is a customer of the target.
Also note one-hop paths: the connector knows someone at the account who is not the buyer (for example the Chief Product Officer when you want the CRO).
### Step 3. Score strength
| Score | Tier | Signals |
|---|---|---|
| 1 | Potential | LinkedIn connection only. No other evidence. |
| 2 | Likely | Two or more signals: connected plus emails, or a meeting, or school overlap, or engagement. |
| 3 | Confirmed | Worked together at a small company in the same location and function, or the connector says "I know them well." |
Downgrade rules:
- Same large company but different country, office, or function: drop one tier.
- Overlap more than 8 years ago with no recent interaction: drop one tier.
- Anyone on the suppression list: remove the path entirely.
### Step 4. Prioritize
Priority score = fit (A=3, B=2, C=1) x intent (signals present=2, none=1) x strength (1 to 3). Max 18. Work top down.
### Step 5. Pick the play
| Play | When | What happens |
|---|---|---|
| **Direct intro** | Strength 3, or 2 with recent interaction, to the buyer | Connector forwards your note to the buyer. Highest conversion. |
| **Indirect (one hop)** | Connector knows someone else at the account | Connector sends your note to their contact, who passes it to the buyer. |
| **Mention** | Target used to work with one of your happy customers | You reach out yourself and reference the shared connection. Use sparingly and only with the customer's OK. |
Expect mentions and one-hop paths to drive most of the volume, and direct intros to drive most of the pipeline.
### Step 6. Batch the monthly ask
Once a month per connector. Never weekly. Up to 10 targets, ranked. Ask them to reply with the numbers they know well enough to introduce. Protect their social capital: they stay in control, nothing goes out behind their back.
### Step 7. Draft forwardable notes
For every target the connector says yes to, write a note they can forward without editing.
### Step 8. Track it
In your CRM, set lead source = "Network intro" and record the connector name and play. Review monthly: asks sent, yeses, intros made, meetings, pipeline, closed-won. Retire connectors who never respond; thank the ones who do with results.
## Templates
### Monthly ask to a connector (send yourself, from your own inbox)
```
Subject: 10 intros for [Month], reply with numbers
Hi [Name],
Here are the 10 accounts we'd most like to meet this month where it looks like you may know someone. Reply with the numbers you know well enough to introduce, and I'll send a short note you can forward as-is.
1. [Person], [Title] at [Company] ([how you likely know them])
2. ...
10. ...
No pressure on any of them. Thanks for helping.
[You]
```
### Forwardable note (direct or one hop), under 120 words
```
Subject: [Connector] suggested we connect
Hi [Target first name],
[Connector] thought we should talk. [Company] helps [ICP] [specific outcome], and [one concrete reason it's relevant to them now: hire, funding, launch, initiative].
[One proof point: customer like them plus result.]
Open to 20 minutes in the next two weeks? If the timing's wrong, no worries at all.
[You], [Title], [Company]
```
### Mention (you send, only with the customer's OK)
```
Hi [Target], saw you worked with [Customer contact] at [Old company]. We've been working with [them / their team] at [Current company] on [outcome], and they suggested you might find it useful too. Worth a quick look?
```
### The vouch capture (quarterly, 5 minutes per person)
Give each connector a list of 200 to 300 people from their own network in a relevant function. Ask them to mark the 10 they would personally vouch for. Store the marks as tags. Use them for intros and for recruiting.
## Output format
1. **Path table**: Account | Target contact | Connector | How they know each other | Strength | Play | Priority
2. **Monthly ask emails**, one per connector, ready to send
3. **Forwardable notes**, one per approved target
4. **Tracking sheet columns**: Date asked | Connector | Target | Play | Yes/No | Intro date | Meeting date | Opportunity | Amount | Stage
5. **Gaps**: accounts with no warm path (route these to cold outbound and ads)
## Guardrails
- Always draft; never send on anyone's behalf. Connectors send their own forwards.
- Only use network exports the connector gave you with permission.
- Respect the suppression list in every step.
- Cap asks at 10 per connector per month.
- Treat strength scores as hypotheses. The connector's answer overrides the score.
Credit: Inspired by David Connors, The Swarm (theswarm.com), on the GTM AI Podcast. GTMAIpodcast.comTop 1% Sequia Strategy
---
name: vouch-capture
description: Run the Sequoia "top 1%" exercise. Turns each person's network export into a short review list, captures who they would vouch for, and puts those vouches to work for intros and recruiting.
