Implementation · We build it · Your team runs it

Pick the systems.
We install them in 90 days.
Your team runs them after.

The infrastructure that runs our own consumer brands, installed inside yours. Twenty-plus systems across six capability areas, or a custom build. Each one ships running on your stack, documented to extend, trained to your people.

Start a conversation See the capabilities
01 · Which fork is this

This is the one
where you keep the keys.

You already have the team. What's missing is the machinery.

Implementation is for brands with operators already in seat. We install the systems, run them in shadow beside your existing process until the edge cases surface, train the people who'll own them, and hand over the documentation. Then we leave. You run it, extend it, and own it.

If you'd rather not staff it at all, Growth Partner is the same infrastructure with our team running the day-to-day. Brands move between the two. Picking wrong isn't a thing that happens here.

Every engagement opens with a two-week audit — a diagnostic of your stack ending in a written install plan, sized and ordered. It's folded into the engagement, not billed as a separate project.

02 · Capabilities & skills

Six capability areas.
Twenty-plus skills.

Pick across capabilities or stack within one — whatever fits the leverage points in your business. Each skill is install-ready inside 90 days and ships with documentation. The first call surfaces the right ones for you.

Capability · 01

Brand

Foundation
SKILL

Brand Foundation

Voice rules, brand guidelines, and a copy-rules enforcement skill that flags voice violations on any new copy.
Brand
SKILL

Visual System

UI kits, design tokens (CSS + Tailwind), design-system docs mapped to your stack. Token-driven so the brand stays coherent across web, email, social, decks.
Brand
SKILL

Audience Map

Cohort segmentation using the intimidation framework, validated through your customer data. Each cohort with its dominant fear, approved language, banned language.
Brand
Capability · 02

Marketing

Acquisition + Retention
SKILL

Paid Media

Nightly Meta + Google optimizer skill, tier 1/2/3 ad evaluation, Triple Whale validation, auto-pause guardrails. Replaces 15–30 hours/week of media-buyer time.
Marketing
SKILL

Creative Production

Image and video generation pipelines, batch ad runbook, asset pack templates. Model-agnostic — built to swap providers as the AI stack evolves.
Marketing
SKILL

Comment Marketing

Multi-platform comment outreach plus engagement skill pair. Runs every two hours across Instagram, TikTok, YouTube, Facebook. Nightly dashboard rebuild.
Marketing
SKILL

Email + SMS

Klaviyo flow design, segmentation, campaign cadence, agency-attribution model so reports name owners explicitly.
Marketing
SKILL

Competitive Intel

Sweep skill (Mon/Thu briefing email), threat-tier classification, opportunity signals. Eliminates the "did you see what they launched?" Slack interruption.
Marketing
Capability · 03

Ecom

Conversion
SKILL

Storefront Build

Shopify (or equivalent) build, design system implementation, content/copy review. The skill when the front-end is what's holding back conversion.
Ecom
SKILL

CRO Program

The full Stix testing framework. Ten parallel test slots, progressive traffic allocation (Learning → Signal → Validation → Finalization), Shopify metafield-driven runtime config (no theme push for weight changes), GA4 event structure, weekly auto-report.
Ecom
SKILL

Landing Page Pipeline

SEO + paid LP build process, content brief template, A/B-ready scaffolding. Turns "we need a new LP" from a two-week task into a two-day one.
Ecom
Capability · 04

Operations

Backbone
SKILL

Daily Numbers

Data Integrity Daily skill, source-of-truth hierarchy, P0/P1/P2 alerting. Catches reporting bugs before they reach exec inboxes. The defensive moat under everything else.
Operations
SKILL

Weekly Funnel

Monday 8am pacing report. Anchors to forecast, not last year. Names channel and agency owners explicitly. Replaces the manual report your team builds Sunday night.
Operations
SKILL

Supply Chain

Inventory management, freight monitoring, supplier oversight. Transcript-ingest skill turns weekly supplier calls into structured action items + dashboard.
Operations
SKILL

FP&A

Financial modeling, demand planning, monthly close support. Operator-grade workbook plus a skill that runs daily checks against forecast pacing.
Operations
Capability · 05

Product

Pipeline
SKILL

Concept Validation

Six-pillar evaluation framework, weighted scoring, opportunity sizing. Takes a free-text product brief and produces a decision-ready evaluation.
Product
SKILL

Build & Source

Prototyping, tech packs, supplier collaboration. Gets a new SKU from concept to first prototype without losing weeks to email back-and-forth.
Product
Capability · 06

Custom

Built for you

We built every skill above for our own businesses. We can build new ones for yours.

Custom · 01

The catalog is the menu, not the limit.

If your business has a workflow worth automating that isn't represented here, we'll scope it on the call and build it inside the same 90-day frame.

Custom · 02

Examples we'd consider building.

Review-response (Trustpilot, Yotpo, Google), wholesale order intake, customer-success triage, supplier RFQ pipelines, custom forecast-pacing dashboards, partner outreach engines.

Custom · 03

How custom is scoped.

A net-new system is scoped on the call and sized like any other. Genuinely novel builds — new architecture, new integrations — we scope separately. Documented for handoff either way.

Custom · 04

What we won't build.

Skills that touch financial transactions, healthcare records, or other high-stakes domains where operator credibility doesn't transfer.

03 · Always included

The standards layer.
What keeps the skills working after we leave.

