AI Roadmap · Media Scaling

What's possible, and where to start.

Built from our conversation. You're already AI-first, so this isn't a pitch to adopt AI. It's the map of where the orchestration gaps are, what they're costing you, and the fastest path to your 3x launch goal.

The bottom line
Your 12-month goal, in your words

"Triple our monthly client launch capacity while saving costs due to increased team productivity."

Media Scaling intake

The fastest path: install the operating system everything runs on (and finally syncs across your team), then build the content-audit funnel you're already moving on, take the busywork off sales and onboarding so launches climb toward your 3x goal, and give your writers a QA co-pilot that protects the craft instead of replacing it.

This is built from one discovery call plus research. It gets sharper once Walt interviews your leaders and we lock the numbers to your real launches.

What we're solving for

Three things, underneath the tools.

Strip away the stack and the tactics, and the roadmap comes down to three moves.

01

Ship faster

Compress the two things gating your launches, scripting and onboarding, so monthly launch capacity climbs without a proportional jump in headcount.

02

Run it as one system

You have the tools, the AI, and now a developer. What's missing is the layer above them, and the one piece nobody has solved: syncing all of it across 140 people.

03

Keep the human edge

Kill the AI slop, not the writers. Protect the craft your brand is built on while taking the busywork off your team.

Current state

Your workflow today.

Four of your core workflows, side by side. Each one breaks at the same kind of place: a manual, by-hand step where time leaks and launches slip.

Funnel build Marketing manager
Claude drafts the HTML Rebuilt element-by-element in GHL by hand Publish

Where it breaks: every funnel is a manual rebuild, so a 2-week idea takes far longer and never ships clean.

Scripting Script writers + editors
Idea / inspiration AI draft Human rewrites the slop Multiple revision rounds QC Publish

Where it breaks: revision cycles pile up, writers fall behind on other clients, and launches slip.

Onboarding Spencer / ops
Sign Content access Scripts Manual team assignment, not by capacity Social setup Tracking + content DB built by hand Launch

Where it breaks: 2 to 4 weeks per launch, and it's the ceiling on how many clients you can take on.

Sales follow-up Sales team
Call ends 12 to 13 field form filled by hand 2 to 3 day data lag CRM updated

Where it breaks: data shows up late or not at all, so the numbers your dashboards depend on are already stale.

Your stack GHLNotionSlackGoogle DriveQuickBooksHyrosClaudeChatGPTLinkAlawareMetricoolSocial ProofPoppy AIFathom
The cost of today

What that is costing you.

Conservative estimates of the hands-on time those manual steps eat, framed for a 140-person team. These are placeholders to be tightened after Walt interviews your leaders, not billed hours.

Capacity you're spending on work the OS can take

~9 full-time roles

of capacity going to work the operating system can carry

≈ 350 hours/week recovered across the team (conservative: ~2.5 hrs/week/person × 140)

Scripting revision tax

~half of editing time

on the #1 bottleneck · recoverable

Funnel rebuilds

days per funnel

rebuilt by hand in GHL → minutes

Sales data lag

2 to 3 days per call

+ manager review time · recoverable

Launch ceiling

2 to 4 weeks per launch

caps new clients · the lever on 3x

The real prize isn't the saving, it's hitting your 3x launch goal without 3x the payroll. These are conservative estimates we confirm with your leaders.

The orchestration gap

More tools, or one system they plug into.

You're already AI-first. The next gain isn't another tool, it's the layer that ties the ones you have together.

The instinct

More tools, more dashboards

Stack another AI tool or build another standalone dashboard. Each one helps, and each one is another island.

vs
What actually scales

One system they all plug into

An operating system that orchestrates your tools, your AI, and your dashboards, and syncs across the whole team.

Your developer is building dashboards. The OS is the thing they plug into.
An OS, not more tools

Four parts and a swappable engine.

Five components, wired into one system that already knows Media Scaling.

The architecture

The layer above your stack, synced across your team.

Not another tool to log into. A system that runs across the platforms you already own, works the same for every person, and gets smarter as it grows.

Bolt AI onto a messy setup and you get a faster mess. The OS is what makes it scale to 140 people instead of one laptop.
01
Instructions

Your brand rules and standards written once, so every workflow pulls in the same direction instead of inventing its own.

02
Workspace + team-sync The piece nobody solved

Three repos and one sync command, so improvements spread across 140 people and nobody overwrites anybody. This is the part nobody has solved for a team your size, and the reason your developer's dashboards become a system instead of islands.

