Marketing Skill Orchestrator
Route one marketing goal through the right sequence of specialist skills
- Difficulty
- Advanced
- Time to result
- ~weeks to results
- Steps
- 7
- Confidence
- 97%
The Marketing Skill Orchestrator acts as a routing and context-management layer above specialist skills. It first inspects the active project and asks for the operator's primary goal. It then proposes a campaign path and invokes the appropriate capabilities in sequence—for example, competitor research, positioning, lead-magnet design, direct-response copy, and email creation. Each stage consumes relevant earlier outputs rather than starting from scratch. Durable findings are saved into a shared brand folder containing positioning, audience, voice, and other context, allowing subsequent skills and future sessions to reuse them. The human remains responsible for selecting directions and approving recommendations, while the orchestrator handles sequencing, handoffs, and context loading. The output is a coordinated campaign rather than a collection of disconnected AI assets.
Origin
Extracted from Marketing Against The Grain, where James Dickerson demonstrated an orchestrator that routed a LinkedIn SaaS project through research, positioning, lead-magnet, copy, and email skills.
Core principles
- 01Begin with the campaign goal
- 02Inspect existing project context
- 03Route work to specialist skills
- 04Pass outputs forward between stages
- 05Persist brand knowledge for reuse
- 06Keep a human decision-maker in the loop
How to run it
- 1
Inspect the Workspace
Examine the active project folder to identify existing assets, files, and campaign context. Use those findings to avoid redundant questions and work.
Pro tip Summarize the detected project before proposing actions.
Watch out Do not infer missing brand facts from a landing page alone.
- 2
Clarify the Primary Goal
Present a concise set of campaign outcomes based on the available context and ask the operator to select one. Convert the selection into a specific deliverable path.
Pro tip Use numbered choices to reduce interaction friction.
Watch out Do not launch every available skill without a goal.
- 3
Plan the Skill Sequence
Select and order specialist skills according to dependencies. Research and positioning should precede offer creation, copywriting, and follow-up messaging.
Pro tip Make each stage consume a named output from the previous stage.
Watch out Running downstream creation before upstream strategy creates inconsistent assets.
- 4
Load Shared Brand Context
Before each specialist skill runs, load the relevant positioning, audience, voice, and learning files from the shared context directory.
Pro tip Load only the context relevant to the current task.
Watch out Stale or contradictory brand files can contaminate every downstream output.
- 5
Execute and Route Outputs
Run each specialist skill and pass its validated findings to the next stage. Pause at meaningful strategic choices so the operator can approve a direction.
Pro tip Recommend a default while preserving the operator's choice.
Watch out Do not confuse autonomous execution with removing human judgment.
- 6
Persist Durable Findings
Save reusable outputs such as market gaps, positioning, audience definitions, and voice guidance in a structured brand folder.
Pro tip Separate durable brand knowledge from disposable drafts.
Watch out Saving every intermediate thought creates noisy context.
- 7
Validate the Campaign Loop
Review the final assets together to ensure the lead magnet, landing page, call to action, and follow-up sequence express the same positioning and promise.
Pro tip Trace the audience journey from first impression to conversion.
Watch out Individually strong assets can still form an incoherent campaign.
In the wild
The orchestrator inspected an existing LinkedIn growth SaaS landing page, researched competitors, identified underexploited positioning, selected a systems-focused angle, proposed lead-magnet concepts, and built an interactive LinkedIn audit. It preserved strategic context in a brand folder for downstream work and future sessions.
→ A landing page concept evolved into a coordinated inbound campaign with differentiated positioning and an interactive lead magnet.
An orchestrator could inspect a consultancy website, research competing firms, establish a differentiated diagnostic angle, design an assessment, rewrite the landing page, and prepare a welcome sequence. Each specialist stage would load the approved positioning before creating its asset.
→ The consultancy receives a coherent acquisition funnel instead of unrelated AI-generated drafts.
Common mistakes
Using Skills as Isolated Prompts
Independent specialist outputs drift when they do not share research, positioning, and brand context.
Skipping Dependency Order
Creating copy or offers before completing market and positioning work forces downstream assets to rely on generic assumptions.
Saving Context Without Structure
A disorganized accumulation of drafts makes it difficult for later skills to identify authoritative brand knowledge.
Is it for you?
Best for
Marketers building multi-stage campaigns that require research, positioning, offers, copy, and follow-up assets.
Not ideal for
Tiny one-off tasks that need only a single prompt or specialist skill.
From the transcript
“The orchestrator skill knows what skills are available for all kinds of different use cases. So it routes you to the skill to get the…”
“So as outputs are created from individual skills, it saves the outputs into a brand folder so that it can be re-referenced every single time.”
“So, first we're going to go and find a differentiated angle uh and research competitors. We're going to take that data and then we're going…”
From the episode
Claude Code: Landing Page to Lead Magnet in 50 Minutes