For brands doing $3M to $100M a year
Your growth department, installed.
AI employees. A human growth executive. One team.
Your Slack, from 6:00 AM
- Johnson
- Nova
- Emily
- Carley
growth-dept
Johnson, Nova, Emily and Carley at work. Roles and numbers are examples.
Meet Johnson, Nova, Emily and Carley: AI employees working together in Slack. Paid media, Amazon, creative, email and reporting, with a human growth executive responsible for the numbers.
$800M+ a year in revenue managed across six brands we run ourselves.
Built on brands we run ourselves
$800M+a year in revenue managed.
Six brands, six ad platforms, nine Amazon marketplaces. We built the department for our own brands first and run it for them every day. The demos show how AI employees share reports, propose changes and carry out approved work in Slack.
Lives in your Slack
Connected to Claude
- Runs on a machine you own
- A named growth executive on the hook

Bedding · Meta, Google, Amazon, TikTok Shop creators, email

Swim and activewear · Meta, TikTok

Nutrition · Meta, Amazon

Hotel bedding · Amazon, Google, Meta

Mattress protection · Amazon, Google

Home · Meta, Snapchat, AppLovin
- 6
- brands we run today
- Bedding, home, swim and nutrition
- 6
- ad platforms, every day
- Meta, Google, Amazon, TikTok, Snapchat, AppLovin
- 9
- Amazon marketplaces watched
- US, CA, MX, UK, DE, FR, IT, ES, AU
- 102
- connections running today
- Stores, ad accounts, email tools, spreadsheets
Real ads our AI team staged for these brands, served on Meta.
See every adSwipe to see more.
What we do
A whole team. Ten jobs. Every day.
Your AI employees share your numbers and work across the department. Johnson watches paid media, Nova brings the brief, Emily stages creative and Carley runs email. Here are six jobs they cover.
Johnson · AI employee
Buys your media
Meta, Google, TikTok, Snapchat, AppLovin and Amazon — every ad judged daily, budgets moved to what's working, nothing changed without your OK.
Nova · AI employee
Runs your Amazon
Listings rewritten and kept healthy, ads run daily, best-seller rank and Buy Box watched every hour, restocks planned before you sell out.
- Thu 8:40 · 6 new ads stagedEmily
4 from creators, 2 from your Drive folder · $50/day each
@thecozyhome
@dormlifedani
@sleepwithsarah
@linenlover
Approve all 6Review one by oneEmily · AI employee
Makes and launches ads
New ads every week from your creators, your Drive and your best sellers — named, tagged, budgeted and staged for a one-tap approval.
- @thecozyhome92k2 videos live$4,210
- @dormlifedani310k1 video live$1,880
- @sleepwithsarah184kSample shipped—
Emily · AI employee
Finds creators on TikTok Shop
Your AI team finds creators who already sell in your category, sends the samples, tracks what they post and what it sells, and turns the winners into ads.
Email + SMS share of revenue
Carley · AI employee
Runs your email and SMS
Welcome, browse, cart, post-purchase, win-back and VIP flows built from playbooks that already run for six brands, then tested and tuned for yours.
Which products drove the jump in order size last month?
Three products explain 78% of it. 61% of those orders came from Meta prospecting.
- King bamboo set38%
- Queen duvet bundle27%
- Wrap-around bedskirt13%
- Pillow pair9%
- Everything else13%
Nova · AI employee
Answers anything, in Slack or Claude
“Which products drove the jump in order size last month?” Ask your AI team in Slack or ask Claude on your laptop. Same numbers, same answer, in seconds.
The other four — reports, competitors, your store pages and the daily schedule — are on everything your AI team does.
Lives in Slack
See it the way your team will.
Your AI employees work in the Slack channels you already use. Each proposal has a named owner, the numbers behind it and an approval from your team before anything changes.
Harbor & Pine is a made-up brand. Agent roles and numbers are examples.
reports
Blended return
2.7×
+0.2 vs 7d
Meta
2.9×
+0.4
2.1×
−0.1
Amazon ad cost
11%
of sales · flat
Email + SMS
$5.9k
27% of revenue
- 2 ad sets under your floor
- Queen White sheets · 9 days of stock left
- Welcome flow B: $2.54 per person, +38%
Your numbers, every morning
Your numbers, in the tools your team already opens.
Your AI team posts the numbers that matter in Slack, keeps them in a Google Sheet that refreshes itself, and builds the Monday deck. Nothing to log into.
Revenue · last 30 days
$612k
+11% vs prior 30
Ad spend
$228k
+4%
Blended return
2.7×
+0.2
Email + SMS revenue
$171k
28% of total
Revenue by channel
Monthly, last six months
- Amazon
- Shopify
- TikTok Shop
- Wholesale
Where revenue comes from
Last 30 days
- Paid ads48%
- Email22%
- Organic & returning24%
- SMS6%
| Week | Revenue | Ad spend | Return | New customers |
|---|---|---|---|---|
| Week 31 | $126,400 | $49,100 | 2.6× | 1,640 |
| Week 32 | $131,900 | $51,300 | 2.6× | 1,700 |
| Week 33 | $136,200 | $52,700 | 2.6× | 1,810 |
| Week 34 | $148,050 | $54,400 | 2.7× | 1,930 |
| Last 4 weeks | $542,550 | $207,500 | 2.6× | 7,080 |
Harbor & Pine · weekly review
Best week since Q1. Amazon led, email caught up.
Revenue
$148k
Ad spend
$54k
Return
2.7×
Nova in #reports, Sunday 9:00 PM
The Monday deck is ready. Fourteen slides, three things to decide.
The sheet refreshes itself every hour. The Monday deck is ready Sunday night. Example numbers for a made-up brand.
Creative and creators
One winning video. Every platform. Automatically.
A creator posts on TikTok Shop and it sells. Your AI team notices, pulls the video, and turns it into ads on Meta, TikTok, Snapchat, AppLovin, Google and Amazon — tagged, budgeted and staged for your approval. Winners on one platform become tests on the next.

