Eleven agents. One engagement. Research through phone attendant.

The pipeline, stage by stage

Every engagement runs the same sequence. Agent 11 closes it. Operator review happens between every stage.

Tool: Anthropic web search + custom scoring model

Maps the SERP landscape for your vertical and geography. Surfaces keyword clusters, content gaps, and the exact competitor positions worth targeting. Flags which ranking goals are realistic in 90 days versus 12 months.

Operator gate: Research summary reviewed and annotated before any brand work begins.

Tool: Anthropic web search + structured scrapers (BBB, BuildZoom, state license boards, review platforms, trade associations)

Pulls every third-party credential your business actually has. License numbers, accreditation records, platform ratings, trade memberships. Builds the verified-credential inventory that anchors all downstream site copy and trust signals.

*Operator gate: Verification report reviewed.

Tool: Anthropic web search + competitive positioning model

Estimates local and vertical share-of-voice. Identifies the gap between where your business sits today and where the top three competitors are positioned.

Operator gate: Competitive landscape annotated and reviewed before positioning work begins.

No downstream agent runs until it is approved.

Tool: Claude + Cloudflare Pages + wrangler deploy

Pages include home, about, services, service areas, and contact. Deploys to Cloudflare Pages via wrangler, with the layout already tuned for the trust signals and SERP structure specific to your vertical.

Operator gate: Every page reviewed before the domain goes live. Nothing deploys without explicit sign-off.

Tool: Claude + positioning brief

Produces a phased marketing strategy document: channel prioritization, 30/60/90-day action plan, budget ranges by tactic, and KPI targets. Grounded in the SERP data and competitive map from Agents 1 and 2.5, not pulled from a generic template.

Operator gate: Strategy document reviewed with the client before any execution begins.

Tool: Claude + vertical growth playbook

Builds the detailed growth execution plan. Specific keyword targets, content briefs, backlink prospecting lists, Google Business Profile optimization checklist, and a review-solicitation sequence. Output is an actionable playbook the operator delivers directly to the client.

Operator gate: Growth plan reviewed and sequenced before any execution.

Tool: Claude + Meta Graph API (for OAuth-connected accounts)

Generates a populated 30-day social calendar for the platforms relevant to your vertical. Post copy, caption frameworks, hashtag sets, and posting cadence. For Meta-connected accounts, queues posts directly via the Graph API. All others receive a publish-ready export.

Nothing auto-ships.*

Tool: Apollo.io API

Sources a qualified prospect list matched to your ideal customer profile. For B2B service businesses: company size, revenue band, geography, and decision-maker title filters applied. Output is a verified, scrubbed list ready for outreach.

Operator gate: Lead list reviewed before any outbound campaign begins.

Tool: Claude + job-board posting frameworks

For businesses that need to staff alongside the marketing buildout. Generates job descriptions, interview question sets, and candidate scoring rubrics tuned to the vertical. Includes a sourcing strategy for the most effective channels in your trade or service category.

Operator gate: All hiring materials reviewed before any posting goes live.

Tool: Vapi (voice AI platform)

Provisions a trained voice AI attendant for your inbound calls. Handles after-hours inquiries, routes calls, captures lead information, and escalates to a human when the situation calls for it. Trained on your service list, geography, and the most common inbound questions for your vertical.

Operator gate: Script and routing logic reviewed. Test calls completed before the number goes live.

Agent 10: Health Pulse

Tool: Claude + Anthropic web search + analytics connectors

Weekly automated audit of your marketing health. SERP rank movement, review platform activity, site uptime, social engagement delta, and any new competitor moves in your geography. Flags anomalies for operator review before they reach you.

Operator gate: Weekly report reviewed by the operator. Recommendations go to the client only after operator sign-off.

Agent 11: Outreach Manager

Tool: Claude + your own email domain (not ours)

Builds and runs a multi-touch outbound cadence from your own domain. Prospect-specific personalization, human-reply detection, and automatic sequence pausing when a real conversation starts. No blasting. No spoofed senders.

Operator gate: Every sequence reviewed before the first email sends. Human replies route to you, not to an AI.

One deployment, not a parts menu

Some products publish a catalog of four named agents and call it a complete AI marketing stack. This pipeline runs thirteen stages across eleven agents, from the first SERP pull to the weekly health pulse, as a single delivered engagement.

Ten-plus engagements have run this exact pipeline end-to-end across nine verticals.

This is what an AI agent stack for small business marketing looks like when it is built for delivery, not for a feature comparison grid.

The operator gate is not optional

Every agent in this stack has a named stop where the output is reviewed before the next stage opens.

The operator reads every post. Then it publishes. Agent 11 builds the outreach sequence. The operator reads every email. Then it sends. The operator reviews every page. Then it deploys.

Nothing auto-ships. That is not a limitation. That is the product.

The speed of an AI stack, with the editorial accountability of a human reviewer at every stage. That is what the operator model delivers.

See how this becomes one engagement   Book an intro with the operator