Get Found in AI Search Results

Coordinating international SEO on a 1M-page marketplace

Malt, SEO Manager, June 2023 to January 2026. Approximately 1 million pages across all domains, 6 markets.

Starting point

The platform had solid SEO foundations and an agency had delivered a roadmap of both technical and on-page recommendations. But nobody owned SEO internally and there were limited resources available to implement changes. Much of my initial role was aligning product, web, engineering and content teams around fixing technical debt and improving key pages.

What I did

• With no development resourcing, I optimized the targeted templates by hand with colleagues helping on localization.
• I worked with the website team to redesign the templates, until we secured a developer.
• Ran an internal linking project connecting the main templates to blog content clusters, the header and the footer.

Result

Over six months, targeted templates grew by +17% YoY clicks in France, +18% clicks in Spain and +23% clicks in the UK. These figures apply to the optimized templates, not the whole site.

YoY clicks on optimized templates

What I learned

Resourcing was the main constraint. Where I lacked writing capacity I used AI to help produce content, and implemented it in the CMS by hand. I used a human-in-the-loop strategy for revision, recruiting help from international teammates to review text and revise prompts for better outputs. That process led directly to the next project.

Scaling long-tail SEO with programmatic pages

Malt, SEO Manager, June 2023 to January 2026. Approximately 1 million pages across all domains, 6 markets.

Starting point

Optimizing core templates was successful, but the process was not scalable. Led by Malt’s Head of Growth, we looked for ways to automate and expand production of new pages with programmatic content. Malt’s internal search engine shows what clients actually look for. Using that data to decide which pages to create could capture long-tail demand our existing pages missed.

What I did

I worked with engineers and helped design the specification: on-page requirements for each page type, schema markup for each template, and the rules defining when a page should exist, to avoid thin or redundant pages at scale.

Result

The result was 4,000+ pages created and optimized, and a content automation process that kept running after launch.

What I learned

Doing the first project manually meant I knew the templates well enough to assist with a precise spec for product and engineering teams. At this scale the main risk is thin or duplicate pages, so the creation rules matter most.

Paying down technical SEO debt across a multi-market marketplace

Malt, SEO Manager, June 2023 to January 2026. Approximately 1 million pages across all domains, 6 markets.

Starting point

Technical issues had built up over a long time with little progress. I did not have the engineering background to solve them alone, so my job was to find the problems, scope them, and agree on the right fix with engineers.

What I did

I ran regular Screaming Frog crawls, triaged the errors and took each to a developer to diagnose the cause. Over my tenure, 50+ optimizations shipped. Three of the most important:

• Crawl waste. Google kept trying to reach old URLs we no longer wanted crawled. They were blocked in robots.txt, so Google could not read any instructions on the pages and kept remembering them. We unblocked them and added a noindex tag so Google could see it and drop them properly.
• Deleted freelancer profiles. When freelancers deleted accounts, their pages were flagged as errors instead of handled cleanly. We made them return a proper 404.
• Page speed. With engineering we resized images across the site and removed dynamic images that slowed loading.

I also worked through duplicate URL and localization issues caused by UTM parameters, as an ongoing workstream.

Result

50+ fixes shipped over my tenure, and improvement of core web vitals score.

Experiment-led testing and optimization

HubSpot, International AEO/GEO Strategist, February 2026 to present. Five markets, with four experiments shipping per six-week sprint.

On-page optimization is necessary to be referenced in AI search, but it isn’t the only factor at play. AI assistants draw on a wide mix of sources, so I expect off-site signals to matter a great deal: independent reviews, educational content and video. That means AEO work has to reach beyond the SEO team into the teams that produce those assets.

How I work

• Build a set of prompts that represent real customer questions.
• Track which sources and brands get cited across assistants over time.
• Compare citation patterns before and after changes to reviews, educational content, competitive content, video and other experimental content sources.

Where I am now

We run on a six week sprint model with four international experiments shipping per sprint. As project owner I lead by designing the experiments, coordinating with EN and international teams, and replicating wins across markets.

Growing the blog 697% in a year

Livestorm, Content Manager, June 2019 to May 2023. B2B webinar platform.

What I did

I owned the blog and its SEO output, managing an agency and a team of freelance writers, and built strong on-page content practices. I conducted topic research and implemented a cluster strategy complete with blog articles, pillar pages, webinars, ebooks and videos.

Result

• Organic blog traffic grew 697% year over year.
• The blog became 46% of English organic website traffic.
• Demand generation grew 200% year over year.

What I learned

How to structure an entire content flow and build the machine to make it happen on a continuous basis.