@thecozyhome
sheets that actually stay cool #tiktokshop
From TikTok Shop
@thecozyhome
posted Tuesday · 92k followers · sample sent by your AI team two weeks ago
Emily · Creative
Cuts and tags each version. Johnson sets the $50 test budget. Both wait for your OK.
To six platforms
Creators found
Your AI team finds creators who already sell in your category on TikTok Shop, sends the offer, ships the sample.
Videos tracked
Every post is watched and every sale credited to the creator who made it — so you know who's worth a second box.
Winners recycled
The best videos become ads everywhere else. Fatigue is spotted early and the next batch is briefed.
Real CGK Linens creative · creator and sales figure are examples.
How did last weekend's sale do compared to Memorial Day, and which products should we feature next time?
Last weekend did $184k across Shopify and Amazon — 22% more than Memorial Day on 9% more ad spend, so it was more efficient, not just bigger.
Revenue
$184k
+22%
Ad spend
$58k
+9%
New customers
2,410
+31%
Feature the bamboo sheet set and the duvet bundle next time — they were 61% of new-customer orders and both still have 40+ days of stock. I'd skip the bedskirt promo; it sells fine at full price.
From your own sales, ad and inventory data · live to the hour
Claude, connected
Your whole business, one question away.
We connect Claude to your numbers. Anyone on your team opens Claude, asks a plain question, and gets an answer from your real sales, ads, inventory, email and reviews. The same numbers your AI team uses in Slack.
- Ask like you'd ask a colleague. No dashboards, no exports.
- Answers come from your own data, live to the hour.
- Charts, sheets and slides on request.
- Claude can look, never change. Changes still go through your AI team and your approval.
Harbor & Pine · built for you · ships day 30 · yours outright
- Built for your brand
- Lives in your office
- Yours, outright
- Keeps running if you leave
To
Harbor & Pine
14 Mercer St, Suite 3
From
Growth Dept
Day 30
Contents: Your AI team, your numbers, every playbook, the manual.
You own it
We build your machine and ship it to your office. It's yours.
Every Growth Dept client gets a dedicated machine, built and configured for their brand. It runs your AI team, holds your numbers and connects to everything. It lives with you, and if you ever move on it keeps running and everything we built stays yours.
Built for you
Configured for your brand, your accounts, your Slack.
Lives in your office
Your data stays on your premises, under your control.
Yours, outright
The machine, your AI team, the data and every playbook are your property from day one.
Keeps running
If you leave, nothing switches off. We hand you the keys and the manual.
Proven, then fitted
Nothing starts from a blank page.
Every email flow, store page, ad structure and Amazon listing comes from a playbook already running for the brands we manage, then fitted to your products, your customers and your voice.

