A 10-agent AI marketing system that runs campaigns from brief to business impact. Three phases (Intelligence → Execution → Performance), human approval at every gate, brand memory that never drifts. Your team does the thinking. The system does the production and analysis. You keep the margin.
We take one brief from one of your existing clients and run it through the entire Command Center workflow — alongside your normal process. Same brief, same timeline. You see the delta.
Token costs included during pilot. If you convert to a full engagement within 30 days, the €2,500 credits toward the setup fee.
| Item | Detail |
|---|---|
| Brand onboarding | Client brand voice, ICP, banned vocabulary loaded into global_brand.json |
| Connector setup | Notion, Slack, Gmail, Google Calendar wired up for the pilot client |
| Full 3-phase run | Brief → Intelligence → Strategy (PAUSE) → Campaign plan → Creative + Media (PAUSE) → Performance → Report |
| Human checkpoints | 2 review sessions (PAUSE 1 + PAUSE 2) where your team approves before the system proceeds |
| Deliverables | Strategy doc, content calendar, ad copy variants, media plan, performance report, experiment backlog |
| Comparison report | Side-by-side: what your team produced vs what the Command Center produced, with time/cost delta |
Risk for you: near zero. You invest one brief and €2,500. If the output isn't meaningfully better or faster than your current process, we shake hands and you walk away with the deliverables.
Brand voice configured, connectors wired, historical data seeded, channel mix calibrated, team trained on review workflow. Complexity determines where in the range.
Infrastructure (hosting, database, API baseline), system maintenance, model updates, ongoing support. Includes a token allowance covering standard campaign volume.
Transparent pass-through of Claude API token costs with a fixed margin. Heavy months (launches, multi-variant testing) cost more. Light months cost less. You always see the breakdown.
The agency's commercial case isn't about the technology — it's about what it does to your margin.
| Scenario | Without Command Center | With Command Center |
|---|---|---|
| Campaign setup time | 3–5 days (strategist + copywriter + media planner) | 4–8 hours (system + one review session) |
| Copy variants per cycle | 3–5 (what the copywriter has time for) | 15–25 (system generates, human curates) |
| Performance analysis | Weekly manual pull, analyst writes narrative | Automated trend narrative + hypothesis register, analyst reviews |
| Client monthly cost | €8,000–€15,000 (team hours) | €8,000–€15,000 (same price to client) |
| Your delivery cost | €6,000–€12,000 (salaries, overhead) | €2,000–€4,000 (platform + review time) |
| Gross margin | 25–35% | 60–75% |
Your people stop doing production. They start doing curation and client relationships — the parts that actually retain accounts. The Command Center handles the grind.
One client brief, brand guidelines, access to the client's historical campaign data (performance metrics, past creative, audience segments), and one point of contact for the two review sessions.
Connector access (Notion workspace, Slack channel, Google Calendar), client brand files per onboarding, and commitment to the review cadence — the system pauses for human approval, so response time affects throughput.
| Asset | Ownership |
|---|---|
| Campaign outputs | Yours. Every strategy doc, ad variant, media plan, and report belongs to the agency and its clients. |
| Client data | Yours. Never leaves your infrastructure. The system reads it, never stores it beyond the active run. |
| Brand configurations | Yours. The global_brand.json and all client brand files are your IP. |
| The system itself | Kirk & Blackbeard. Agent architecture, skill definitions, orchestration logic, and product roadmap remain K&B intellectual property. |
| Performance data | Shared. Anonymised, aggregated system performance data (not client data) may be used to improve the platform. |
K&B runs the Command Center on your behalf. You send briefs, we run the system, you review at checkpoints. Lowest effort for you. Highest per-client fee.
Best for: agencies testing the model with 1–3 clients before committing team resources.
Your team learns to operate the system. K&B handles setup, training, and ongoing support. You run day-to-day. Lower per-client cost. Your team builds the capability.
Best for: agencies scaling to 5–15 clients who want the competency in-house.
The Command Center runs under your brand. Your clients see your name, your portal, your workflow. K&B provides the infrastructure and ongoing development. Revenue share model.
Best for: agencies building AI into their core proposition at 15+ clients.
Every engagement starts with the pilot. Model A, B, or C is a conversation we have after you've seen the system work on a real brief. The pilot is designed to give you the evidence to make that decision.
Pick one client. Send us the brief. We'll have the pilot scoped and scheduled within 48 hours.
Mark de Kock
mark@brandpwrdmedia.com · +31 65 587 00 78
Kirk & Blackbeard — Captains of Done