How a two-person SEO shop is building an engine to run twelve clients in thirty minutes a month

Overview

Adrian Martinez, who runs the Toronto agency Hire Adrian with his wife, is building an AI delivery system to serve roughly twelve SEO and answer engine optimization clients while keeping the team at two people. The goal is to shrink hands-on work from 10 to 15 hours per client each month down to about 30 minutes, without hiring account managers. He pairs Claude for planning and decisions with Zapier MCP, which connects the agent to WordPress, reporting tools, and image generation so plans turn into real production work.

Key Takeaways

Hire Adrian is a two-person digital marketing shop in Toronto offering website design, SEO, and answer engine optimization (AEO) for about twelve clients.
The bottleneck is time, not demand: each account takes 10 to 15 hours of manual work a month across research, drafts, technical SEO, and reporting.
Martinez built an SEO/AEO engine that runs from a single client intake form through published WordPress content to a monthly client report.
The architecture splits roles clearly: Claude handles planning and context across clients, while Zapier MCP executes the steps inside connected apps.
The target is roughly 30 minutes of hands-on time per client per month, down from 10 to 15 hours, with no added headcount.
Next on the roadmap are ElevenLabs voice agents, wired through the same Zapier path, to answer missed calls for trades clients running Google Ads.

Stats & Key Facts

#About 12 active clients served by a two-person team.
#10 to 15 hours of manual work per client per month under the current process.
#Target of roughly 30 minutes of hands-on time per client per month, a reduction of more than 95 percent from the high end.
#Team size of 2 people: Martinez and his wife, with no plan to add account managers first.
#Dozens of connected apps reachable through Zapier MCP that lack a native Claude connection.

Hire Adrian: a two-person Toronto agency built around SEO and AEO

The story centers on a small agency facing a scaling math problem.

Adrian Martinez runs Hire Adrian, a digital marketing agency in Toronto that focuses on website design, search engine optimization, and answer engine optimization (AEO), the practice of getting a client cited by AI answer tools. He and his wife make up the entire team and deliver work for about twelve clients.

The growth ceiling is time. Each account consumes 10 to 15 hours of hands-on work every month spread across keyword research, content drafts, technical SEO fixes, and reporting. At that rate, the agency runs out of hours long before it runs out of potential clients. Rather than hire a dozen account managers to break that ceiling, Martinez decided to build a delivery system that keeps quality high while the team stays at two people.

How the SEO/AEO engine moves a client from intake to published content

The system follows a repeatable path that captures setup details once and reuses them.

›A seed form collects the basics up front: location, services, positioning, and target service areas, the same fields a kickoff call would gather.
›From that form, the system produces a master research document covering keyword direction, article targets, and overall content strategy.
›A Claude artifact acts as a control surface, working like a lightweight project board for each client.
›Martinez decides which pieces move to draft, which pause for his review, and which run on their own.
›Approved articles flow into WordPress with technical SEO handled inside the pipeline.

Monthly reporting pulls live data instead of static spreadsheets

At month close, the engine assembles a client-facing performance report.

When the month ends, the system gathers site performance, traffic numbers, Google Business Profile activity, and the rest of the campaign picture into one report for the client. Because the report is built from live data rather than spreadsheets updated by hand, Martinez expects the finished system to produce sharper output than manual delivery.

Image work runs through the same setup. For a moving company client, generated visuals place the client logo on trucks and boxes, and each account carries its own visual brief inside the shared hub. This keeps creative assets consistent per client without separate tools.

Why Claude plans and Zapier MCP executes

The design draws a firm line between the thinking layer and the action layer.

Martinez sums up the split as Claude thinks and Zapier MCP executes. Claude is where he plans, iterates, and holds context across all his clients. Zapier MCP is what turns those plans into updates in WordPress, rows in a report, or assets in a client stack. In his framing, the language model is not the integration layer; Zapier is.

He had used Zapier for years on classic automations before connecting Zapier MCP so the agent could reach the apps his clients actually use, including WordPress, reporting sources, image generation, and dozens of tools with no native Claude connection. In his words, he probably will never make another automation by hand, preferring to describe the outcome and parameters to Claude and let it run.

Native MCP versus Zapier as the connection bridge

Martinez uses whichever connection is most reliable for each app.

›When a platform already exposes a solid native MCP, he connects to it directly.
›When a platform lacks one, he reaches for Zapier because the app catalog and trusted connections are already in place.
›He still opens Zapier directly when needed, but most of the work now happens on the agent side.
›He describes the overall experience as very positive and calls Zapier a solid bridge to APIs and MCPs that are not native to Claude.

What comes next: voice agents for missed calls

The SEO engine is one of several builds Martinez has planned.

Many of his clients are trades businesses that run Google Ads to generate phone calls. The problem is that calls go unanswered while crews are out on job sites, which wastes ad spend. Martinez is wiring ElevenLabs voice agents through the same Zapier MCP path so those calls get answered automatically.

The pattern mirrors the content engine: he describes the workflow in Claude, and Zapier MCP carries it into the connected tools. The same architecture that publishes blog posts is being reused to handle live phone leads, which shows how a small team can stack new services on one foundation.

Plain-language takeaway for service businesses

The case is a model for any small services firm weighing AI.

The lesson for a non-technical owner is that the AI model and the automation tool play different roles, and keeping them separate matters. Claude holds the judgment and the running context, while a tool like Zapier MCP does the actual clicking and posting inside the apps you already pay for. You do not have to replace your software stack to get value.

The headline number, going from 10 to 15 hours to a target near 30 minutes per client, is the kind of efficiency a two-person shop needs to compete with larger agencies. It is a target Martinez is building toward rather than a finished result, so the realistic read is a direction of travel, not a guarantee.

Frequently Asked Questions

What is answer engine optimization (AEO)?

AEO is the practice of getting a business cited and recommended inside AI answer tools, not only ranked in traditional search results. Martinez tracks LLM citations as one of the metrics on his agency dashboard.

What is the difference between Claude and Zapier MCP in this setup?

Claude is the planning and decision layer where Martinez holds context across clients and chooses what to publish. Zapier MCP is the execution layer that carries those decisions into connected apps like WordPress, reporting tools, and image generation.

How much time is the agency trying to save per client?

Each client currently takes 10 to 15 hours of hands-on work a month. The target Martinez is building toward is roughly 30 minutes per client per month, a cut of more than 95 percent at the high end, without adding staff.

Does the system publish content automatically?

Partly. Martinez reviews a Claude artifact that works like a project board and chooses which pieces run automatically, which pause for his review, and which move to draft. Approved articles then flow into WordPress with technical SEO handled in the pipeline.

What is the agency planning to build next?

Martinez is wiring ElevenLabs voice agents through the same Zapier MCP path to answer missed phone calls for trades clients who run Google Ads while their crews are on site.

Hire Adrian shows how a two-person agency can stretch capacity by letting Claude plan and Zapier MCP execute, turning a single client form into published content and monthly reports. The 30 minute per client target is still a goal in progress, but the architecture points a clear path for small service businesses adding AI.