How Proxidize built an AI-powered content workflow to grow its AI search visibility
Summary
- Proxidize provides cloud-managed mobile and residential proxies for technical teams. As a bootstrapped business competing with larger providers, it relies on inbound growth to reach buyers through content and AI search.
- A manual content process and a restrictive WordPress setup slowed the team down. Moving to Prismic let Proxidize connect its website to the AI tools the team was already using and build a custom workflow combining search data, AI research, and writing tools.
- In roughly three months, Proxidize cut article production time from weeks to hours and went from zero to a peak visibility of 22.8% in tracked AI answers. The team also attributed 35-40% of its July and August 2026 signups to AI search.
The challenge: growing content output with a small team
Proxidize grows entirely through inbound marketing. With no outbound sales team, search, and increasingly AI search, is how new customers discover the company.
Their technical buyers research proxies, web scraping, and related topics before choosing a provider. As the business expanded beyond hardware into cloud-managed services, the team also needed its content to explain what Proxidize offers today. That puts a lot of weight on content.
Getting content live involved several stages of manual work. The team researched keywords, studied competitors, wrote and edited articles, then prepared everything for publication.
Zeid Abughazaleh, content writer at Proxidize, estimates that a straightforward article could take around a week from topic selection to publication; a more complex piece could take two. And uploading a finished article into their WordPress setup could take another 20 to 30 minutes: copying in the text, adjusting formatting, uploading images, and setting up internal linking.
With more content topics to explore and a small team to cover them, Proxidize needed a faster path from identifying an opportunity to publishing useful content.
For Yazan Sharawi, growth engineer at Proxidize, the challenge extended beyond getting articles live. As Proxidize’s offering evolved, he wanted to build custom landing pages, pricing pages, and interactive features, but the existing WordPress themes and architecture made those changes difficult to implement. Content management and website development were also closely connected, and the team wanted writers to publish independently while engineers worked on the site.
The solution: connecting AI to content in Prismic
Proxidize moved its website to Next.js and chose Prismic to manage its content. This gave Yazan more control over the website’s design and functionality, while allowing the content team to create and publish pages independently of engineering.
Yazan set up page types for blog articles, research, and legal documents, giving writers a structure they could work with on their own. They could create pages, reorder sections, and update content without asking a developer for help.
Prismic’s MCP server was a key reason the team chose the platform. It made it possible for Yazan to connect the AI tools the team was already using directly to its website content, so an article could move from research to a page ready for review without someone having to assemble it manually. For Zeid, content now arrives in Prismic ready for him to review, refine, and publish.
“Having Prismic connected to our internal tools and AI agents allows us to generate, update, and publish content programmatically.”
Search data helps the team decide what to write
Yazan brought Proxidize’s research and performance data tools into one workspace to help the team decide what content to create or improve. The workspace brings together several sources of context:
- Peec AI shows where Proxidize appears in AI-generated answers and helps identify content gaps. It also reveals the “fan-out” queries AI models run before answering, giving the team ideas for content they might otherwise have missed.
- Ahrefs shows what people search for and which topics competitors cover. The team uses these insights to assess demand and find opportunities to cover topics missing from its own website.
- Google Search Console shows which articles appear in Google search results and attract clicks. The data reveals what’s performing well and where pages are being seen but aren’t attracting visitors.
- PostHog tracks where website visitors come from and what they do once they arrive. Beyond search visibility, it offers a view of how people engage with the content and move toward signing up.
Together, these insights give the team a clearer starting point: a question worth answering, a topic to explore, or an older article that needs attention. From the same workspace, they can then ask their tool to research the topic or help improve the existing content.
AI helps with research and writing inside a connected workflow
Once the team has chosen a topic, ChatGPT and Claude help research it and turn the findings into a draft. They work from Proxidize’s editorial guidelines, with instructions to find supporting sources and follow the team’s writing style. Proxidize’s custom tool connects to Prismic through its MCP server, so content can move into the website’s content structure without the same manual transfer work.
Internal linking shows the practical difference. Writers previously had to find relevant articles and insert links individually. The workflow now identifies related articles already in Prismic and includes links in the draft. The Proxidize team still checks that the links make sense, but much of the repetitive work is already done.
People stay in control because nothing gets published unreviewed. Zeid carefully and critically reviews every piece of content before it gets published. He also uses additional AI checks to flag potential factual errors. Zeid gave the tool examples from his past articles and more than 200 editorial rules covering tone, word choice, and the Oxford comma. And he continues to refine those instructions.
“I never tell our tool to write an article and then immediately publish it. I always want to make sure that whatever is written is still me.”
The editorial care reflects a wider priority for Proxidize. As Yazan puts it, “we’re building an honest brand.” The team wants its content to help technical buyers make informed decisions and trust the company behind it.
The results: faster publishing and a measurable presence in AI search
With research, writing, and publishing connected, Proxidize’s team spends less time on repetitive content tasks and more time on strategy and putting ideas into practice. In roughly three months, the team reported:
- AI search visibility from zero to a peak of 22.8%: the percentage shows how often Proxidize appears in the AI responses Proxidize tracks.
- Around 35-40% of its signups in July and August 2026 attributed to AI search tools.
- Article time down from one to two weeks to one to two hours: research, drafting, and linking that used to fill a week now fit into a morning.
- Roughly 3x the monthly output, with no new hires: the team went from about 10 to 15 articles a month to around 40.
- Around 200 articles created or substantially rewritten, covering both new topics and improvements to existing content.
The Proxidize team is still learning what earns them a place in AI search answers. The prompt set Proxidize tracks evolves as the team targets more commercial searches, so figures fluctuate and don’t represent a fixed comparison over time.
What has changed in the team’s day-to-day work is more tangible. With less time spent preparing pages manually, there is more room to revisit underperforming content, strengthen internal links, and explore new content ideas. It also leaves more freedom to test landing pages and content structures as Proxidize develops its cloud offering. The team can spend more time deciding what to improve next, and putting those ideas into practice.
“It’s really saved us an ungodly amount of time.”
Bonus tip from Proxidize
Bring an engineer into your growth team and start with the tasks that slow people down. For Proxidize, having Yazan work alongside the content team meant he could build around their everyday needs. His advice is to build on an existing open-source tool where possible, then adapt it to your workflow. Involve writers from the start so they can shape the editorial rules and stay in control of what gets published.