AI Agents

mkt36z uses a multi-agent architecture where specialized AI agents collaborate to produce high-quality marketing output.

Agent Architecture

Each request is processed by a pipeline of agents. Agents run in sequence or parallel depending on the task. Every agent run is logged to the governance audit trail.

Core Agents

Agent	Role	Used In
`strategist`	Market analysis, positioning, competitive intelligence	analyze, campaign, plan
`content-engine`	Content generation across all formats	generate, campaign, template
`quality-guardian`	QA scoring, brand voice checks, compliance review	generate, campaign
`seo-auditor`	SEO analysis, keyword research, content optimization	analyze seo, generate seo-content
`email-specialist`	Email copy, sequence design, deliverability analysis	generate email, analyze email-health
`competitive-mapper`	Competitive landscape mapping, battle card creation	analyze competitors, generate battle-card

Running Agents Directly

You can run any agent individually for fine-grained control:

# Run the strategist agent
mkt36z agent run strategist "B2B SaaS market entry"

# List all available agents
mkt36z agent list

# Check a previous run
mkt36z agent status <run-id>

Multi-Agent Campaigns

Campaigns orchestrate multiple agents working together. The CLI shows a live dashboard tracking each agent's progress:

mkt36z campaign create "Launch campaign for new AI product"

The dashboard shows real-time status for each agent:

Launch Campaign

> ✓ Strategist           Research complete (2.3s) [QA: 82/100]
  ◎ Content Engine       Generating landing page copy...
  ○ Quality Guardian     Pending
  ○ SEO Auditor          Pending

2/4 complete, 1 running  [j/k] navigate  [enter] expand  [q] quit

QA Scoring

The Quality Guardian agent scores every piece of generated content on a 0-100 scale:

Score	Rating	Meaning
80-100	Excellent	Ready to publish with minimal edits
70-79	Pass	Good quality, minor refinements suggested
Below 70	Needs Improvement	Significant revisions recommended