← All work
Growth AI Product Outbuzz AI

AI Storefronts for Agencies

A reusable system I designed and built across multiple agencies: instant, branded AI storefronts: prompt-chained research and templated generation that turn a prospect's own website into a tailored plan in 60 seconds. Outbuzz AI is the example below. No existing ICP, no targeting playbook going in: 2 contracts in 1.5 months from cold outbound coming out.

Company
Outbuzz AI
Role
GTM / Growth Engineer
Stack
Apollo · Clay · Gemini · PhantomBuster · SmartLead
Not a one-off build
I designed this once, as a system. Then deployed it for multiple agencies.
Before
✕ Generic "Contact us" form
  • 24–48 hr response time
  • 3% avg conversion
  • Disappears into a queue
  • Same generic reply for every visitor
After
✓ Self-serve AI storefront
  • Instant value = instant engagement
  • 15% avg conversion
  • Tailored insights based on their own data
  • Clear next steps, not a generic pitch
  • A CTA built to convert
  • Personalized nurture from there: AI-powered emails based on their company context
Case in point

The Outbuzz AI Example

Real funnel, real numbers: measured end to end at one of the agencies this system was built for, from first LinkedIn touch to signed contract.

Outbound funnel · actual results · 1.5 month campaign
540 contacted
Baseline
205 accepted
38%
56 positive replies
27% of accepted
13 meetings
2.4% overall
2 signed contracts
in 1.5 months
01 · The insight

No ICP, so I built one from the client's own wins

There was no predefined targeting, no messaging playbook, no list to start from. Instead of guessing, I looked at the client's historical wins: every strong project was connected to conferences. That became the buying signal, not a demographic assumption.

The core insight
The conference is the buying trigger
There was no predefined targeting, no messaging playbook, no list to start from. Instead of guessing, I looked at the client's historical wins: every strong project was connected to conferences. Companies attending a major event have a concrete, time-bound marketing moment where OOH becomes strategically relevant. That's the trigger, not "who buys OOH" in general.
Why it worked
Relevance did the heavy lifting. The message landed when the prospect was already thinking about their conference presence, not cold, not random, not generic.
The formula
Conference (the trigger)
+
Company fit: 50–200 employees
+
Timing: inside the event window
+
Contact: owns budget or execution
Buying signal detected
02 · The AI Storefront

Every channel pointed at the same storefront.

This wasn't a generic outbound campaign. Paid, organic, community, and 1:1 outreach all funneled attention into one thing: an AI storefront that turns a prospect's own website into a personalized OOH plan in 60 seconds.

Get Your Conference OOH Plan in 60 Seconds
1Visitor enters email + website, no form, no call
2AI researches the company and matches with upcoming conferences
3Branded OOH campaign plan generated as a PDF, under 60 seconds
Channels driving traffic to the storefront
📣
Paid
Conference ads
🌐
Organic
LinkedIn
💬
Community
Slack · Discord
✉️
Outbound
Proposal-first
One AI Storefront
60-Second Personalized OOH Plan
03 · The AI workflow

AI personalized the offer, not just the message

The interesting part wasn't using AI to write messages. It was using AI to turn company research into a tangible, company-specific OOH campaign concept, in clearly separated stages.

Research → Strategy → Proposal Engine
Research · Interpret · Generate · Personalize
< 60 seconds
🔍
Step 01
Company research
What they do, their positioning, the event they're attending, likely audience, relevant public context.
🧠
Step 02
Strategic interpretation
Gemini reasons from structured research + a defined strategy framework to generate a company-specific OOH strategy.
📄
Step 03
Proposal generation
Branded PDF via a pre-designed template: campaign concept, placement rationale, expected outcome.
✉️
Step 04
Message generation
Same research context feeds the LinkedIn outreach so message and proposal stay consistent.
🚀
Output
Personalized proposal sent
The proposal IS the value prop. Show the campaign, don't pitch OOH.
The key insight: "I wasn't using AI to personalize a sentence. I was using AI to personalize the actual offer."
04 · Outreach sequence

Four touches, timed around the event

Every message was built around event timing, not a generic pitch. The LinkedIn message's job was to create enough curiosity to open the proposal, not to explain OOH.

Touch 1
Relevant trigger
Establish immediately why this company, why now.
"I saw you're attending [conference]…"
Touch 2
Contextual idea
A relevant thought on how they could use OOH around the event. Give them something to think about, not just a cold ask.
Touch 3
Personalized proposal
The branded, company-specific OOH plan as the strongest value touch. Their actual campaign concept, not a generic deck.
Touch 4
Soft CTA
Short and conversational throughout. Create curiosity, not pressure.
"Worth a quick conversation?"
05 · Learnings

What I'd take to the next campaign

Historical data > assumptions

The ICP came from the client's own data, not from a "who buys OOH" assumption. Strong outbound starts with finding the observable signal that tells you why this account might care now.

Offer tangibility drove conversion

The 27% positive reply rate came from prospects receiving something specific and useful, their own campaign concept, not a generic "let's explore OOH" message. The proposal was the conversion mechanism.

Full-funnel compounds

Running paid, organic, community, and outbound simultaneously meant prospects often encountered the brand from multiple angles before a cold LinkedIn message arrived: warming the cold.

Timing is the unlock

The same message, sent to the same person, three months before or after the conference would have performed far worse. The event created urgency the pitch didn't have to manufacture.