---
# Vouch Capture
Inspired by David Connors, Co-Founder and CEO of The Swarm, on the GTM AI Podcast. At Sequoia, David's team gave partners and founders a list of 200 to 300 people from their own network and asked them to mark the 10 they'd vouch for. It took 5 minutes per person, and the result was proprietary data nobody else had. This skill runs that exercise for your team. GTMAIpodcast.com
## What this does
1. Turns each person's LinkedIn export into a focused review list of 200 to 300 names in a relevant function.
2. Captures who they would personally vouch for, in one sitting of about 5 minutes.
3. Combines everyone's vouches into a shared list you can use for warm intros, hiring, and referrals.
## What you need
- **Participants**: leaders, investors, advisors, or anyone willing to spend 5 minutes.
- **Each participant's LinkedIn export**: Settings > Data privacy > Get a copy of your data > Connections. Produces `Connections.csv` (First Name, Last Name, URL, Email Address, Company, Position, Connected On).
- **The purpose of this round** (pick one): pipeline (buyers at target accounts), hiring (a specific role), or referrals (people who'd send you customers).
- **Optional**: target account list (for pipeline rounds) or job description (for hiring rounds).
## Process
### Step 1. Define the filter
Write the round's filter in one line before touching data:
- Pipeline: "Directors and above in Revenue, Sales, or RevOps at our tier A and B accounts"
- Hiring: "Account Executives and Sales Managers, 3+ years, at B2B SaaS companies"
- Referrals: "Founders, CROs, and consultants who advise companies in our ICP"
### Step 2. Build each review list
From each participant's `Connections.csv`:
1. Keep rows whose Position and Company match the filter. Use title keywords and, for pipeline rounds, match Company to the target account list (normalize names: drop Inc, LLC, punctuation).
2. Remove anyone on your suppression list.
3. If more than 300 remain, keep the most relevant 300 (tier A accounts first, then seniority, then most recent connection date). If fewer than 50 remain, loosen the filter one notch.
4. Sort alphabetically so the participant scans quickly.
### Step 3. Send the review
One message per participant with their list. Ask for three marks only:
- **V**: "I'd vouch for this person." Limit: 10.
- **K**: "I know them well enough to intro, but wouldn't vouch."
- **X**: "Don't route through me to this person." (Goes to the suppression list, no reason needed.)
Everything unmarked stays unmarked. Do not ask for reasons; speed is the point.
### Step 4. Combine the results
Merge all participants' marks into one table. A person vouched for by two or more participants is a top priority. Record who vouched, the date, and the round's purpose.
### Step 5. Put it to work
- **Pipeline**: V and K marks at target accounts become Confirmed paths in the Warm Intro Engine. V marks get direct intro asks first.
- **Hiring**: V marks get a personal outreach from the participant who vouched, with the role attached.
- **Referrals**: V marks get a thank-you and an ask for one intro.
### Step 6. Refresh
Re-run each quarter or when someone new joins the cap table, advisory board, or leadership team. Vouches older than 18 months get re-confirmed before use.
## Templates
### Review request
```
Subject: 5 minutes: who would you vouch for?
Hi [Name],
Below are [n] people from your network who fit [filter in plain words]. Could you scan it and mark:
V = you'd vouch for them (up to 10)
K = you know them well enough to intro
X = please never route through you to them
Takes about 5 minutes. Nothing goes out without your OK.
[List]
Thanks,
[You]
```
### Vouch-based outreach (for the participant to send)
```
Hi [Target], I'm helping [Company] [find a great Sales Manager / meet a few leaders in your space], and you were one of the first people I thought of. Would you be open to a short call with [Your name]? Happy to share more first.
```
## Output format
1. **Review lists**: one per participant, ready to send
2. **Combined vouch table**: Person | Company | Title | Vouched by | Mark | Round | Date
3. **Top priorities**: people with 2+ vouches, and V marks at tier A accounts
4. **Suppression additions**: all X marks
5. **Next actions**: drafted intro asks or outreach for the top 10
## Guardrails
- Only use exports participants give you, and use them only for this purpose.
- Keep vouch data internal. Never tell a target who marked them, or how.
- X marks are permanent unless the participant removes them.
- Participants approve every message that uses their name.
Credit: Inspired by David Connors, The Swarm (theswarm.com), on the GTM AI Podcast. GTMAIpodcast.com