03
Memory

Every meeting and client Slack channel ingested into one second brain, so anyone can ask what's happening on any account and get a real answer.

04
Connections

Wired into GHL, Fathom, Hyros, Link, Alaware, your real stack, so every build runs on true data instead of a sandbox.

05
Dashboard

The whole company on one screen, with buttons instead of the terminal so the whole team can use it, not just your developer.

Future-proof

The engine that runs it (Claude today) is swappable without rebuilding the system. Last year, picking an AI was a strategic decision. Today it's a billing decision. You own the system; the model just plugs in.

What it runs for Media Scaling

What it can already do.

A library tailored to your workflows. Tap any card to see the specifics. We push new skills into your OS every month, so it only grows.

Content & scripting
Script QA co-pilot

Protects the writer, kills the slop

  • Scores every draft against your brand rules + format buckets before it reaches your editor
  • Catches the slop so your writers spend their time on craft, not cleanup. The human stays the author
  • Self-annealing: when a script gets kicked back, the fix gets written in
Reel / video transcript puller

The Poppy replacement

  • Pulls reel and YouTube transcripts straight into the system
  • Feeds voice-training and format analysis inside Claude Code
  • Replaces the standalone Poppy AI step
Voice training

From a creator's channel

  • Learns a creator's voice from their own channel
  • Keeps every sub-account on-brand across 100+ writers
  • Consistency across 24 to 56 sub-accounts per client
Format-bucket planner

Plans by proven format

  • Plans against your format buckets: green screen, reaction, stitch
  • Objection role-play and industry-news formats included
  • Keeps the mix intentional, not random
Hook testing

3 hooks per video

  • Generates three hook variations per video
  • Built to feed your views performance guarantee
  • Stops the team guessing on the most important line
Sales & CRM
Fathom → GHL autofill

Post-call form, filled

  • Fills the 12 to 13 field post-call GHL form from the Fathom transcript
  • Kills the 2 to 3 day data lag
  • Your dashboards finally read off fresh numbers
Near-miss call triage

For the sales manager

  • Analyzes every sales transcript
  • Flags the almost-closed calls worth training on
  • A feedback loop on the revenue you nearly captured
Deep prospect research

Replaces thin notes

  • Real research notes in GHL, not three thin paragraphs
  • Replaces the current OpenRouter output
  • Reps walk in knowing the account
Call-coaching summaries

Sharper reps

  • Turns transcripts into coachable summaries
  • Surfaces patterns across the whole team
  • Training material that writes itself
Onboarding & ops
Capacity-based team assignment

By load, not by who's free

  • Assigns new accounts by real capacity, not who's free in Slack
  • Removes the manual judgment call from every launch
  • One owner handles materially more launches
Auto content-database build

Drive + timestamps + ranking

  • Builds the content DB from Drive automatically
  • Timestamps, clip ranking, and viral-reuse flags
  • Replaces a labor-intensive, perpetual task
Setup scaffolds

Social + tracking dashboards

  • Scaffolds social-account setup per launch
  • Spins up the tracking dashboard automatically
  • Takes days of by-hand setup off the launch path
Launch-status tracker

Nothing sits idle

  • Triggers the next step automatically
  • No launch sits stuck between stages
  • Live view of what's moving and what's stuck
Marketing & funnels
Funnel + landing page builder

On Vercel, in minutes

  • Spins up funnels and landing pages on Vercel
  • Stops the manual element-by-element rebuild in GHL
  • The next funnel is an afternoon, not two weeks
Audit-PDF generator

From a call

  • Turns each audit call into a branded PDF automatically
  • Brand, content-volume gap, hooks-per-video, view benchmarks
  • Revenue-left-on-table calc built in
Coupon / checkout mechanic

The "free with code" anchor

  • The checkout-coupon anchor mechanic
  • Revenue-qualification form into calendar booking
  • The exact flow Walt runs on his own audit funnel
Podcast site

A spin-up, on the same system

  • A podcast site stood up on the same foundation
  • Wired into the OS, not a one-off project
  • Another channel without another vendor
Reporting & dashboards
Live company pulse

One screen

  • The whole business on one live screen
  • No ten tabs and a month-end scramble
  • The same view for Logan, Spencer, and the team
Client-health view

Spot risk early

  • Live client-health across the book
  • Surfaces accounts at risk before they churn
  • Reads off the same source of truth as everything else
Wire in your dev's dashboards

