The AI growth operating system
A full growth function, run by one person. At its center is an AI Growth Director you steer in plain language. You set direction, it runs the growth work you never get to: content, follow-ups, website updates, newsletters, reporting. It works through modules on your tools, or ones we recommend.
Growth is the job that always waits
At a small company, growth is somebody's second job. It waits behind client work, runs in bursts, and stops right when business picks up.
A growth hire is a full salary before the pipeline earns it. Another tool is another thing to run. So the work stays with the same few people.
Built for whoever owns growth at a startup or SMB
From a founder wearing every hat to a small growth, marketing, or sales team.
Founder
You are the growth department, on top of delivering the work.
Director
Growth, marketing, or sales, and you still run it yourself.
Small team
A few people carrying all of growth at once.
How it works
You direct and approve. An AI director runs your playbooks and works through your tools, or ones we recommend, one capability at a time.
Strategy, positioning, voice rules, and process live as plain files. Any capable AI model can read them.
It boots from those files and holds your whole company in context. You work with it in the AI app you already prefer: Claude, ChatGPT, Mistral, or another.
Workflows that run on their own, on schedule, on tools like Notion, n8n, Metricool, Strapi, and Slack. The director is part of some of them and can run any of them on demand.
Everything reports back. Nothing new to buy. Nothing to migrate.
What you get
One AI Growth Director you steer, and six modules that do the work. The director is where you plan and set direction; each module takes a piece of the growth work off your plate and does it end to end. Start with the one that removes the most work, add more as they earn it.
Growth Director
Directs every moduleThe module you talk to. Discuss plans, test ideas, and set direction in plain language. It writes your decisions into the playbooks every other module runs on.
What's included
- A standing conversation where you already work: Slack, Teams, or email
- Discuss plans, priorities, and new ideas whenever they hit you, in plain language
- Your decisions written into the playbook files every module reads
- Change direction once and the whole system follows, no reconfiguring
- A weekly review of what the system did, with proposals for what to do next
- You stay the director: it plans and proposes, you decide
-
CRM Intelligence
Every new contact captured, researched, and scored on how well they fit. You get a morning digest of who is worth your time today.
See the moduleWhat's included
- New contacts land in your CRM automatically, no manual entry
- Every record enriched with profile and company context
- Each contact scored against your ideal customer profile
- A morning digest delivered where you already work: Slack, Teams, or email
- Your CRM stays the system of record
-
Pipeline Follow-up
People who engage with your posts become contacts with context. You get a shortlist of who to reach out to, and when.
See the moduleWhat's included
- People who engage with your posts captured as contacts, with the context of what they engaged with
- Duplicates matched against contacts you already have
- Engagement logged as signals on each record
- A shortlist of who to reach out to, and when
-
Autonomous Website
An AI developer builds your website changes from a task board and posts a live preview for each one. You approve before it goes live. This is how we ship our own site.
See the moduleWhat's included
- A task board for website changes, filled by you or by the system when your marketing needs a page changed
- An AI developer builds each change in your website's repository
- Every change ships with a live preview you can click through
- Nothing goes live without your approval
- This loop ships every page change on our own site
-
Content Engine
Your completed work mined for post ideas, drafted, and scheduled. A steady rhythm without the blank page. We run it ourselves: 3 posts a week, 38 in Q1 2026.
See the moduleWhat's included
- Your completed work mined for post ideas every week
- Posts drafted in your voice, from playbook files written for your company
- Scheduled on a standing cadence; ours is 3 a week, 38 shipped in Q1 2026
- Performance stats flow back to each post
- Every post approved by you before it publishes
-
Newsletter
A newsletter that assembles itself from what you already published, ready to send on schedule.
See the moduleWhat's included
- Each issue assembled from what you already published
- Drafted and rendered, ready to send
- On a standing schedule; ours goes every two weeks
- You review before anything goes out
-
Finance and Reporting
Client reports drafted and invoices generated on schedule, with revenue and workload forecast from your pipeline. Month-end admin done, and a clear read on where the month is heading.
See the moduleWhat's included
- Client reports drafted on schedule, ready for review
- Invoices generated from the work delivered
- Revenue and workload forecast from your pipeline
- A clear month-end read on where the business is heading
Everything runs in one of two modes
Scheduled work that arrives ready for review, and a director you work with directly on everything in between.
Autonomous, with reports
The workflows run without you: the morning digest, the drafted posts, the newsletter, the reports, the invoices. Every run reports what it did, and the output lands ready for review. You always see what happened.
Deep collaboration
The director holds your whole company in context, so you can work with it directly on whatever comes up between the workflows: a proposal draft before tomorrow's call, prep for a discovery conversation, research on a prospect.
You collaborate on a shared growth task board. Offload the tasks that can be offloaded, keep the ones that need you. Each morning a brief lands in Slack with the day's plate: what is due, what the director finished, what is waiting for your review.
When a task on the board starts repeating, it becomes a new scheduled workflow. The system grows the way your company works.
A week with the system
What it looks like in practice.