Your first night, sorted
WELCOME10Day 0Your first night, sorted 
Why bamboo
Day 2Why bamboo 
Code ends tonight
ends tonightDay 5Code ends tonight Email & SMS flows
Welcome, browse, cart, post-purchase, win-back and VIP, rebuilt for your brand and tested against each other.

Cooling Bamboo Sheet Set
4.8 · 2,140
$109 $129
TFQKAdd to cartStore pages that convert
Product pages, bundles and reviews laid out the way our best-selling stores do it, then fitted to your catalog.
- $50Thu$150Day 7$400Day 14WinnerDay 21PausedDay 7
Ad testing system
How many ads to launch, how much each gets, when to call a winner, when to pull a loser. A system, not a guess.
- BeforeNo A+AfterA+
Title, bullets and images rewritten with the words that rank and sell.
Amazon listing formula
Titles, bullets and images built from what actually ranks and converts across nine marketplaces.
How it works
In your Slack in a week. Running everything in thirty days.
No long onboarding, no 40-slide kickoff. Four weeks from the first call to a department that runs every day.
Your first 30 days
5 milestones- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- 24
- 25
- 26
- 27
- 28
- 29
- 30
- Day 1The call
- Day 7Your AI team in your Slack · first brief
- Day 14Everything connected
- Day 21Your AI team proposing
- Day 30Machine in your office · weekly review
- Week 1
Step 1: A call, then your AI team joins your Slack
Forty-five minutes on your numbers. We build your machine, your AI team joins your Slack and the first morning brief lands.
- Week 2
Step 2: Everything connected
Stores, marketplaces, ad accounts and email connected. The morning brief, which started the day your AI team joined, now covers all of them.
- Week 3
Step 3: Your AI team starts proposing
Ad changes, restocks, new ads and email tests arrive in Slack for your approval.
- Week 4
Step 4: You set the limits. The team gets to work.
Three numbers per channel. The machine ships to you. The weekly review starts.
Your part: a few sign-ins, three numbers per channel, one signature.
Meet the team
The people behind your department.
Growth Dept brings growth strategy, implementation and day-to-day support together. Your engagement has a named growth executive, with the team behind them.
Holden Russell
CEO
CEO of Growth Dept. Holden built the Growth Dept operating system around the day-to-day work of running ecommerce growth across paid media, Amazon, creative, retention and analytics.
Eli Lazar
Sales and partnerships
Works with brands to understand the business, discuss the scope and find the right fit for Growth Dept.
Matt Block
Engineering and implementation
Works on client setup, connections and ongoing maintenance, and the systems that support each department.
Questions
The ones founders ask on the first call.
Who are the AI employees?
Johnson, Nova, Emily and Carley are AI employees shown across our demos. They work together in Slack on media buying, Amazon, creative, email and reporting, using your own numbers. We set up your team around your brand, with names, responsibilities and approval rules you choose.
How is this different from Viktor or another AI assistant?
Viktor gives you a capable assistant you still have to manage, billed in credits. Growth Dept installs the whole department. AI employees, your numbers in one place, media buying on six platforms, Amazon, creators and email, on a machine we build and ship to you. A human growth executive reviews the numbers with you every week.
Who actually changes my ads?
Your AI team proposes; someone on your team approves. Every change shows up in Slack with before and after, and nothing moves until you tap Approve. Once you trust it, you can let it act inside three numbers you set per channel — a daily budget ceiling, a return floor, a daily scale cap — and it reports every hour. Even then, a new campaign, a price change, a listing edit or a customer email always waits for a tap.
What happens when your AI team makes a bad call?
Every change is logged with its before and after, so any of them can be undone with one message: 'put it back' is enough. Spend already out the door cannot be undone, which is why the ceiling exists: with a $9,000 ceiling the most it can spend on a platform in a day is $9,000, the biggest single move is your daily cap, and an ad set under your floor is paused inside three days. When a person on your team changes something, your AI team leaves it alone for two days. And your growth executive walks through every miss with you on Monday, so the same mistake is not made twice.
Next step
Forty-five minutes with our team. Your AI team in your Slack a week later.
Bring your store URL and the numbers you already trust. You'll leave with a written plan: what your AI team takes on first, what it costs, and when it's in your Slack.
- Johnson
- Nova
- Emily
- Carley
growth-dept
Your department's first week. Roles are examples.