Self-updating

  • Your developer's dashboards plug into the OS
  • Self-updating instead of standing alone
  • We don't rebuild them, we connect them
Ask anything

Across every system

  • Ask in plain language across every connected system
  • "Which launches are stuck and why?" answered live
  • Pulled from every tool at once
Knowledge & team-sync
Meeting ingest

Every call, captured

  • Every call flows into the second brain automatically
  • No one re-explains an account from scratch
  • Searchable across the whole team
Slack ingest per client

Channels into memory

  • Each client's Slack activity captured per account
  • Material items rolled into a triage view
  • Nothing important buried in a busy channel
One sync command

Across the team

  • Push your work to the whole team in one step
  • Checks what everyone else changed first
  • Improvements spread, nobody clobbers anybody
SOP capture

Process, written once

  • Captures how work actually gets done
  • New hires onboard against real SOPs
  • Tribal knowledge stops walking out the door
ClickUp + Claude task automation

Auto-flag missed tasks

  • Tasks flow from meetings and Slack automatically
  • Auto-flags the tasks that slipped through
  • Sequenced after the wedge, never on day one
How we'd build it

You've seen what's possible. Here's what I'd build with you.

You're already AI-first and you've got a developer. So this isn't about adopting AI. It's the operating system that ties it all together, synced across your team, and exactly how we'd put it in your hands.

What you're getting

One system that runs like a department that never sleeps.

Not another tool to log into. A system that already knows Media Scaling, runs across every platform you own, and works the same for every person on the team.

Onboarding

Assigns the team by capacity, builds the content database, and gets the first batch moving so launches happen in days, not weeks.

Scripting co-pilot

Clears the AI slop before it reaches your writers, so they spend their time on craft, not cleanup. The human stays the author.

Sales

Fills the post-call CRM form from the Fathom transcript, and flags the almost-closed calls your manager should be training on.

Marketing

Spins up funnels and landing pages on Vercel in minutes, and turns each audit call into a finished, branded PDF.

Reporting

Wires the dashboards your developer is building into one live brain the whole company reads from.

The builder

Your developer, multiplied. Spins up new tools, agents, and sites on request, on a foundation that's already wired in.

Every one of those is throughput you don't have to hire for.
A day with it

A launch that used to take a month, moving in a day.

Overnight
A new client's content database gets built and timestamped from their Drive, the script co-pilot clears the slop on the first batch against their brand rules, and the near-miss sales calls from yesterday are queued for the manager.
9:00 am
The scripting team opens drafts that already passed QA. They sharpen and approve instead of grinding through revision rounds. Editing time on the batch is cut roughly in half.
11:00 am
A brand signs from the audit funnel. The branded audit PDF was already generated from their call. The new account is assigned by capacity, not by who's free in Slack.
Any time
Anyone asks in Slack, "which launches are stuck and why?" and the answer comes back live, pulled from every system at once, the same for Logan, Spencer, and the team.
The copy-paste, the revision tax, the chase across ten tools. Gone.
What we set up

Everything we build into your system.

Five real builds, wired together into one operating system. The middle one is the piece nobody has solved for a team your size, and the reason your developer's dashboards become a system instead of islands.

01 / 05

Custom system instructions

It runs Media Scaling the way you would, not the way a generic bot guesses.

The problem. You already know this one. Generic AI gives generic answers, which is exactly why every script comes out as the same slop in slightly different words. Without your standards, your formats, and your voice written in, it drifts, and your team spends more time correcting it than using it.

This is the part almost everyone overbuilds. Getting it right is what makes the whole system feel like you.

Your brand rules, written once

Standards and format buckets encoded so the script co-pilot and every workflow pull in the same direction instead of inventing their own.

Slim and sharp on purpose

Most setups bloat until every task drags a wall of text, slower and quietly burning money on every run. We keep the core lean.

Self-annealing

When a script gets kicked back, the fix gets written in, so the panel gets smarter every time instead of repeating the same misses.

Routing, not detail

It knows where everything lives and who owns what, then hands off cleanly across departments.

02 / 05

The workspace & team-sync

The one nobody has solved: all of this, synced across 140 people, without overwrites.

The problem. You said it yourself. The hard part for a team is keeping everyone in sync, so when one person improves something, everyone gets it, and nobody overwrites anyone. Bolt AI onto a messy setup and you get a faster mess: files scattered, every machine on a different version, your developer's dashboards living as islands.