- Morning
A digest to read
A digest is waiting in Slack: who is worth your time today, and the day's plate of tasks. Every contact already captured, researched, and scored.
- Midweek
A post to approve
A post is waiting, drafted in your voice from your own completed work. You read it, adjust a line, approve. It publishes on schedule.
- Any day
A preview to click
Marketing needs a page changed, so the task went on the website board on its own. An AI developer built the change. You click the preview and approve. It goes live.
- Month end
A report to check
Client reports are drafted, invoices are generated, and the forecast shows where the month is heading. You review and send.
That is your whole side of the work. The system carries the rest.
We map before we build
The modules are starting points, not a fixed menu.
How growth moves through your company:
- What works
- What stalls
- Where AI takes work off your plate
Then we pick one:
- Deploy a module
- Leave what already works
- Offload a manual workflow to AI
The Growth Mapping Workshop
The mapping is its own engagement: two working sessions with whoever owns growth. You leave with your growth process mapped, the friction ranked by cost, and a clear first move. Yours to act on, with us or without us.
How it's different
The three usual ways to move growth forward each come with a catch. The AI Growth Director doesn't.
Most growth tools are a supporting investment: they hand you data or lists, then leave the operating to you. This system is the operating layer itself. It turns sources into finished work: the morning digest, the drafted follow-up, the scheduled post, the proposal ready for tomorrow's call.
- An agency
Brief it, then wait in their queue.
A retainer every month.
- A hire
Someone to manage and pay.
A salary before the pipeline earns it.
- A tool
It hands you data. You do the work.
One more dashboard to run.
You steer it in plain language, like briefing a person, and it works through your own tools.
- No retainer, no hire, no new dashboard
- It proposes, you approve, it runs the work end to end
- Runs on your tools, not locked to one AI model
You stay in control
The system drafts, scores, reports, and invoices. Nothing goes out without your approval. AI does the work; the judgment stays yours.
Nothing runs in a black box. Every run lands on a timeline in PhantomSwarm, a tool we built and use ourselves, with a link back to the source for every result. It only observes; it never runs your workflows.
We run it on our own growth
This system runs our own growth today. Every post, newsletter, follow-up, report, and invoice we send comes out of it. It is the same operating layer we build for clients, on the everyday tools we already had. The first client deployment is running now.
Marketing
- Post ideas mined from our completed work, weekly scheduled
- Posts drafted in our voice and scheduled, 3 a week scheduled
- Performance stats pulled back to every post, daily scheduled
- Newsletter compiled and rendered, every two weeks scheduled
- Articles drafted for the website on demand
Sales
- New contacts captured, enriched, and scored against our ideal customer profile, overnight scheduled
- People who engage with our posts turned into contacts with context and a follow-up shortlist scheduled
- Weekly prospect sourcing through the same enrich and score pipeline scheduled
- A digest in Slack by 06:00 with who is worth time today scheduled
- Inbox mined into the CRM: interactions logged, inbound leads flagged scheduled
- Outreach messages, proposals, and call prep drafted with the director on demand
Website
- Change tasks filed on a board, by us or by the system scheduled
- An AI developer builds each change and posts a preview we approve scheduled
Finance and ops
- A morning brief in Slack with the day's tasks scheduled
- Client reports drafted and invoices generated on schedule scheduled
- Revenue and workload forecast from the pipeline scheduled
Every scheduled workflow can also be run on demand in the collaboration.
How to start
- 1 Book a 20-minute walkthrough.
Watch the system plan and execute real work, live.
- 2 We map your growth together.
We find what stalls and where AI offloads the most work.
- 3 We build what wins.
A module, or an AI layer on a workflow you already run. Fixed scope, fixed price, in weeks.
Questions founders ask
What does a module cost?
Each module is fixed scope and fixed price. We quote the exact price after mapping your setup, because it depends on the tools you already run.
What commercial terms do you offer?
We shape terms to the company: a scoped project, an ongoing engagement, or a performance-based arrangement where it fits.
What tools do we need to have?
None to start. The system runs on lightweight tools most teams already have. If you run a CRM or a publishing tool, it plugs in and uses them. If you run nothing, we set up only what the work needs.
Which AI model runs it?
Whichever capable model fits your setup: Claude, Gemini, or another. Your playbooks live as plain files any capable model can read, so you are never locked to one vendor.
Where does our data live?
In your tools and your accounts. Your CRM stays the system of record, your website stays in your repository, posts ship from your publishing tool. Nothing moves into a platform of ours.
How long until the first module runs?
Weeks, not months. We map your setup, build on your tools, and the module starts producing.
How much of our time does it take?
A scoping call up front, then approvals. The work arrives ready for review: a digest to read, a post to approve, a report to check.
Do we need technical people?
No. We build the module, run it, and maintain it. You review and approve.
What happens if we stop?
Everything we built stays yours: the playbook files, the workflows on your tools, and your data. There is nothing to migrate off.
Can we start with several modules at once?
You can, but we advise starting with the one that removes the most work. Prove it, then add the next. The mapping call tells us which one that is.
Insights
How the growth system shows up in real work.