A skill without standards underneath is a one-off workflow that breaks in six months.

Most AI consulting ships a workflow and disappears. Within a quarter, the model API changes shape, the prompt drifts, the team that owned it leaves, and the workflow quietly stops getting used. We've watched it happen at companies we've been customers of.

The standards layer is the layer underneath every skill that lets your team adapt as the AI stack evolves. It's documentation, conventions, tooling, and a wiki — built so a new operator can read it on a flight and pick up where we left off. It's free in every implementation. It's also the moat.

LAYER · 01

The framework library

Eleven canonical docs that govern how every skill we install operates. Batch architecture (collect → filter → batch-generate → execute, never per-item AI calls). Source-of-truth hierarchy (which data system wins when two disagree). Model selection rules. Browser automation best practices. Tier-evaluation logic. The data-validation library. Plus five more. When a new model lands, your team reads the framework and updates skills accordingly.
11 docs
LAYER · 02

Brand & copy standards

Brand context, voice rules, copy-rules enforcement skill — forked from LF canonical and customized to your business. Becomes the source of truth for everything customer-facing: ad copy, landing pages, email, social, support replies.
3 docs + 1 skill
LAYER · 03

Project conventions

Folder structure, file naming, skill authoring standard, engagement cadence, what gets measured weekly. The conventions that let your team write skill N+1 without us. Modeled on the operating systems we use at Stix and Host Modern.
1 doc + scaffolding
LAYER · 04

The wiki

A local-served HTML hub that organizes every doc, framework, dashboard, and skill spec. Pre-rendered from markdown — opens via file://, no server, no internet. Works on a flight. Anyone on your team can browse the entire operating system from one launch point.
Local-first
LAYER · 05

Operating tooling

The CLAUDE.md context file that auto-loads in any AI session your team runs. The wiki build script (markdown → HTML on every doc edit). The approval queue and run log when a skill produces decisions needing human review. The per-skill runbook your ops person reads at 6am when something breaks.
Scripts + state
LAYER · 06

Training

Two ninety-minute sessions with the team owning the work. One operational — how to monitor, when to intervene, what each runbook covers. One editorial — how to update prompts, how to author the next skill, how to extend the framework library when your business changes.
2 sessions

Six layers, free in every implementation. They're the difference between a workflow that runs for a quarter and an operating system that compounds for years.

04 · How a 90-day implementation runs

Twelve weeks.
Same shape every time.

Week 0

Kickoff

Both operators in the room. We agree in writing on what success looks like at day 90, who gives feedback, what's out of scope. Pinned to your wiki by end of day.
1 hour
Weeks 1–2

Read and decide

We read your stack — APIs, dashboards, Slack history, agency reports, the spreadsheet that's actually how the business runs. We interview operators. End of week two: skills confirmed, order set.
Audit
Weeks 3–7

Build and shadow

Each skill ships in shadow mode first — running parallel to manual processes, outputs to a draft inbox. Shadow surfaces edge cases the briefing missed. Most weeks we're rewriting prompts based on what shadow produced.
Build
Weeks 8–10

Live and tune

Skills go live one at a time. Shadow logs run a week after each go-live. Tuning happens in three places: the prompt, the script, the runbook. The runbook is what a future ops person reads at 6am when something breaks.
Launch
Weeks 11–12

Handoff and write-up

Two 90-minute training sessions — one operational, one editorial. Framework library reviewed against your context. Everything committed to your repo. We hand the keys and step out.
Handoff
05 · After the install

Three ways it continues.
Or none.

The install is designed to end. Most clients keep us close for the first few months, then run on their own. Month to month either way. If you'd rather assign it to an internal hire, we'll write the job description and help you interview.

AFTER · 01

Maintain

System upkeep, prompt-drift correction, framework refreshes as our library updates, and a monthly review call. We don't run anything — your team does. The lightest possible touch.
Light
AFTER · 02

Co-operate

Maintain, plus we run selected systems for you while your team gets its footing — the nightly media review, say, or the weekly reporting. One operator embedded part-time, same-business-day Slack.
Shared
AFTER · 03

Hand it over entirely

If the answer turns out to be that you'd rather not run it at all, the install converts into a Growth Partner engagement and we take the day-to-day. Nothing gets rebuilt — it's the same infrastructure, different hands.
Growth Partner
06 · Fit

Best for.
And not for.

Best for

  • Consumer brands (DTC + omnichannel) at $5M–$200M annual revenue.
  • Operators who've tried a one-off AI tool that quietly stopped getting used.
  • Founders or VPs who want a defensible AI strategy their team can run after we leave.
  • Teams already using Shopify, Klaviyo, Triple Whale (or willing to migrate).
  • Brands at a 1.5x+ growth posture — the operating system pays back when volume scales.

Where we'll say no

  • Pre-revenue startups. Optimizing for problems you don't have yet.
  • Manufacturing, healthcare, financial services, regulated industries — operator credibility doesn't transfer.
  • Outsourcing of judgment. Every skill is operator-in-the-loop, not autonomous.
  • Teams with nobody to hand it to. If there's no operator to own it, take Growth Partner instead.
  • Procurement-scale projects. If it needs a committee to approve, we're the wrong firm.

Free 30-minute call.
If we're a fit, you'll know in twenty.

Start a conversation Or see Growth Partner →