How Aptible grew organic qualified signups 3.4× after publishing content with Protocol

The situation

Aptible sells infrastructure to startups that have to prove HIPAA compliance.

Technical founders and early engineers are Aptible’s primary users, and the primary growth channel is organic search leading to a self-service, product-led growth signup flow.

HIPAA content requires real technical and security expertise. Aptible operates as a lean team, so that expertise was concentrated in a few engineering and security leaders. Producing new organic content was slow, expensive, and frustrating.

“I couldn’t publish as fast as I wanted to, at the quality I knew we needed to grow traffic and earn trust with technical founders. I’d used ChatGPT and Perplexity and written longer and longer prompts, and still got mediocre content every time.” Kelsey Ellis, Senior Product Marketing Manager at Aptible

First, we wrote the company down

Before building anything, we gathered the documents, calls, and operating rules that shaped Aptible’s content, and turned them into files a machine could read:

• Product messaging organized into a clear architecture, including a detailed ICP definition
• Customer call transcripts, support tickets, and product docs
• Internal subject matter interviews with the engineering and security leaders
• Previously published content, and the workflows for producing more of it
• Voice, tone, and anti-slop rules
• Traffic, signup, and conversion data from the data warehouse

That is the part nobody can download. Aptible’s judgment about what a good HIPAA guide looks like, which claims their security team would stand behind, and which proof they were cleared to name lived in about four people’s heads. Writing it down was most of the work.

Then we built the agents

An agent is a repeatable job with its data, instructions, checks, and review points attached. Each one declares what it needs before it runs:

id: hipaa-guide
produces: guide
channel: public
gathers:
  - messaging/components
  - definitions/icp
  - proof/security-interviews
  - voice/aptible-editorial
  - sources/ahrefs
constraints:
  - policies/banned-phrases
review: kelsey

Because the requirements are declared rather than remembered, the work does not depend on whoever is at the keyboard. If a required file is missing, the run stops and says so instead of inventing something that reads well.

We connected Ahrefs so the agent could compare live keyword research against Aptible’s approved context, suggesting improvements to existing pages and proposing new ones for real search opportunities.

The sources and rules became reusable context. Kelsey or another team member could ask Claude Desktop, ChatGPT, or Claude Code to:

• Find 10 easy keyword opportunities Aptible didn’t rank for that were likely to attract its ICP
• Draft a guide from the last 15 support tickets on a topic, including the CISO’s expert opinion
• Create a changelog entry for the feature in the latest PRD
• Identify which content pages appeared most often in signup paths over the last 30 days

Protocol kept the context and workflows current in every team member’s AI tools.

The second agent took a fraction of the first

The organic content agent came first. The product content agent came after it — and it started from the same messaging architecture, the same ICP definition, the same proof library, and the same voice rules.

Nobody re-explained the company. That is the whole argument for writing the context down before building anything: the first agent pays for the foundation, and every agent after it starts from a foundation that already exists.

What happened

Over the next few months, Aptible used Claude Code connected to Protocol to write new organic and product-page content.

Traffic to the revamped HIPAA guide grew 5.3× over the quarter. Two newly launched HIPAA AI guides became Aptible’s highest traffic drivers.

Organic qualified signups grew 3.4× at the same time.

After the homepage and product page content was updated, the share of signups that qualified rose 2.8×, and 15% of those qualified signups were attributable to the new guide pages.

How it runs now

Aptible continues to report record levels, attributed to the organic and product content agents powered by the OCL. Traffic and signups continue to build.

Strong performance continues. In the five months following the launch of the marketing site Aptible drafted from its Protocol context layer, visitor to qualified signup conversion ran at 3.8× its historical average.

Aptible’s own growth accelerated over the same period. The company passed its entire 2025 new-customer total in July, with five months of the year still to run.