Inside the Harness

Scope wins
over generality.

A general “marketing agent” is too vague. Vague agents just make vague output faster. Scoped agents — with their own tools, voice rules, approval gates, and a clear definition of done — get very good. This is what’s inside a harness.

SEE THE PLATFORM →
∴ Why Scope Wins

Scoped agents get very good.

Each harness is purpose-built for one domain.

Avoice rules · approval gates
Lifecycle email agent

Knows the right campaigns for the right segments. Writes inside enforced voice rules, with examples of what good looks like — and nothing sends without passing an approval gate.

right campaignsvoice rulesapproval gates→ gets very good

Not one giant agent — a company of scoped agents, each earning their role.

∫02 · Inside the Harness

One brain, four layers of depth.

When you look at this hierarchy you’re looking inside a harness. Here’s how it works inside a marketing agency.

INPUTS THAT FEED THE BRAIN
transcripts · chats · past campaigns · client learnings · strategy docs · workflows · examples of what good looks like
COMPANY BRAIN
everything we know, in one place. maintained by a human champion + the orchestrator.
synced via universal MCP · Claude Desktop · mobile · Code · ChatGPT
ORCHESTRATOR AGENT
routes work · reads the brain
content
lifecycle email
technical SEO
paid ads
department verticals — each one drills down the same way
SPECIALIST AGENT
own tools · voice rules · gates
e.g. blog writer · content research · repurposing
SCOPED SUB-AGENT
one narrow job, one clear definition of done
the narrower, the easier to tune
∫ Verticals

Same structure, every department.

Each vertical drills down the same way: orchestrator, specialists, sub-agents. Only the tools and rules change.

Content
Blog writer
drafts from briefs, matches brand voice
Content research
finds sources, extracts insights
Repurposing
long-form into social, email, clips
Lifecycle Email
Campaign builder
right sequence for right segment
Copy specialist
voice rules, A/B variants
Send-time optimizer
maximize open rates
Technical SEO
Site auditor
crawls, flags, prioritizes by impact
Schema markup
generates and validates structured data
Internal linking
suggest contextual links
Paid Ads
Budget allocator
distributes spend across channels
Ad copy agent
brand + platform constraints
Performance tracker
flag underperformers daily
∫ Specialists

Not a chatbot with a system prompt.

A specialist connects to only the tools it needs. Its voice is enforced, not suggested. Approval gates keep the human in the loop.

Own Tools

The blog writer sees your CMS and content calendar — not your billing system. Each specialist’s toolset is wired through the universal MCP.

Voice Rules

The lifecycle email agent writes in your brand voice because those rules are baked in, with examples of what good looks like. An SEO agent is terse and data-driven. Different voice, same architecture.

Approval Gates

The blog writer can draft 20 posts, but nothing publishes without human approval. Gates define where the human stays in the loop.

Configured during Phase ∫02. Once set, the specialist runs continuously.

∫ The Brain

The brain compounds.

Not a document dump. Every record is typed (facts, preferences, skills, sessions), carries its source, and is segmented by function — structured memory that agents can traverse like a knowledge graph. Synced through the universal MCP to Claude Desktop, mobile, Code, and ChatGPT.

transcripts from client callschats across Slack, Teams, emailpast campaigns and performance dataclient learnings and feedbackstrategy docs and quarterly plansinternal workflows and SOPsexamples of what good looks like

The agent at six months is dramatically better than day one.

∫ Humans + Agents

Every segment has a champion.

The brain is segmented by function — marketing, ops, finance, legal. Each segment is curated by its own human champion. The orchestrator reads the whole brain and escalates each segment’s ambiguity to the right champion.

THE CHAMPIONS · ONE PER SEGMENT
Each curates their slice of the brain
Reviews and approves learnings in their domain
Sets priorities for their vertical
Catches when their segment drifts from reality
THE ORCHESTRATOR AGENT
Reads the whole brain continuously
Routes incoming work to the right vertical
Escalates ambiguity to that segment’s champion
Logs decisions for the brain to learn from

This isn’t one hero maintaining everything. It’s domain owners, each championing their slice, with an agent company working for them.

∫ Your Pod

Your own agent company.

Your pod is a harness deployment — stood up on your cloud, in your own namespace of the universal MCP. Your brain, your data, your approvals. Completely isolated from every other deployment.

YOUR POD · ISOLATION BOUNDARY
your brain · segmented by function
your orchestrator
your specialist agents
your workflows · approvals · memory
your cloud · your MCP namespace
AGENCY CAPABILITY LIBRARY
Proven harness templates, trained patterns, quality gates — battle-tested across deployments. Your pod reaches back for capabilities when it needs them.
templates & skills flow in · your data never flows out

Isolation where it matters. Leverage where it helps.

∫ Scale

Build one vertical well. Fork it.

You start at ~75%, not from zero. The remaining 25% — your voice, your tools, your brain — that’s the Deploy phase.

Client 1
built from zero
Client 2
forked at 75%
Client 10
mostly config

What gets forked is the architecture — the agent hierarchy, the quality gates, the proven patterns. Your brain, your data, and your memory never leave your pod. That’s why the tenth client is dramatically more profitable than the first — and why being one of them costs less, not more.

∫ Beyond

Same architecture, any industry.

The marketing agency is one harness. The same four-layer architecture works anywhere work can be scoped.

Legal ops
Contract review, clause extraction, risk-flagging
Healthcare ops
Scheduling, insurance verification, prior-auth
Customer support
Ticket triage, response drafting, SLA tracking
Financial services
Compliance monitoring, reporting, anomaly detection

Different verticals. Same architecture. One brain. Scoped agents. Isolated pods.

See how we deploy it →

Get Started

Request an architecture review.

Tell us where you'd start, which clients your team uses, and whether you need private inference.