For in-house marketing teams
Ship the backlog you never had headcount for
Every marketing team has a list of things that would obviously work and never get done. The constraint is not ideas or agreement. It is hands.
What is actually hard about this
The backlog is not a prioritisation problem
You already know which twenty things are worth doing. You are doing four of them because four is what the team can carry, and the other sixteen are not less valuable, just unstaffed.
Agency work arrives without your context
Briefing an agency properly takes most of the time the agency saves. Briefing them badly produces work that is competent and wrong, which takes longer to fix than to redo.
Quality control does not scale with volume
More output only helps if someone reads it. Teams that increase production without changing how review works publish more and learn less.
What changes
The agents named below are the ones that actually do this work, from a fleet of 50.
One context, every channel
Positioning, ICP, competitors and voice live in one place, and the content, email, ads and social agents all draw from it. The consistency problem stops being a policing problem.
Missions for the multi-part work
A launch, a campaign, a content engine. Several agents working in sequence, each seeing what came before, ending in one document rather than nine loose ones.
Review that survives a busy week
Notes pinned to the artefact, unresolved notes carried into the next round, and a recorded decision against every version. Reviewing never costs credits, so there is no incentive to skip it.
Your stack, connected
HubSpot, Attio, Slack and Apollo connect natively, so agents work from your real CRM data and deliver where the team already looks.
A sensible first month
Not a rollout plan. Just the order that tends to work.
- 1Week oneBuild the world and have two people who disagree about positioning review it. Their argument is the most valuable input the system will get.
- 2Week twoTake the three oldest items off the backlog. Not the most important, the oldest, because those are the ones nobody will miss if the first attempt is poor.
- 3Week threeSet a hard cap on how much you will publish per week, below what the system can produce. Volume without review is the failure mode here.
- 4Week fourMove one recurring job onto a schedule: the monthly competitor review, the weekly social plan, the quarterly ICP refresh.
Where most marketing teams start
The Team plan
Five seats and five worlds, which covers a core team plus the sub-brands and product lines that always turn out to exist.
15,000 credits a month · 5 seats · 5 worlds. Your first world is free and needs no card.
Questions this raises
Does this replace our agency?
It replaces the part of agency work that is production against a clear brief. It does not replace the part that is genuine outside judgement, and if your agency is giving you that, keep them and give them more time for it by taking the production off their plate.
How do we stop this producing a pile of unread documents?
Cap what you publish below what you can produce, and treat the send-back rate as a health metric. If nothing is ever sent back, review has stopped happening and volume is about to become a problem.
Can different people own different areas?
Yes. Seats have roles, and there is a client role for reviewers who should see the Dock and nothing else. Chats are private by default; the Vault is shared.