This is the difference between a system that scales to your whole team and a clever setup that only works on one person's laptop.

Three repos, clean separation

Approved skills and agents, live work and sites, and the memory. Nothing collides, nothing gets lost.

One sync command

Push your work to the whole team in one step. It checks what everyone else changed first, so improvements spread and nobody clobbers anybody.

Your developer, leveled up

The dashboards they're building plug into this instead of standing alone, and they learn the architecture to extend it.

Backed up by default

Every change versioned and pushed. Nothing important sits in one spot a bad day can erase.

03 / 05

A living memory

It remembers every client, every decision, so the team never re-explains anything.

The problem. A chatbot forgets you the moment you close it. Every session restarts from zero. That's why your OpenRouter prospect notes are three thin paragraphs and your AI never feels like it actually knows the account.

This is the one thing that turns a pile of AI tools into an operating system.

Every meeting and channel, ingested

Calls and client Slack channels flow into one second brain automatically, so anyone can ask what's happening on any account and get a real answer.

Structured, not a junk drawer

Captured in a shape it can actually find and use later. Raw memory is a trap; the structure is the hard part.

Synthesized over time

Notes distilled into a connected knowledge base that gets smarter as it grows, not heavier.

One source of truth

A single brain the whole team and every tool reads from, so sales, ops, and leadership work off the same facts.

04 / 05

Connected to your stack

It works on your real business, on the tools you already run.

The problem. Most AI lives in a sandbox, disconnected from the tools you actually use. The hard part isn't connecting, it's knowing what each platform can truly do, and most people promise things their tools simply cannot deliver.

It's the difference between a promise and a result.

Meets you where you are

Plugs into GHL, Link, Alaware, Hyros, Fathom, QuickBooks, Notion, and Google, instead of asking the team to move to something new.

We know which are dialed in

Every platform has a ceiling. Knowing which connect cleanly and which need a workaround is the whole game, and we've already mapped it.

The right ones first

Wire the few that unlock the most (the funnel, Fathom, GHL), then expand. Not everything at once.

On real data

Once connected, every new build has something true to run on, and each one after gets easier.

05 / 05

Your command dashboard

The whole company on one screen, and the buttons that run it without the terminal.

The problem. The intimidating part for a 140-person team isn't the power, it's that not everyone wants to live in a terminal. So the capability sits with a few people and never reaches the team.

This is the layer that makes the OS usable by everyone, not just your developer.

The whole business, one screen

Company pulse, client health, workload, and pipeline live, instead of ten tabs and a month-end scramble.

Buttons, not commands

Onboard a client, build a funnel, run an audit, end-of-day wrap, all one click, so the team uses the OS without learning the terminal.

Built around your roles

Each person sees the views and the skills that matter to their seat.

The layer that ties it together

Instructions, memory, and connections all surface here, in one place you own.

How we work together

For a team your size, we'd build the engine with you.

Three ways in. For where Media Scaling is right now, the Department Engine is the one we'd point you at: we build the flagship engine alongside your developer, co-build the wins with your team, and roll it out, all of it built to train your people so you own it. The other two are here if you want to start smaller or go further.

The foundation

The Base Install

We give you the system, you run with it

The OS installed on your three core seats, both founders and your developer, so the people who matter feel it daily.

  • OS + team-sync set up on 3 seats (Logan, Spencer, your developer)
  • Wired to your priority tools: GHL, Fathom, Drive, Hyros
  • The 3-repo + one-command sync, meeting + Slack ingest, and the command dashboard
  • The starter skill set switched on, ready to build from
$11,0003 core seats · extra seats $3K each
★ What we'd recommend

The Department Engine

We build it with your team and train them to own it

The base, plus one flagship engine built alongside your developer, the quick wins co-built live, and the OS rolled out to your people. The full plan is below.

  • Everything in the Base Install
  • One flagship engine built with you and your developer: the Launch Engine or the Sales Engine
  • A handful of core skills co-built live with your team
  • Fast people-onboarding + rollout beyond the core seats
  • Your developer and leaders trained to own and extend it
$30,000over 3 months: $13K, then $8.5K + $8.5K
Optional · longer-term

The Scaling OS

The whole operation, when you're ready

The expansion, once the engine has proven out. Where Media Scaling becomes the AI-native agency in your category.

