Systems on autopilot · open to new projects

I engineer the systems
that fill pipelines.

Cole Bonawitz

I'm Cole, a GTM engineer. I wire Clay, Python and AI agents into outbound that books meetings, and I measure every step so the system gets better each week.

See the systems Get a free outbound teardownLinkedIn
~/systems/swoop-outbound — swoop tutors
running
  1. ClayClay tablessource · 10k cos
  2. Clean + dedupepython · pandas
  3. ICP scoringattrs → tier
  4. ClayEnrich contactsclay · waterfall
  5. HubSpot synccrm · lifecycle
  6. A/B sequenceszapier · 100/var
  7. Reply dashboardreact · segment
Live diagram of the Swoop Tutors outbound engine. Open the system →10k in3k qualified50 meetings
Stack in production
ClayClayApolloApolloHubSpotZapiern8nClaudeOpenAIOpenAIPythonpandasPostgreSQLInstantlyInstantlyHiggsfieldHiggsfieldMetaTikTokLinkedInLinkedInShopifyGoogle SheetsReact
01System output

What the systems produced.

Every number here comes from a system on this page. Click one to see the pipeline behind it.
10,000+
records enriched + deduped
↳ swoop-outbound
5,000+
prospects sequenced
↳ swoop-outbound
15%
cold reply rate
↳ swoop-outbound
50
meetings booked
↳ swoop-outbound
$135K
DTC revenue driven
↳ rize-growth
3M+
organic impressions
↳ rize-growth
02Systems · 05

Systems I've built and run.

Outbound, enrichment, content and lifecycle, across edtech, DTC, hospitality and consumer hardware.
01Swoop Tutors·Outbound
shipped

Outbound engine→

The first outbound engine for a 4-person startup: Clay enrichment, Python cleaning, ICP scoring, HubSpot and Zapier sequencing, and a dashboard that told us who actually converts.

ClayClay
10,000+
records cleaned
15%
cold reply rate
40%
warm reply rate
50
calls booked
02RiZE Maps·Outbound
live · autopilot

Hotel outbound agent→

Race-timed B2B outbound to hotels and resorts in F1 host cities: researched targets, verified contact routes, AI-drafted first touches, compliance gates and a human approval queue. It drafts and never sends on its own.

118
target properties
3
US race markets
14
compliance rules
0
unapproved sends
03RiZE Maps·Content + Paid
live · autopilot

Content + paid growth engine→

Ingests every Instagram reel, scores it against the account's own baseline, classifies the hook, writes a weekly brief, and turns winners into Higgsfield video briefs and Meta ads with hard spend guardrails.

Higgsfield
$135K
lifetime revenue
3M+
organic impressions
$55→$90
average order value
135
reels scored
04Personal / agency·Outbound
building

Founder outreach engine→

A Claude Code + Clay + Python system that scores startups against an ICP, enriches founder contacts, and runs A/B-tested sequences tracked all the way to booked meetings.

ApolloClayInstantly
ICP
agent-scored accounts
A/B
tested sequences
→ mtg
tracked to meetings
Live
in build
05Momnt·Lifecycle
live · autopilot

Retention + discovery loop→

300+ discovery interviews defined the ICP; per-session usage in Firestore plus automated win-back surveys turned a hardware + app product into 85% two-month retention.

300+
discovery interviews
300+
units sold
85%
2-month retention
70%
6-month retention
Open slot · GTM audit

Your company could be next.

I'll map your current outbound (data, tooling, sequences, reply rates) and send back the system I'd build. It's free and takes 30 minutes.

Request a teardown
03About

Engineer by training, operator by habit.

Cole Bonawitz speaking at an eventChattanooga, TN → Durham, NC

I'm a Computer Science student at Duke (minors in Statistics and Philosophy) who'd rather ship a pipeline than talk about one. I founded RiZE Maps, a DTC brand that did $135K in revenue, and I've spent the years since building the machinery behind growth: enrichment, scoring, sequencing, content engines and the dashboards that keep them honest.

I work best with early-stage founders who need outbound that runs without them, built by someone who can write the Python, design the Clay table and get on the sales call.

Education
Duke CS · 3.95 GPA
HackDuke
1st of 100+ teams
Founded
RiZE Maps · $135K
Co-founded
Momnt · 300+ units
Resume (PDF)LinkedIn Follow on LinkedIn
04Method

Every system runs the same loop.

ICP in, meetings out, with data flowing back into targeting.
  1. 01

    Define the ICP

    Turn closed-won patterns and interviews into attributes a machine can score.

  2. 02

    Source + enrich

    Clay tables and waterfalls, then Python to dedupe and normalize.

  3. 03

    Score with agents

    LLM agents research and qualify each account against the ICP.

  4. 04

    Gate

    Compliance rules and human approval before anything sends.

  5. 05

    Sequence

    A/B-tested copy and timing, synced to the CRM.

  6. 06

    Measure + loop

    Reply and meeting data feed back into targeting every week.

05Writing

Notes on modern outbound.

All writing →
Oct 16, 2026Human-in-the-loop by default: how I gate AI-drafted outboundThe agent drafts and a person approves. Here are the guardrails I put between an LLM and a prospect's inbox.Oct 14, 2026Clean before you enrich: the boring step that made our outbound work10,000 Clay records became 3,000 qualified accounts. Most of the lift came from pandas, not prompts.Oct 7, 2026A/B testing cold email when your list is small100 sends per variant per week won't give you statistical significance. Here's how to still learn something.
Replies within 24h

Have a pipeline problem?
Let's build the system.

Tell me your ICP and what you've tried. I'll come back with the workflow I'd build and how I'd measure it.

colebonawitz@gmail.com LinkedIn Message on LinkedIn