Home / Case Studies / GetBoat

GetBoat Travel marketplace SEO · AI agents

An SEO network for GetBoat, built and run by AI agents.

33
expired travel domains restored
526
pages back online
65
AI-written articles

The challenge

GetBoat competes with getmyboat.com for boat and yacht rental searches in many countries. Building topical authority with relevant travel sites by hand means months of manual work per site.

The goal

Build and run a network of relevant travel sites that supports getboat.com in search, with as little manual work as possible.

The approach

01

Restore

33 expired travel domains restored from the Web Archive as clean static sites without scripts: 526 pages back online.

02

Write

65 new articles, each grounded in the original content of its site through retrieval, each with a contextual link to getboat.com.

03

Amplify

A tiered link layer on top (33 second-tier and 33 third-tier placements) and IndexNow submission to Bing, Yandex and Seznam: 31 of 31 hosts accepted.

04

Keep growing

A drop hunter scans newly expired domains every day and filters them for relevance: 212 candidates found so far.

The results

33
sites live
65
articles published
212
candidate domains found

The whole pipeline is scripted: restore, topic planning, article generation, publishing and index checks run as agents, with a staging step before anything goes live. We publish only numbers we measured ourselves; ranking impact on getboat.com is tracked on the client side.

How it was built

Running a network of sites with agents — and the guardrails it needs.

Automating content production at scale is only defensible when each site has a real audience and each page answers a real query. The engineering problem is throughput; the risk problem is quality control, and the second one decides whether the work survives an algorithm update.

Sites with their own purpose

Every property needs a reason to exist for a visitor. Properties built purely to pass authority are the pattern search engines penalise.

Agents for volume, humans for judgement

Research, drafting and publishing scale well. Deciding what deserves a page, and whether a claim is true, does not.

Review before publish

Fact checks, duplicate checks and internal-link placement run as a gate, not as a clean-up afterwards.

Monitoring per property

Indexation, impressions and quality flags are tracked per site, so one problem does not propagate across the network unnoticed.

The figures on this page describe what was built and published in the stated period. Volume alone is not a result — the metric that matters is indexed pages that earn impressions for queries the business cares about.

Lessons

What transfers to other projects.

Automate the pipeline, not the decision about what to publish.
Give every generated page a specific query and a specific reader, or it will be treated as filler.
Build fact checking into the workflow; a model's confident error costs more than the time it saved.
Keep technical footprints clean — shared analytics, identical templates and duplicate content connect properties in ways you did not intend.
Measure indexation and impressions per property, weekly, so problems surface while they are small.
Keep a human owner per property who can answer why any page exists.

Ready when you are

Want a straight line from visibility to revenue?

A free audit shows where your pipeline is hiding — and how to reach it.

or book a call →