  • Every priority engine built: launch, sales, content DB, reporting, plus the scripting co-pilot
  • Team-sync OS across all 140 people, wired to your dashboards
  • Full leadership + dev enablement, every build done with your team
  • Embedded partnership through the rollout
$65K–$95Kquoted to scope + embedded retainer
Add-on
The Funnel Sprint  +$5,000. Bolt your two-week priority onto the Base or the Engine: the content-audit funnel built done-with-you on Vercel (form, revenue qualification, checkout-coupon, calendar) plus the AI workflow that turns each audit call into a finished, branded PDF, wired to GHL, Link, and Alaware. You walk away with the system that builds the next funnel in an afternoon.

The OS install and team-sync are flat, no matter how many people are on it. We price by what we build for you, not per head, so a 140-person team isn't penalized for being big.

The Department Engine · the plan

A sprint in month one, then we calibrate.

Not a drop-and-go, and not us building in a corner and tossing it over. The whole point is training: we build it with your developer and your team so you own it, not us. Month one is a focused build sprint. Months two and three are calibration, tuning it on your real work and rolling it out. (Want us to just build it for you instead? We can, as an extra.)

Month 1The Sprint
1

The foundation

Week 1

We get the bones in. The OS and team-sync installed on your core seats, your transcript tools set up (Fathom plus the reel and video puller), meeting and Slack ingest running, and your custom dashboards stood up. The connections that matter, GHL, Fathom, Drive, Hyros, wired in.

The foundation is live, and everything is being captured and synced.

2

Quick wins, co-built with you

Weeks 2–3

We build a few high-value skills live with your team, so they see how it works and own it: the Fathom-to-GHL autofill that kills your post-call form, deeper prospect-research notes, and a script QA co-pilot that clears the slop so your writers stay on the craft. (If you take the Funnel Sprint add-on, we build that here too.)

Your team has real wins in hand, and starts to see how to build.

3

The flagship engine, built with you

Weeks 3–4

We build the one big engine together, with your developer driving and us guiding, so your team learns to build it and owns it. The default is the Launch Engine: compress the 2 to 4 week client launch so you assign by capacity, auto-build the content database, and get the first batch moving in days. The most direct lever on your goal to triple launch capacity. (Prefer the Sales Engine? We build that one with you instead.)

Onboarding stops being the ceiling on how many clients you can take.

Months 2–3Calibration
4

Tune it on real work, roll it out

Months 2–3

The engine meets reality. We tune it on your actual launches and scripts, bring more of the team onto the OS through the dashboard (buttons, not the terminal), and give each person the views and skills for their seat. We stay on calls with you, fixing and sharpening as you use it.

It stops being a demo and becomes how your team actually works.

5

Hand-off & keep it growing

End of month 3

Your developer and leaders are now self-sufficient on the architecture and building their own. Through GrowthHub we keep pushing new skills into your OS every month, so it compounds instead of freezing the day we finish.

Capability you own, that gets better every month.

Month 1 · Sprint
$13,000
foundation + engine built
Month 2 · Calibrate
$8,500
tuned + rolled out
Month 3 · Embed
$8,500
handed off, yours to run

One engagement, split across the three months we are working together. No big lump sum up front.

Built in month one. Dialed in and yours by month three.
Keeping it alive

And it gets better every month.

What you start with is the worst it will ever be. New skills and updates ship into your OS every month, instead of a tool that freezes the day you buy it. Pick the level of hands-on you want.

GrowthHub

$2,500 / mo

We keep your OS current: new skills and drops pushed in every month, plus a regular check-up on what you've got so nothing goes stale. Includes five of your team in the weekly group calls, and a monthly check-in with us. The system keeps getting better instead of freezing.

Embedded Partner

$6,000 / mo

Everything in GrowthHub, plus a dedicated Slack channel and a direct line to us, so we're connected to your real setup, answering and building alongside your team as you go. The peer-to-peer partnership tier.

Most start with The Department Engine + GrowthHub: one engine built with your team, the wins co-built, the OS and team-sync rolled out, your developer trained to extend it, and a system that keeps growing, for $30K over three months plus $2,500/mo.

The Setup Guarantee

If your OS isn't set up and running the way we promised within two weeks of kickoff, we keep working until it is, or your money back. The risk is on our delivery, not on you.

Let's pick the way in and build it with you.

A quick call to choose the first engine, lock the numbers to your real launches and hours, and answer anything open. We can even build a piece of it live on the call.